OPINION AND ORDER DENYING PLAINTIFF’S MOTION TO COMPEL RETURN OF ‘DELOITTE REPORT’
BRAZIL, United States Magistrate Judge.I. INTRODUCTION: ISSUE, TEST APPLIED, AND HOLDING
The principal issue we address is whether plaintiff has shown that a document it inadvertently produced in response to a discovery request was “prepared in anticipation of litigation” as that phrase is used in Federal Rule of Civil Procedure 26(b)(3)(“Rule 26(b)(3)”) to fix the outer boundaries of the work product doctrine. To resolve this issue we have had to grapple with the meaning of the test that has been adopted by several courts of appeal, most notably the Second Circuit in the case of United States v. Adl-man, 134 F.3d 1194 (2d Cir.1998). In part because the nature of the document in the case at bar is so different from the nature of the document in issue in Adlman, we have had to explicate carefully the critical passages in that opinion and to determine how to extend the test so it can be used in the setting we face.
For reasons set forth in detail below, we have concluded that when a court is trying to decide whether a document was “prepared in anticipation of litigation” it should apply a two-stage test. In the first stage, the court should determine whether the party trying to invoke work product protection has shown that the prospect of litigation was *331a substantial factor in the mix of considerations, purposes, or forces that led to the preparation of the document. If, but only if, the party trying to invoke the protection makes this showing, the court proceeds to the second stage of the analysis. In this second stage, the court focuses on the policy objectives that the work product doctrine has been developed to promote — then determines whether (and to what extent) denying Rule 26(b)(3)’s protections to the document would harm those objectives (or, the extent to which conferring that protection would advance the policy purposes that inform the work product doctrine.) The court would conclude that the document comes within the ambit of the Rule (was “prepared in anticipation of litigation”) on a showing that a contrary conclusion would likely frustrate or interfere (more than minimally) with the promotion of the principal objectives this doctrine is designed to serve.
In the case at bar, we have concluded that plaintiffs showing satisfies the requirements of the first stage of the test but does not satisfy the requirements of the second stage. We therefore hold that the document does not fall within the ambit of Rule 26(b)(3). It follows that defense counsel need not return it. Whether, or how, defense counsel might use the document in this litigation is, of course, an entirely separate question that is not before the court at this juncture.
II. FACTUAL 1 AND PROCEDURAL BACKGROUND
Plaintiff JumpSport is a small company engaged in the business of designing, manufacturing and selling trampoline enclosures, trampolines, and related games and accessories. Complaint For Infringement of Patents, filed December 19, 2001, (“Complaint”), ¶ 3. It owns two United States patents for trampoline safety enclosures, U.S. Patent No. 6,053,845, issued on April 25, 2000 (“the 845 patent”), and U.S. Patent No. 6,261,207, issued on July 17, 2001 (“the 207 patent”). Supplemental Declaration of Steven W. Moulton In Support Of Motion For Return Of An Inadvertently Produced Privileged Document (“Supplemental Moulton Declaration”), ¶ 3.
JumpSport began marketing its trampoline enclosure product in 1997, before filing its first patent application, and continued to market and sell trampoline enclosure products throughout the pendency of those applications. Supplemental Moulton Declaration, H 4. By the end of 1998, defendant Jumpking, Inc. (“Jumpking”), a large manufacturer of trampolines, began offering its own trampoline enclosure product. Thereafter, defendants Hedstrom Corporation (“Hedstrom”) and Variflex, Inc. (“Variflex”), both large sporting goods manufacturers, also brought competing trampoline enclosure products to market. Supplemental Moulton Declaration, II5.
At all times relevant to the instant motion, JumpSport believed that the trampoline enclosures sold by defendants were copied from JumpSport’s product, and that defendants’ products infringed its pending patent application for the 845 patent (and the subsequently filed continuing application for the 207 patent). Supplemental Moulton Declaration, ¶ 6. On a few occasions in 1999 and 2000, prior to the issuance of the 845 patent, JumpSport management spoke with Jumpking management to discuss the possible formation of a business relationship between the two companies as a means of resolving the anticipated patent infringement dispute without the necessity of litigation. Supplemental Moulton Declaration, ¶ 8.
While plaintiff was interested in negotiating business solutions with its competitors, it knew that it might be forced to sue in order to protect its intellectual property rights. It also knew that, as a modest sized start-up company, its resources would be considerably strained by the high cost of such litigation. With these high costs in mind, in February of 2000 JumpSport began preparing an application for “patent enforcement” or “patent pursuit” insurance coverage through a company called Litigation Risk Management, Inc. (“LRM”), which specializes in preparing the materials necessary to complete an applica*332tion for this kind of insurance. Supplemental Moulton Declaration, ¶ 11. Between February and October of 2000, JumpSport did not actively pursue the insurance application because of financial constraints. In October of 2000, when talks with Jumpking regarding a possible business relationship appeared to be going nowhere,2 JumpSport decided to go forward with the application. Supplemental Moulton Declaration, ¶ 11.
JumpSport was required to submit several documents as part of its insurance application, including an assessment of the value of the company and of the patents it held, as well as an independent legal analysis of its patent rights by a lawyer with expertise in intellectual property. Supplemental Moulton Declaration, ¶ 13. LRM required JumpSport to commission the first report (the financial assessment report) from the accounting and consulting firm Deloitte & Touche (“Deloitte”). Supplemental Moulton Declaration, ¶ 13. Accordingly, in the late fall of 2000, JumpSport commissioned Deloitte to prepare a report assessing the value of JumpSport and its intellectual property. Supplemental Moulton Declaration, ¶ 16.
At the end of January of 2001, Deloitte submitted the first draft of its report, entitled, “JumpSport, Inc., — Fair Market Value Of Invested Capital and Intellectual Property As Of November 15, 2000.” This draft contains a brief description of the economy in the United States and of the sporting goods market, a company overview of JumpSport, a valuation of JumpSport’s invested capital and its intellectual property, and a valuation synthesis and conclusion, including dollar estimates of potential damage settlements for alleged- patent infringement.3 JumpSport, Inc. — Fair Market Value of Invested Capital and Intellectual Property as of November 15, 2000 (“JumpSport Report”). The report discloses no reasoning about law or litigation.
Because of the high cost of the patent litigation insurance, JumpSport ultimately chose not to proceed with its application. Supplemental Moulton Declaration, ¶ 23. It did decide, however, to proceed with litigation. In December of 2001, JumpSport filed suit against defendants Jumpking, Hedstrom and Variflex for patent infringement and misrepresentation.
In late July of 2002, in response to discovery propounded by defendant Jumpking, plaintiff produced a large quantity of documents that inadvertently included the first draft of the Deloitte report. Declaration of Christine A. Amatruda In Support of Motion To Compel Return Of An Inadvertently Produced Privileged Document (“Amatruda Declaration”), ¶¶ 3,4. When it discovered this error in early November of 2002, JumpSport promptly asked Jumpking to return the document — contending that it was protected by the work product doctrine. Amatruda Declaration, ¶ 7. Counsel for Jumpking, whose damages expert had examined the report before JumpSport sought its return, was not persuaded — and refused to return the document. Declaration of C.J. Veverka In Support of Defendants’ Opposition To Plaintiff JumpSport’s Motion To Compel Return Of An [Allegedly] Inadvertently Produced Document (“Veverka Declaration”), ¶ 4; Amatruda Declaration, ¶ 14. JumpSport then filed this motion, asking the Court to order Jumpking to return the draft report.'1 Thereafter, Jumpking’s expert refined the opinions he was preparing for the damages phase of this litigation, citing the Deloitte draft twice in his lengthy rebuttal report. Declaration of Lance E. Gunderson In Support of Defendants’ Supplemental Brief Filed in Opposition To Plaintiff JumpSport’s Motion To Compel Return Of An [Allegedly] Inadvertently Produced Document (“Gunderson Declaration”), ¶¶ 4-8.
III. DOCTRINAL CONTEXT
Plaintiffs motion requires the court to address difficult questions about how to determine whether the work product doctrine reaches or covers given material. We will attempt to refine our understanding of the *333elements of the analytical path that courts are to follow when addressing this threshold question.
At the outset, we emphasize that our focus will be on whether the doctrine applies at all to given material (or parts thereof), not on the additional issues a court must address in determining whether, ultimately, to deny a party discovery access to a document under this doctrine. Courts are required to address these additional issues only if they first conclude that the doctrine reaches or covers the material in question. If they answer that question in the affirmative, courts go on to determine which of two categories the contested material falls within: is it “opinion” work product or “non-opinion” work product? 4
In most (but not all) instances, if the material is deemed “opinion” work product, the analysis is over — the document will be protected against discovery probes absent proof of waiver or some other extraordinary circumstance. Fed.R.Civ.P. 26(b)(3). But if the material is deemed “non-opinion” work product, the court proceeds to determine whether the party seeking the discovery has shown (1) that is has substantial need of the materials in the preparation of its case and (2) is unable, without undue hardship, to obtain the substantial equivalent of the materials by other means. Fed.R.Civ.P. 26(b)(3). These additional issues are not the focus of our concern. Instead, we struggle with the first question courts must consider in this doctrinal arena: how to reason reliably about whether particular material qualifies as work product at all, i.e., whether the doctrine extends any coverage to the material in question.
Under Rule 26(b)(3), the concept that dictates whether the work product doctrine extends any coverage at' all to given material is couched in these words: was the material “prepared in anticipation of litigation or for trial” by a party or its representative or agent? The rub, which has produced quite a fuzzy genie, is the difficulty of determining how to give meaning to the phrase “in anticipation of litigation”.
Before exploring the meaning of this phrase, however, we note that in some of the cases there appears to be some play (sometimes unacknowledged) between the meaning that is given to this phrase (how far it is permitted to extend, and the test used to fix its boundaries), on the one hand, and, on the other, the category within which the court deems the material to fall. In other words, some courts may vary the test (or how vigorously they apply the test) for determining whether the doctrine reaches the material in question with the category of material in issue: if the material is core “opinion” material, the test for determining whether that material was prepared “in anticipation of litigation” might be less demanding than if the material in issue is “non-opinion”.
The real world upshot of this is that the work product doctrine is more likely to be available to protect core opinion material than the less sensitive non-opinion material. As a matter of logic, permitting such “play” to occur could be dangerous. If not restrained, it could make reasoning about how to fix the scope of the phrase “in anticipation of litigation” circular, at least when the material in issue consists of opinions about a legal dispute. On the other hand, as a matter of common sense it is not surprising that courts look to the content of the material in issue to see if that content offers clues about what the anticipations, needs, possible uses, or purposes were that animated the decision to have the document prepared. Sometimes the content of the material will be so explicitly focused on what to do in particular litigation that, for all practical purposes, the content itself answers the question about whether the document was prepared in anticipation of litigation. But sometimes the content will be uninstructive with respect to this issue — or even misleading. The content might make it appear that the material was prepared in anticipation of litigation — but the court might learn from other evidentiary sources that multiple purposes inspired the *334preparation of the material, none of which, at least directly, was rooted in dealing in some way with possible litigation.
The point here is that we must be aware of and open about what we are doing analytically. We must recognize that the first issue in work product analysis (i.e., does the doctrine even reach the material in question) is formally separate from issues that arise later in analytical sequence (e.g., the issue of which of the two categories the material falls within). And if we are going to permit the nature or elements of the test that we use when we are trying to make the first determination (whether the document was “prepared in anticipation of litigation”) to be influenced by the category in which the material would fall, we must say so clearly and explain why.
Moreover, any such explanations should be rooted in and driven by the principal policy objectives that the work product doctrine was designed to promote. Those policy objectives, as we shall see, might well justify at least some play between the kind of material in issue and the nature of the test used to determine whether that material was “prepared in anticipation of litigation.”
IV. POLICY OBJECTIVES OF THE WORK PRODUCT DOCTRINE
As we understand the authorities, there are two (related) sets of policy objectives that the work product doctrine is primarily intended to advance. One is to free lawyers, their clients and their clients’ other agents to think frankly, critically, and creatively about litigation-related matters — and to be able to record their thoughts so they are not lost, or inaccurately remembered, or developed illogically or clumsily because they are not written down (or otherwise preserved in a form that assures access to them and permits their manipulation and development.) Thus, one goal of the work product doctrine is to create a “zone of privacy” within which such matters can be secure from discovery by parties with adverse interests. See United States v. Meyer, 398 F.2d 66, 74 (9th Cir.1968); see also United States v. Nobles, 422 U.S. 225, 238, 95 S.Ct. 2160, 45 L.Ed.2d 141 (1975). Such protection, it is assumed, enhances the creativity and improves both the candor and the quality of thinking by parties and their agents about litigation-related matters — and thereby improves the quality of legal services and enables clients to consider squarely all aspects of their situations and all the implications of alternative litigation-related paths they might follow.
The second primary policy objective of the work product doctrine is to preserve and enhance the set of incentives that are said to be essential to the effective operation of an adversary system of civil justice. Hickman v. Taylor, 329 U.S. 495, 511, 67 S.Ct. 385, 91 L.Ed. 451(1947). That system is primarily party and lawyer driven — unlike continental systems, where the court or some other instrumentality of the “state” assumes considerable responsibility for uncovering and interpreting the evidence and for identifying accurately all the legal principles that bear on civil disputes. If a party/lawyer driven system is to work effectively, each party and lawyer must work hard, from their independent perspectives, to find the relevant evidence and to explicate the relevant law. The theory of our system is that adversaries will have different perspectives on and different evidentiary and legal interests in any given dispute — and that because of those differences they will search at least in part in different ways or in different places for evidence and law. The system further assumes that highly motivated searches simultaneously undertaken from different perspectives are likely to generate the most comprehensive data (evidentiary and legal) and the widest array of interpretations of it — and it is that comprehensiveness of data, and the clash of views about what inferences should be drawn from it, that enable the adversary system to deliver reliable results.
For this system to function as intended, each party must be motivated to work hard and independently. The work product doctrine has been crafted, in substantial part, to help protect both this motivation and this independence. Thus, Justice Jackson endorsed the work product doctrine out of concern that its protections were necessary to prevent both the demoralization of the legal profession and the erosion of investigatory *335and analytical independence that would occur if courts permitted one party or lawyer simply to sit back and wait until his opponent had done his legal and evidentiary homework, then “borrow” (through civil discovery) the product of that work and the “wits” that informed it. Id. at 516, 67 S.Ct. 385 (Jackson, J., concurring). The prospect of such easy “borrowing” would discourage both the conscientious and the lazy practitioner from devoting the effort to pretrial preparation on which the system depends. Admiral Ins. Co. v. United States District Court, 881 F.2d 1486, 1494 (9th Cir.1989). Thus, in theory, the ultimate purpose of the work product doctrine is to help protect the quality of the truth finding process — and, thereby, the integrity of our system of justice.
How close the connection might be between this theory and the realities of everyday civil practice is well beyond this court’s capacity to judge. But it is significant, for the dispute before us, that our confidence in the reality of the connection grows in direct correlation to the proximity of the material in issue to core opinion work product. The more remote the material is from the heartland of work product sensitivity, the more likely it is to consist of straightforward historical facts or data about real world events — material that exists independent of the litigation process, in something like the public domain — material likely to be anticipated and pursued by any rational mind trying to develop a picture of what really happened — material to which the modern rules of civil disclosure and discovery are more likely to provide access (at least once separated from the process that opposing counsel used to find the material and the packages in which she stores it).
Core opinion work product, on the other hand, consists of confidential analyses, impressions, inferences, predictions, strategies, plans — material that would not exist absent the litigation, material that is in no sense in the public domain, material that is not part and parcel of the key historical events on which parties’ rights turn. Clearly, it is this kind of material whose forced disclosure could do the greatest harm — both to the morale of the profession and to the candor and quality of the thinking on which parties must depend when they make important decisions about whether and how to try to protect their rights in the civil justice system.
V. APPROACHES TO DETERMINING WHETHER A DOCUMENT WAS “PREPARED IN ANTICIPATION OF LITIGATION”
Armed with these understandings of the fundamentals, we turn to the task of giving meaning to the phrase “prepared in anticipation of litigation.” The court of appeals for the circuit in which we sit (the Ninth Circuit) has not definitively addressed this issue— and has not expressed a preference for either of the two competing general approaches that have been endorsed by other federal courts of appeals. Nor, to our knowledge, has any court of appeals had occasion to decide how to go about determining whether a document of the kind that JumpSport inadvertently produced in this case was “prepared in anticipation of litigation.” Thus, we work in an arena without clear and controlling doctrine — an arena in which no court has provided a detailed analytical guide for resolving the specific problem we face.
This is not to say that we work in a tabula rasa. Many courts have tried to clarify the meaning of this phrase, and federal courts of appeal have adopted at least two quite different standards, both rather general, for use in determining whether particular documents were “prepared in anticipation of litigation.” The first of these general standards dates from at least 1981, when the Fifth Circuit issued its opinion in United States v. Davis, 636 F.2d 1028 (5th Cir.1981). Under Davis, courts are to determine whether a document was “prepared in anticipation of litigation” by identifying the “primary motivating purpose” that underlay the preparation of the document. Under this standard, the work product doctrine extends to a document only if the party trying to invoke the doctrine’s protection first establishes that the “primary motivating purpose behind the creation of the document was to aid in possible future litigation.” Id. at 1040. Thus, even if a secondary purpose was directly related to *336litigation, no protection is available under this formulation of the test if the primary-purpose underlying the preparation of the document was to inform a business decision, to facilitate a business transaction, or to satisfy a regulatory requirement. Id.
This approach is accompanied by no small number of difficulties, one of which is uncertainty about how to distinguish primary from secondary purposes. Are courts to focus on the use of the document that was or would be closest in time to the preparation of the document, or should they try to identify the document’s “ultimate” purpose? How are courts to ascribe relative weight or significance to multiple purposes? If either of two purposes clearly would have been sufficient to cause the document to be prepared, on what basis should courts decide which one of the two was the “primary” purpose?
The second (and now apparently more widely endorsed) general standard for determining whether a document was “prepared in anticipation of litigation” is associated most closely with the Second Circuit’s decision in United States v. Adlman, 134 F.3d 1194 (2d Cir.1998). In Adlman, an aerospace manufacturer, Sequa, contemplated merging with two of its subsidiaries. The contemplated merger was expected to produce an enormous loss and tax refund, which Sequa expected the IRS to challenge, resulting in litigation. Sequa asked an accountant/lawyer from the consulting firm Arthur Anderson to evaluate the tax implications of the proposed restructuring. Adlman, 134 F.3d at 1195. In response to this request, the Anderson accountant/lawyer prepared a fifty eight page legal memorandum analyzing likely IRS challenges to the reorganization. The memorandum discussed statutory provisions, IRS regulations, legislative history, and prior judicial and IRS rulings relevant to the claim. It also proposed possible legal theories or strategies for Sequa to adopt in response, recommended preferred methods of structuring the transaction, and made predictions about the likely outcome of the litigation. Id. After studying the memorandum, Sequa decided to proceed with the restructuring. As anticipated, litigation with the IRS resulted. During discovery, Sequa referenced the Anderson memorandum in its privilege log, but refused to produce the document on the grounds that it was protected by attorney-client privilege and the work product doctrine. The IRS moved to compel production of the memorandum. Id. at 1195-96.
The issue, as framed by the Adlman court, was “whether Rule 26(b)(3) is inapplicable to a litigation analysis prepared by a party or its representative in order to inform a business decision which turns on the party’s assessment of the likely outcome of litigation expected to result from the transaction.” Id. at 1197.
The Adlman majority assumed that if it used the “primary motivating purpose” test to resolve this issue the document would be entitled to no protection at all under the work product doctrine. The document’s most immediate and obvious purpose, after all, was to “inform a business decision” — not to assist a party prepare for or participate in litigation. But because of the nature of the document — analysis of and strategies for litigation that was virtually certain to ensue if the transaction was consummated — the court stepped back and subjected the “primary purpose” standard to critical review. In that review, the court identified several scenarios in which use of the “primary purpose” test would force courts to deny any work product protection to documents that consisted of pure litigation analysis — just because the most immediate purpose of their preparation was to help a party decide whether to proceed with a transaction or to develop a reliable picture of a party’s financial health. Id. at 1199-1200.
It followed, in the view of the Adlman majority, that use of the “primary purpose” test would create obstacles to business transactions and to developing accurate information needed to inform sound judgments about future courses of action. More significantly, use of the “primary purpose” test in these settings would contravene core policy objectives of the work product doctrine. If courts insisted on using this test, there would be no “zone of privacy” in which parties who actually anticipated litigation — litigation with potentially enormous consequences — could secure the frank and careful assessments of *337how that litigation was likely to play out that they need to move forward with their plans. Id. at 1201-02.
Concluding for these reasons that the “primary purpose” test was inadequate, the Adlman court turned to the standard that is articulated in the Wright, Miller & Marcus treatise,5 and that had been adopted in several other circuits. See In re Grand Jury Proceedings, 604 F.2d 798, 803 (3d Cir.1979); National Union Fire Ins. Co. v. Murray Sheet Metal Co., Inc., 967 F.2d 980, 984 (4th Cir.1992); Binks Mfg. Co. v. National Presto Indus., Inc., 709 F.2d 1109, 1118-9 (7th Cir.1983); Simon v. G.D. Searle & Co., 816 F.2d 397, 401 (8th Cir.), cert. denied, 484 U.S. 917, 108 S.Ct. 268, 98 L.Ed.2d 225 (1987); Senate of Puerto Rico v. United States Dep’t of Justice, 823 F.2d 574, 586 n. 42 (D.C.Cir. 1987). Under that standard, a document is deemed “prepared in anticipation of litigation” if “in light of the nature of the document and the factual situation in the particular ease, the document can fairly be said to have been prepared or obtained because of the prospect of litigation.” Id. at 1202 (emphasis in Adlman).
The Second Circuit panel in Adlman squarely endorsed this formulation of the standard after concluding that it would enable courts to extend protection in appropriate circumstances to documents reflecting core litigation analysis and strategy — and thus would enable courts to advance, rather than frustrate, the policy objectives of the work product doctrine. According to the Adlman majority, using this “because of’ test also would enable courts to deny protection in circumstances that would not advance the underlying policy objectives of the work product doctrine — e.g., as would be the case, in that court’s view, for documents “prepared in the ordinary course of business6 or that would have been created in essentially similar form irrespective of the litigation.” Id.
VI. ANALYZING ADLMAN AND REFINING A TWO STAGE TEST FOR DETERMINING WHETHER A DOCUMENT WAS “PREPARED IN ANTICIPATION OF LITIGATION”
As we have indicated in another setting, United States v. Chevron Texaco Corp., No. 3:01-CV-04243, 2002 WL 31956025 (N.D.Cal. 2002), the opinion by the Adlman majority persuaded us that the Wright, Miller & Marcus formulation of the standard better serves the policy considerations that inform work product doctrine than the “primary purpose” test endorsed by the Fifth Circuit.7 When we tried to apply Adlman’s teachings in the case at bar, however, we found passages within the opinion that we could not reconcile. We also identified important questions that the opinion leaves unanswered. More significantly, we found that some passages in the opinion, if taken out of context, could be used to deny work product protection to documents whose disclosure would strike at the heart of work product policy. Similarly, we found other passages that could be used (again, out of context) to extend the reach of work product protection to documents whose disclosure would not implicate core work product policies at all. Our struggle with these problems persuaded us that there was a need to clarify and restructure the test that Adlman endorsed for determining whether a document was prepared “in anticipation of litigation”.
The most important question that the Adl-man opinion, at least on the surface, seems to beg is this: what does “because of’ mean (in work product analysis)? What criteria are courts to use when they try to decide whether a document was prepared “because of anticipated litigation”? How elastic or rigid is that concept? If there is some elasticity in this phrase, what are the considerations or factors that determine how much *338play it should be given in any particular situation?
The phrase “because of’ suggests an inquiry that would focus on something in the nature of “causation”. The key word is “because” — a combination of “be” and “cause.” And according to Mr. Webster, the phrase “because of’ means “by reason of; on account of.” Webster’s II New Riverside Dictionary 161 (1984 ed.). But if we are to focus on causation, what kind of causation are we talking about? Cause in fact? Legal cause (synonymous for our purposes with “proximate cause”)? Some other kind of “causation”?
The significance of this last set of questions is graphically exposed when we observe that the plaintiff pushes us energetically (but without complete conceptual self-awareness) to use a definition of causation that is rooted in the notion of “cause in fact,” while the defendant urges us with comparable vigor to adopt an approach that has much more in common with the notion of “legal cause” or “proximate cause.” Plaintiff advocates a “cause in fact” approach because it is easier to satisfy — and would leave little or no additional room for assessing how disposition of the issue would affect the policy objectives of the work product doctrine. Defendant, in contrast, contends that it is only through application of a broader, more flexible test, akin in spirit, at least, to “proximate cause,” that the court can do justice to those policy objectives. To these party opponents, it seems that the outcome of the whole matter turns on which of these two approaches we adopt.
We begin our effort to address the issues just raised by examining the two, baek-to-back sentences near the beginning of Adl-man in which the court sets forth the key legal propositions that it endorses in this opinion. In the words of the majority opinion: “We hold that a document created because of anticipated litigation, which tends to reveal mental impressions, conclusions, opinions or theories concerning the litigation, does not lose work-product protection merely because it is intended to assist in the making of a business decision influenced by the likely outcome of the anticipated litigation. Where a document was created because of anticipated litigation, and would not have been prepared in substantially similar form but for the prospect of that litigation, it falls within Rule 26(b)(3).” Adlman, 134 F.3d at 1195. Evidencing how great the differences between the notions suggested by these two sentences might be, JumpSport (the plaintiff) would have our disposition of this motion turn on the second sentence,8 while Jumpk-ing (the defendant) would have that disposition turn on the first sentence.
When we consider these two sentences carefully we reach two conclusions. First, only the first of the two sentences articulates a holding — because only that sentence addresses the kind of document that was in issue in that ease. Our other conclusion is that the second sentence, if taken literally and out of context, announces a principle of much broader reach — but the reasoning and the examples in the opinion are not sufficient to support that apparent breadth. In other words, the court’s careful consideration of policy and its creative exploration of hypo-thetieals amply support its actual (rather narrow) holding, but offer little or no support for using the much broader test that the *339court seems to set forth in the second sentence.
In studying Adlman we also noticed that when the court used its own words to articulate the broader test it added a concept to and omitted some potentially important qualifiers from the version in the Wright, Miller & Marcus treatise. The treatise, as quoted with express approval by the Adlman majority, declares that a document is “prepared in anticipation of litigation” if “in light of the nature of the document and the factual situation in the particular case, the document can fairly be said to have been prepared or obtained because of the prospect of litigation.” Id. at 1202 (underlines are ours; italics are Adlman’s). In contrast, the Second Circuit panel articulated the broader test in these words: “Where a document was created because of anticipated litigation, and would not have been prepared in substantially similar form but for the prospect of that litigation, it falls within Rule 26(b)(3).” Id. at 1195.
This articulation by the Adlman majority fails to remind courts to take into account (1) the nature of the document and (2) the factual situation in which the document was generated. It also omits the potentially significant phrase “can fairly be said” that qualifies the words “because of’ in the formulation in the treatise. These three concepts (nature of document, factual situation, and “can fairly be said”) expressly call for case-specific flexibility in applying this standard. They also may imply that the “because of’ that the treatise endorses is closer in spirit to the tort notion of “legal cause” (a.k.a. “proximate cause”) than to “cause in fact.”
In many tort law settings, the boundaries of proximate cause are set primarily by policy considerations — not by our often uncertain understandings of physics or “principles” of social dynamics. W. Keeton, D. Dobbs, R. Keeton & D. Owen, Prosser and Keeton on Law of Torts, § 41 (5th ed.1984). Similarly, by inserting the otherwise unnecessary concept of whether it “can fairly be said” that a document was prepared because of the prospect of litigation, the Wright, Miller & Marcus formulation appears to invite courts to give appropriate weight (through the word “fairly”)9 in this analysis to the policy objectives that underlie and justify the work product doctrine. Instructively, the Adlman majority gives considerable weight to such policy matters when it explains its decision not to use the “primary purpose” test to determine whether the document in issue in that case was “prepared in anticipation of litigation.”
We recognize that it is dangerous (risks analytical error) to import concepts or terms from one arena of legal doctrine (e.g., tort law) into a quite different arena (work product) — at least without first carefully considering the implications of the differences in *340environments. We also recognize that the Second Circuit panel neither expressly admitted to any such importation nor attempted to assay the contextual dissimilarities that might counsel against any such move. The vocabulary and concepts the Second Circuit panel chose to endorse, however, (e.g., “because of’ and “but for”) suggest the influence of “causation” ideas developed in tort law— and that it might be instructive to consider those kinds of ideas as we try to refíne our thinking about what the Wright, Miller & Marcus test means and how to apply it in the setting we confront.
The “but for” proviso in the Second Circuit’s version of the “because of’ test seems to draw on the concept of “cause in fact” from tort theory. In tort law, the “cause in fact” requirement generally operates much less as a limitation (on the scope of exposure to liability) than the requirement of “legal cause” (more commonly referred to as “proximate” cause). Commentators have noted that the “cause in fact” requirement is useful for identifying defendants who are excluded from liability, but is appreciably less useful for identifying defendants to whom liability actually attaches.10 In other words, while “cause in fact” is necessary for liability, it is by no means sufficient. In determining what will be deemed sufficient to establish liability the main analytical chance is much more often located in the arena of “legal cause”.
Thus, if the Adlman majority intended to draw on tort theory, the “but for” requirement would be used as a threshold, screening requirement — to identify documents that could be excluded readily from work product protection. If we follow the analogy to tort theory, however, the fact that a document would not have been prepared “but for” the prospect of litigation would not be a sufficient basis for concluding that it fell within the reach of the work product doctrine— would not necessarily mean that it was prepared “because of’ the prospect of litigation. Courts would need to make additional inquiry before they could reliably answer that question.
If tort theory serves as a source of some instruction here, we would expect that the content of this second level of inquiry would be informed primarily by policy considerations — just as it is in developing the content of “legal cause” or “proximate cause” in various branches of tort law. And in the case at bar, the policies to which we would turn to define the content of this second level of inquiry are the policies that have driven (and been accommodated in) the development of the work product doctrine.
As we will explain momentarily, there are places in Adlman where the court’s reasoning seems to depart from or blur the analysis that would be suggested by the analogy to tort theory — so we cannot be certain how fully the Second Circuit panel would endorse that theory as a source of analogical guidance in the work product arena. But viewed as a whole, Adlman suggests that a two stage inquiry of the kind we have described above is appropriate in this setting — and that in the second stage courts should focus primarily on the policies that inform work product doctrine to determine whether to conclude that a given document was prepared “because of the prospect of litigation.” Evidence of the accuracy of the latter inference permeates the opinion — whose primary preoccupation is with explaining how much harm to the policies that underlie the work product rules would be caused if courts, blindly applying the “primary purpose” test, concluded that no protection could attach to documents that analyze litigation just because those documents were prepared in the first instance to help a party decide whether to proceed with a transaction that the party felt would provoke the litigation that is the subject of the confidential analysis.
As mentioned above, there are two apparently significant places in Adlman where the majority opinion seems to depart from or blur the two stage approach to determining how far Rule 26(b)(3) reaches. Before *341launching our close examination of these passages, however, we should set forth some of the general concerns that we will explicate in the paragraphs that follow. At least by implication, the two passages on which we will focus seem to contradict, or undermine, our inference that the Adlman majority would endorse a two stage test that is something akin to the “cause in fact” and “legal cause” inquiries in the field of torts.11 They also expose some of the analytical mischief that can be provoked, unknowingly, when courts feel constrained to use the “but for” definition of “cause in fact” in the work product setting.
Moreover, there appears to be analytical tension between, on the one hand, the two passages we will examine and, on the other, the majority’s broader formulation of the “test” for determining whether a document was “prepared in anticipation of litigation.” As reported above, Adlman articulates the broad formulation of the test in these words: “Where a document was created because of anticipated litigation, and would not have been prepared in substantially similar form but for the prospect of that litigation, it falls within Rule 26(b)(3).” The addition of this “but for” proviso, and the use of the conjunction “and,” certainly imply that the two inquiries are not coterminous (if they were, one of the two clauses would be superfluous). But in the passages that we will examine in the paragraphs that follow the Adlman majority seems to read the second prong (the “but for” inquiry) of its broad formulation of the test back into what would be the first prong (the “because of’ inquiry) of that test. In fact, for documents whose preparation was directed at a time when a party “had the prospect of litigation in mind,” the Adlman majority seems to subsume the first question in the second.
Thus, we will examine these two passages closely for several reasons: (1) they seem significant to the Adlman majority, (2) taken at face value, they seem to contradict, in part, our inference that the Adlman court would endorse the test we propose, (3) they seem to contradict or undermine other parts of Adlman, as well as what we understand to be the broader and central conceptual thrust of that opinion, and, finally, (4) our close examination will help refine our thinking about how to give a meaning to the phrase “because of the prospect of litigation” that is appropriately responsive to the values the work product doctrine is intended to promote.
The first of the two passages on which we will focus appears in the section where the majority describes and defends the articulation of the test in the Wright, Miller and Marcus treatise. After noting that under this “because of’ formulation of the test a document that was “created because of the prospect of litigation, analyzing the likely outcome of that litigation,” would “not lose protection ... merely because it [was] created in order to assist with a business decision,” the Adlman panel hastens to add the following observations:
the ‘because of formulation withholds protection from documents that are prepared in the ordinary course of business or that would have been created in essentially sim*342ilar form irrespective of the litigation. It is well established that work-product privilege does not apply to such documents, [citing the Advisory Committee Note to the 1970 amendments to Rule 26(b)(3) and National Union Fire Ins. Co. v. Murray Sheet Metal Co., Inc., 967 F.2d 980 (4th Cir.1992)]. Even if such documents might also help in preparation for litigation, they do not qualify for protection because it could not fairly be said that they were created ‘because of actual or impending litigation. See Wright, Miller & Marcus sec.2024, at 346 (‘even though litigation is already in prospect, there is no work-prod-uet immunity for documents prepared in the regular course of business rather than for purposes of the litigation’).
Id. at 1202 (emphasis added).
The Adlman majority does not try to explain (anywhere) how this passage can be squared with the suggestion (clearly made at the outset of the opinion) that there are two separate components to the test for determining the scope of Rule 26(b)(3). In the sequence that is suggested by the Adlman opinion (but that we think should be reversed), the court determines in one component of the inquiry whether the document “was created because of anticipated litigation,” while in the other component the court determines whether the document would “not have been prepared in substantially similar form but for the prospect of that litigation.” Id. at 1195. The court did not explain why, if the document fails the “but for” test, there would be any occasion to consider the “fairly be said” issues. Moreover, the court did not explain why (what reasoning or principles lead to the conclusion) it would not be “fair” to say that such documents were prepared “because of’ the prospect of litigation.
Another difficulty with this passage is that it seems to invite courts to re-inject “primary purpose” reasoning into the equation. In fact, this approach seems to use “primary purpose” reasoning to automatically exclude from any possible work product protection documents that consist of nothing but “pure litigation analysis” — whenever it appears that the documents were prepared “in the ordinary course of business.” We explore these and other concerns after we describe, in the next paragraphs, the second passage that we have had difficulty reconciling with Adlman’s holding and the central thrust of its reasoning.
The second passage is found in part III, where the court explains its decision to remand the work product issue and gives instructions to the district judge. The majority acknowledges at the outset of this section that it cannot tell whether the district judge applied the primary purpose test or the test from the Wright, Miller & Marcus treatise. While in some places the district court seemed (to the court of appeals) to focus on trying to determine the primary purpose for which the document was prepared, “the tenor of the discussion ... suggests [the district judge] may have focused properly on the question whether the [document] ... would have been prepared in substantially similar form regardless of whether litigation was contemplated, and thus was not prepared ‘because of the expected litigation.” Adl-man, 134 F.3d at 1203.
The fact that the district court’s opinion could be understood as applying either test (“primary purpose” or “but for” causation) points to the risk that the latter might, at least in some settings, devolve into the former. That turn of events would frustrate the central purpose of Adlman — which was to empower courts, at least when considering certain kinds of documents (documents analyzing prospective litigation), to look beyond “primary purpose” when determining whether a document would fall within at least the outer limits of the reach of Rule 26(b)(3).
Our confusion is amplified by the paragraphs in which the court of appeals instructs the district judge how to proceed, analytically, on remand. After directing the lower court to apply the test articulated in the Wright, Miller & Marcus treatise, the Second Circuit panel declared: “There is little doubt ... that Sequa had the prospect of litigation in mind when it directed the preparation of the memorandum .... Whether it can fairly be said that the Memorandum was prepared because o/that expected litigation really turns on whether it would have been prepared irrespective of the expected litiga> *343tion with the IRS.” Id. at 1204 (underlines and italics added).
This passage suggests that, at least when a party was aware that litigation might ensue from a possible course of conduct, the entirety of the Wright, Miller & Marcus test would be reduced to a “but for” version of the “cause in fact” inquiry. The upshot would be to foreclose inquiry into the impact on work product policies of ordering the document disclosed. But Adlman does not explain why it would make sense to foreclose any such inquiry for documents of the kind in issue in that ease — documents that consist entirely of litigation analysis and commentary.
It may be that the reason the Adlman majority felt comfortable with the “but for” proviso is that it seemed to be rooted in basic work product policy. The court may have reasoned that if a party would have prepared the document in the normal course of business, whether litigation was anticipated or not, compelling the party to disclose the document would not threaten the set of incentives that underpin an effective adversary system. One significant difficulty with this line of reasoning, however, is that it ignores the other primary policy concerns that inform the work product doctrine (the concerns captured in the “zone of privacy” notion). If a given document consisted solely of analysis of litigation (as it did in Adlman), an order that forced a party to disclose the document would implicate, and might do serious harm, to those other primary purposes of the work product doctrine. This fact should give courts serious pause before refusing to extend work product protection to a document simply because a colorable argument has been made that it would have been prepared even if litigation had not been in view.
Another substantial difficulty with the approach suggested in the two passages on which we have been focusing is that the court offers no guidance about how judges are to go about answering the question on which, according to the Adlman majority, the outcome is to turn: would the document have been prepared in any event, in the ordinary course of business, or would it not have been prepared “but for [the party’s] anticipation of litigation .... ” It is important to emphasize that the “ordinary course of business” language does not appear in the Rule itself— and that the Advisory Committee Note does not try either to define this concept or to suggest how its boundaries should be charted. The Note leaves this phrase naked— undefined, and justified only by the suggestion that documents prepared for some purpose other than litigation should not be covered by the Rule.
This leaves trial courts in a very difficult spot. For example, when a party “had the prospect of litigation in mind when it directed the preparation of the [document],” id. at 1204, what criteria is a trial judge to use when she undertakes to determine whether the document would have been prepared in the normal course of business or would not have been prepared but for the prospect of litigation? In the Adlman setting, does this turn out to be a self-answering question (because the party clearly had litigation in mind)? If so, why remand the case and ask the trial judge to answer this question? If not, won’t the trial judge almost always be tempted to look for what the primary, or at least most immediate, purpose of the document was? If that is what the trial judge ends up doing, have we simply taken a circuitous route back to using the “primary purpose” test — at least when the possibility of litigation was known to the party?
Most important, why should it matter, given all the other things the Adlman court teaches about the importance of work product policy, that the most immediate purpose, or even the dominant purpose, was to make a decision about a non-litigation course of conduct — if the document in issue consists entirely of legal analysis of litigation that likely would be triggered by that course of conduct? This kind of document, after all, provokes the most acute work product sensitivities — and would be the most vigorously protected if it fell anywhere within the boundaries of Rule 26(b)(3). So shouldn’t courts be careful not to adopt “tests” for fixing those boundaries whose application would risk excluding from any work product protection, for no clearly articulated countervailing reason, documents that would invade *344at its core that “zone of privacy” that this doctrine is designed to preserve?
Having pondered these questions for some time, we have identified several considerations that might help explain the apparent internal tensions in the Adlman opinion— some of which also help inform our development of a test that is consistent with the larger themes in Adlman and the primary concerns that animated the court’s holding.
First, it appears that the Adlman majority felt cabined by precedent and by a sentence in the Advisory Committee Note that accompanied the 1970 amendments to Rule 26(b)(3). The court quoted a substantial portion of this sentence in the first of the two passages of its opinion that we discussed above. The sentence in its entirety reads: “Materials assembled in the ordinary course of business, or pursuant to public requirements unrelated to litigation, or for other nonlitigation purposes are not under the qualified immunity provided by this subdivision,(italicized words are those quoted in Adlman). Fed.R.Civ.P. 26(b)(3), Advisory Committee’s note.
In the same paragraph Adlman also quoted the Wright, Miller & Marcus treatise to the same effect: “even though litigation is already in prospect, there is no work-product immunity for documents prepared in the regular course of business rather than for purposes of litigation.” Id. at 1202. Because a goodly number of courts had accepted this notion,12 the judges in the Adlman majority likely felt that they had to acknowledge it squarely as a limitation on the reach of the “because of’ standard.
A few observations about this source of constraint are in order before we turn to the other considerations that may help explain Adlman’s apparent internal tensions. First (not in order of importance), we note one interesting difference between the language in the Advisory Committee’s Note, on the one hand, and, on the other, what is supposed to be the parallel language in many cases (including Adlman) and in the Wright, Miller & Marcus treatise. The Note mentions materials “assembled” in the ordinary course of business, while other authorities discuss materials “prepared” in the ordinary course of business. While this difference may not be significant, the word “assembled” as used in the Note may be more reflective of the spirit that informs the two sentence paragraph of the Note in which the pertinent words quoted by Adlman appear. The second sentence in that paragraph of the Note advises readers that, even with the amendment to 26(b)(3), no change is made in the existing doctrine under which “one party may discover relevant facts known or available to the other party, even though such facts are contained in a document which is not itself discoverable.” As used in this setting, the word “assembled” calls to mind evidence or facts that already existed, independent of anticipation of litigation, and that remain discoverable. The word “prepared,” in contrast, may suggest a process with more interactive and dynamic features — a process in which the content or shape of the materials that emerge is much more the product of purposes or needs beyond those simply of collecting.
Given the fundamentals of the accommodation in the work product doctrine between preserving access to real world, pre-litigation evidence, on the one hand, and, on the other, the goals of protecting zones of litigation-related privacy and the incentives necessary for the adversary system, it is quite a bit clearer that materials that were simply “assembled” would not be protected under this doctrine than it might be that materials that were “prepared” in a setting that seems like the normal course of business could receive no such protection. In other words, the cases and the treatise writers may have read this sentence a bit more broadly than the authors of Rule 26(b)(3) intended. The purpose of this observation in the Note may *345simply have been to recognize that there are kinds of materials that are so clearly prepared for “non-litigation purposes” (to quote the Note) that they cannot receive any protection under this doctrine.
More significantly, we have found no reason to suspect that the Advisory Committee had given any consideration at all to how a document of the kind at issue in Adlman should be treated under Rule 26(b)(3) — or to whether it would be consistent with the policies underlying the Rule to conclude that a document of that character, and prepared in that setting, had been “assembled in the ordinary course of business” and was, therefore, ineligible for any protection. The paragraph of the Note that announces that the new Rule would leave intact the notion that documents “assembled in the ordinary course of business” would not qualify for protection cited only two cases,13 both of which involved documents that were generated in response to a governmental mandate and that had nothing in common with the document in Adlman. The Committee Note offers no clue, by comment or case citation, about what it assumed the phrase “assembled in the ordinary course of business” meant — or what criteria or considerations courts should take into account when trying to decide whether to attach this label to a particular document. While subsequent eases have made it clear that this concept is not self-defining, and that more than occasionally courts will be required to resolve this issue, the Note tells us nothing about what the content of that analysis should be.
A related point is that at least some courts and commentators seem to have assumed that the two categories of documents in question here (those that were “assembled in the ordinary course of business” and those that were “prepared in anticipation of litigation”) are mutually exclusive and will invariably be separated in the real world by real space. Wright, Miller and Marcus, for example, point out that work product protection cannot reach “documents prepared in the regular course of business rather than for purposes of litigation.” Wright, Miller, and Marcus, supra note 4, § 2024, at 346. Unfortunately for people on the front lines of the business and litigation worlds, however, this beguiling conceptual tidiness is chimerical. We need look no farther than the facts in Adlman (and the hypotheticals that court discusses in support of its holding) for proof that the assumption that there can be no overlap (in the real world) between these categories is misplaced. Documents can have multiple sources, purposes, uses, effects and implications. It is dangerous to predicate doctrine on an assumption so at war with the facts of life.
Similarly, a good argument can be made that we needlessly risk harm to the policy accommodations the work product doctrine has been designed to fashion, and to the principal purposes it has evolved to serve, if we force courts on the front lines of the judicial system to choose between one category or the other. In a post-Aristotelian era, we should be able to craft an approach to solving sometimes significant and complex problems that is less crude, rooted in more accurate assumptions about the real world, and more responsive to underlying policies.
Another observation about the Adlman opinion cuts generally in the same direction. One of the factors by which this Second Circuit panel seems to have been moved in the two passages we are discussing is the desire to disarm critics (like the dissenter in Adlman, Judge Kearse) who might fear that adopting the “because of’ test could lead to dangerous (to the truth finding process) extensions of work product protection. It appears that one reason the Adlman majority added its version of the “but for” proviso to the Wright, Miller & Marcus test — and emphasized that documents prepared in the ordinary course of business would not be protected under this approach — was to reassure the legal community that the standard that the Adlman majority was endorsing would preserve real limits on the reach of the work product doctrine and would not undermine hoary lines of authority.
This intention to reassure possible critics is visible in the two paragraphs that follow *346immediately after the majority endorses the standard set forth in the treatise. Both of these substantial paragraphs are devoted to explaining how unthreatening that standard is — and to how, applying this standard, district courts still would be able to order disclosure of all but those fractions of documents that survived all the challenges that a party could mount under Rule 26(b)(3). How these last points could have any real bearing on the fate of the kind of document that was actually in issue in Adlman, however, remains something of a mystery. That document, as characterized by the majority, consisted entirely of legal analysis and strategizing — so if it was deemed to have been “prepared in anticipation of litigation” it would almost certainly not have been vulnerable to challenge under any other provision of Rule 26(b)(3).
All this analytical straining appears to have been occasioned, at least in part, by the majority’s assumption that the two categories of documents had to be mutually exclusive and that the court could not endorse an approach that appeared to undermine the “well established” notion that the work product doctrine simply does not 'reach documents “prepared in the ordinary course of business.” Unfortunately, the document in issue in Adlman directly challenged these assumptions and precedents — and it appears that it was the majority’s disinclination to acknowledge that challenge that resulted in an opinion whose central analytical power is compromised toward the end by some confusion and circularity. Our hope, with genuine respect and deference, is to propose a solution that will enable courts to reason about work product with less artificial constraint and more clarity of purpose when they confront the kind of document that was in issue in Adlman.
Our next (and final) set of observations about the two passages we have been examining helps expose the path that led to our formulation of the first stage of the test for determining whether a particular document was “prepared in anticipation of litigation.” We were inspired to turn to the world of tort theory to see if we could find ideas that might be useful in the case at bar by the Adlman majority’s suggestion that courts adopt a two prong test — with two different kinds of “causation” being in issue in the two prongs. In one prong, the causation in issue would have a more limited focus (the “but for” inquiry), while in the other prong the inquiry into causation questions would have a broader reach (the “because of the prospect of litigation” prong). As we struggled to understand the two passages we have found so difficult, we focused increasingly on the “but for” component of the test — because it seemed to be most responsible for trapping the Adlman majority into apparent analytical inconsistencies. Insisting on using a “but for” test seemed to cut off logical access to considering how a decision to deny protection to a particular document might effect core work product policy goals. So we wondered if it really was necessary in this setting to use the “but for” definition of “causation in fact” — or if there was an alternative definition whose use might free the courts to give appropriate analytical play to fundamental work product policies without threatening an unjustifiable extension of the reach of the Rule.
When we studied tort law treatises we learned that the “but for” phraseology was just one of the ways common law courts had attempted to come to terms with the more generic problem of “causation in fact.” In many tort settings, the “but for” test worked well to determine whether a defendant’s conduct was a “cause in fact” of the harm the plaintiff suffered. But not in all settings. For example, when conduct by two or more separate parties clearly played a significant role in bringing about the plaintiffs injuries, rigid application of the “but for” test would relieve all the culpable parties of liability. It was largely to overcome this clearly unacceptable result that common law courts developed a second, more flexible approach to identifying a “cause in fact.” Under this alternative approach, conduct would be deemed to have been a “cause in fact” of the harm suffered if plaintiff proved that the conduct was “a substantial factor” in bringing about that harm. Using this test cleared the way to impose liability on every party whose conduct played a major role in bringing about the harm — a result that is more consonant with our sense of justice and that *347advances (rather than impedes) one of the primary purposes of tort law: to discourage culpable behavior. W. Keeton, et al., supra note 10, at 265-68. As it turns out, using this alternative approach to identifying a “cause in fact” also can deliver considerable advantages in the work product arena.
We infer, from the logical strains we identified above, that the Adlman majority did not attempt to ascertain all the implications (for work product policy) of its insistence on using the “but for” definition of “cause in fact.” Nor does it appear that the Adlman court looked for or considered the possibility of using some other standard for identifying a “cause in fact.” Feeling constrained by the “ordinary course of business” lines of authority, it appears that the Adlman court simply saw no occasion to consider whether using some other definition of “cause in fact” might enable the court to fashion an approach to applying the Wright, Miller & Marcus standard that (1) would not" threaten to extend the work product doctrine beyond the ranges fixed by well-established traditions but that (2) would eliminate the analytical tensions visible in Adlman and (3) enable courts to be appropriately responsive to the policies that are accommodated in and served by the work product doctrine.
In some settings it will be clear that the prospect of litigation was a “substantial factor” in bringing about the existence of a document (especially when the document consists of legal analysis of that anticipated litigation) even when it is clear that another purpose also would have been sufficient to cause the party to prepare the document. In this setting, if it appeared that the other sufficient purpose would have arisen in the ordinary course of business, the dictum in Adlman would leave a court feeling that it was constrained, on this basis alone, to conclude that the document did not fall within the reach of Rule 26(b)(3). Using the “but for” test in this manner would preclude courts from reaching a second stage of analysis, a stage in which courts would consider how an order compelling disclosure of the document would affect work product policy. In sharp contrast, using the “substantial factor” formulation of the “cause in fact” test would permit courts, even when the document would have been prepared in the ordinary course of business, to proceed to the second stage of the “anticipation of litigation” analysis and to determine in that stage whether the pertinent policies dictate a conclusion that the document in question should be deemed to fall within the reach of Rule 26(b)(3).
These considerations support our view that in the first stage of the analysis for determining whether a document falls within the reach of this Rule courts should use the “substantial factor” test, rather than the “but for” test, to determine whether the prospect of litigation was a “cause in fact” of the preparation of the document — and thus to determine whether it is appropriate to proceed to the second and final stage of the analysis for determining whether “the document can fairly be said to have been prepared or obtained because of the prospect of litigation.” Wright, Miller, and Marcus, supra, § 2024, at 343. Using the “substantial factor” standard in the first stage of the test would leave courts free not to deny protection (without considering any work product policy) to documents that disclose the most sensitive litigation analyses merely because the primary, or most immediate, or ultimate purpose of preparing the documents was something other than anticipation of litigation.
At the same time, using the “substantial factor” formulation for the first stage of the test would not risk an unwarranted extension of work product protection. This follows for two reasons. First, a party seeking protection would be required, always, to prove that anticipation of litigation was at least a substantial factor in causing the document to be prepared. Second, as demonstrated below, a court can use the second stage of the analysis (the stage in which the inquiry is more in the nature of “legal cause” or “proximate cause”) to deny protection to a document that satisfies the first-stage test when the court concludes that disclosing the document would not cause significant harm to the policy objectives that work product doctrine is designed to serve.
*348 VII. APPLYING THE TWO STAGE TEST TO THE DRAFT DELOITTE REPORT
Having attempted to refine our understanding of Adlman and to give policy-sensitive meaning to the Wright, Miller & Marcus formulation of the test for determining whether a document has been “prepared in anticipation of litigation,” we turn to the document in issue in the case at bar. To determine whether JumpSport has shown the draft Deloitte report was “prepared in anticipation of litigation” we examine both the nature of the document and our specific factual situation, in light of the policies that inform work product doctrine, to determine whether this document can fairly be said to have been prepared because of the prospect of litigation.
Our analysis will include two stages. In the first, we will consider whether the prospect of litigation was a substantial factor in the mix of considerations, purposes or forces that led to the preparation of the document. In the second stage, which we will reach only if JumpSport persuades us to answer the first inquiry in the affirmative, we will determine how much harm would be done to the policies that work product doctrine is intended to serve if we were to conclude that the draft report is entitled to no protection (or how much those policy objectives would be advanced if the document were found to fall within the reach of the rule). A finding that denying any protection to the document would risk substantial harm to objectives that the work product doctrine has been crafted to promote would support a conclusion that JumpSport has shown that it could “fairly be said” that the Deloitte report was prepared “because of the prospect of litigation.”
We begin our application of the two stage test by examining the nature of the draft Deloitte report and the specific factual circumstances in which it was prepared.
A. The Nature of the Document
The draft report was not written by lawyers, but by two men with educations in business administration. JumpSport Report, Appendix, Qualifications of Principal Consultants. The “Terms & Conditions” (imposed by Deloitte) under which the report was prepared announce that Deloitte “shall have no responsibility for any assumptions provided by the Client” and “no responsibility to address any legal matters or questions of law.” JumpSport Report, Appendix, Terms and Conditions, 1(c). In the cover letter transmitting the draft report to JumpSport its authors articulate their understanding “that management will use the results of our analysis for internal financial reporting and planning purposes in connection with its business and risk management activities. No other use of this document is intended or should be inferred.” January 30, 2001, letter to JumpSport CEO and President Mark Publi-cover accompanying JumpSport Report (“JumpSport Report cover letter”) (emphasis added.) This understanding of purpose is repeated in the first section of the report. JumpSport Report, Section I.
The core product that the report provides consists of two estimates of fair market value: one of “Invested Capital” and the other of “certain Patents.” Thus, much of the report consists of clearly discoverable financial information. The section devoted to anticipated future growth includes a conclusory (unsupported and unexplained) assertion, expressly attributed to JumpSport’s management, that “other competitive enclosures currently sold in the United States allegedly infringe JumpSport’s patented technology” and that the “Company is positioned to capture a significant market share when management begins to enforce its patent rights.” JumpSport Report, IV. Company Overview, at 8. This section also repeats assertions by management (again unexplained and unsupported) that the company’s patent rights inspired (1) a decision by Sears to make JumpSport its sole supplier of trampolines and their enclosures and (2) a request by Jumpking in the late summer of 2000 to discuss “a potential alliance between the two companies.” JumpSport Report, IV. Company Overview, at 8.
The draft report also contains a description of the one patent that had been issued to JumpSport before the draft was written — but that description simply parrots verbiage from *349the published patent itself. It includes no independent analysis and discloses no legal reasoning. A separate paragraph that follows the description of the patent briefly "identifies “two possibilities to design around the patent” that management passed along to the Deloitte authors. There is no reasoning in this short paragraph about why the identified designs would not infringe, and the only reasons given for the assertion that neither design would threaten JumpSport are practical (not legal). JumpSport Report, IV. Company Overview, at 10.
In the section of its report entitled “Valuation Analysis,” Deloitte opines that the company’s most valuable assets are its intellectual property rights. Proceeding from the untested assumption (provided by management) that JumpSport would have a 100 percent market share for the next ten years, the authors examine publicly available data about royalty rates secured by other companies in other business settings to identify a range of royalty rates that JumpSport might command. JumpSport Report, VI. Valuation Analysis, at 21. In a separate section entitled “Other Valuation Issues” the Deloitte authors disclose that the bottom lines of their valuation estimates include figures for the potential value of damage settlements. These estimates of settlement value also were provided by management of JumpS-port — and no basis for them is articulated. JumpSport Report, VII. Valuation Synthesis and Conclusion, at 23. The report also discloses, with no elaboration, that JumpSport’s management anticipates settling with Jumpking in 2002 and with its two other competitors, Hedstrom and Variflex, in 2003. JumpSport Report, Exhibit VIII.
The only opinions about legal matters that are reflected in the report are completely conclusory assertions by JumpSport management14 that competitors’ products infringe JumpSport’s patents, that those competitors will settle for substantial sums when sued, and about what might constitute successful approaches to designing around the patents. The document discloses no legal analysis or reasoning by anyone. The few assertions that have legal predicates are not explained at all. There is no effort to assess the validity of the patents and no comparison of the patent claims with competitors’ products. Nor is there any discussion of pros and eons of possible lawsuits, or any analysis of litigation that might ensue. No legal strategies or litigation tactics are discussed, no approaches or arguments are outlined, and no potential evidence or sources of evidence are disclosed.
B. The Factual Situation
What do we find when we examine “the factual situation in the particular case”? In 1999 and early 2000, before the ’845 patent issued in late April of 2000, JumpSport and Jumpking discussed the possibility of forming some kind of business alliance. These discussions were unproductive — so JumpS-port realized that it might well be forced to initiate a lawsuit. Supplemental Moulton Declaration, ¶will settle for substantial sums when sued, and about what might constitute successful approaches to designing around the patents. The document discloses no legal analysis or reasoning by anyone. The few assertions that have legal predicates are not explained at all. There is no effort to assess the validity of the patents and no comparison of the patent claims with competitors’ products. Nor is there any discussion of pros and eons of possible lawsuits, or any analysis of litigation that might ensue. No legal strategies or litigation tactics are discussed, no approaches or arguments are outlined, and no potential evidence or sources of evidence are disclosed. will settle for substantial sums when sued, and about what might constitute successful approaches to designing around the patents. The document discloses no legal analysis or reasoning by anyone. The few assertions that have legal predicates are not explained at all. There is no effort to assess the validity of the patents and no comparison of the patent claims with competitors’ products. Nor is there any discussion of pros and eons of possible lawsuits, or any analysis of litigation that might ensue. No legal strategies or litigation tactics are discussed, no approaches or arguments are outlined, and no potential evidence or sources of evidence are disclosed. ¶¶ 8,10.
JumpSport also realized that prosecuting patent litigation could be very expensive — so it approached a broker, Litigation Risk Management, Inc. (“LRM”), to explore the possibility of acquiring insurance that would fund the attorneys’ fees and other costs that such litigation would entail (so-called “patent enforcement” or “patent pursuit” insurance). In the late summer or fall of 2000 LRM advised JumpSport that it would need a “Known Occurrence Patent Enforcement Policy” because it already had identified three known infringers (the three companies that JumpSport eventually named in the complaint that launched the instant litigation).15 LRM also advised JumpSport that the application process would involve several *350elements. One was a valuation of the company and of the patents that would be the subject of litigation. According to JumpS-port, LRM insisted that Deloitte be retained for this purpose. Supplemental Moulton Declaration, ¶ 13.16 So JumpSport agreed to retain Deloitte for this purpose — and that agreement led to the preparation of the draft report.
The insurance application also included two other elements: a cost and risk assessment or summary by LRM, and an “evaluation of the patent portfolio by independent patent counsel selected by Litigation Risk Management.” Supplemental Moulton Declaration, ¶ 13. In the letter that set forth the terms of its engagement, LRM stated: “We recognize that the LRM Risk Assessment is performed at the request and direction of counsel to facilitate the legal opinions and litigation strategies developed on behalf of JumpSport in anticipation of litigation.” Supplemental Moulton Declaration, ¶ 15.
This assertion, apparently written to support a later claim to work product protection, is, of course, not binding on the court. It is additional evidence, however, that at the time JumpSport authorized the preparation of the Deloitte report it understood that there was a substantial likelihood that litigation between it and Jumpking would ensue. In fact, the decision to seek this particular kind of insurance, a “Known Occurrence Patent Enforcement Policy,” constitutes substantial evidence that JumpSport anticipated a very real possibility that this litigation would be necessary. So even though the company continued to hope that it could strike a deal with its competitors, and even though it eventually decided, on cost-benefit grounds, not to acquire the policy, JumpSport clearly had “the prospect of litigation in mind when it directed the preparation” of the Deloitte report. Adlman, 134 F.3d at 1204.
In March of 2001, LRM submitted to JumpSport a draft of the independent patent validity analysis by attorneys at Baker & Botts who had been retained for this purpose, a draft of a report by LRM itself, and a second draft of the valuation report by Deloitte. Supplemental Moulton Declaration, ¶ 22. This second Deloitte report included new material reflecting “information obtained from discussions with independent patent counsel” and a new assertion that the purpose of the report was to provide “litigation risk assessment information to counsel and LRM relative to underwriting recommendations to be made by LRM____” Id.
None of these other draft reports are in issue here. The only document that is in issue is the first Deloitte draft report — initially submitted to JumpSport at the end of January of 2001. It is only that report that was inadvertently disclosed during JumpS-port’s document production. While we have no occasion to determine whether any of these other draft documents could qualify for protection as work product, as comments below will suggest, we would expect our analysis of the work product issues to take a different turn if we were considering the Baker & Botts legal analysis.
It also is instructive to note that a carrier that had issued an insurance policy of this kind apparently would decline to pay anything in response to a claim until it received yet another independent set of legal analyses from an independent counsel. When a carrier received a claim under this kind of policy it could insist, before funding its contractual share of the litigation costs, that it be provided an opinion letter from an independent and appropriately qualified intellectual property lawyer that concluded (with supporting analysis) that the accused products in fact infringed the insured’s patents.17
The fact that industry practices contemplate so much direct input from qualified patent lawyers, both before and after issuance of. a policy like this, reinforces the notion, suggested by the content of the draft Deloitte report, that no participant in the application process looks to a report like the one Deloitte prepared as an independent *351source of any significant or reliable legal analysis or litigation strategy. The valuation report might pass along or accept as assumptions legal conclusions or legal reasoning that originated in independent patent counsel-— but its authors (non-lawyers) would not be expected to engage in legal analysis on their own. We thus feel confident that the assumptions about infringement and settlement that surface in conclusory form in the first draft of the Deloitte report are not the product of any legal analysis by the authors of the report. That finding, coupled with the nakedness of the statements in the report that seem to reflect a legal conclusion by someone, encourage us to infer that all the participants in this insurance application process understood that the kind of general valuation of the business and its patent assets that we see in the draft Deloitte report was intended to serve purposes appreciably farther removed from litigation than the analyses by independent patent counsel — and thus was appreciably farther removed from the center of the work product universe.
Having now examined the nature of the draft report and the factual situation in which Deloitte produced it, it is time to apply the test for determining whether “the document can fairly be said to have been prepared or obtained because of the prospect of litigation.” Wright, Miller, and Marcus, supra, § 2024, at 343.
C. Application of the test
(i) The First Stage of the Test: Was the Prospect of Litigation a Substantial Factor Leading to the Preparation of the Deloitte Report
We first consider whether JumpSport has shown that the prospect of litigation was a “substantial factor” in the mix of considerations and forces that led to the preparation of the report. We think JumpSport has made this showing. As we pointed out, by the time JumpSport decided to commission the report the prospect of litigation was in no sense remote or speculative. While at that point it was not inevitable that litigation would ensue, the likelihood that the parties would find themselves in the court system was quite substantial — especially given the failure, over a substantial period, of the talks that had been aimed at trying to avoid litigation by forming some kind of business alliance with Jumpking. So the odds that litigation would commence were high — and the time was near.
But that is not where our consideration of this question should stop. We also must take into account the nature and magnitude of the role that the prospect of litigation played in causing this report to be generated. One question we ask in this setting is whether there were other, non-litigation considerations or motives at work in the decision to commission the report — and, if so, how substantial they were. The evidence before us supports a finding that the most immediate reason for commissioning this report was to position the company to try to secure patent enforcement insurance coverage. At least in theory, however, a report from a major accounting firm that assesses the value of the company’s invested capital and its intellectual property could be put to many different kinds of business purposes. Moreover, many of those same purposes could be advanced by a company having the kind of insurance that JumpSport was positioning itself to acquire. For example, having this kind of insurance could improve the company’s ability to raise capital. It also might lubricate efforts to merge with or acquire another entity. Yet another purpose could be to increase the price the company might bring if it were offered for sale. Such insurance also might improve a company’s chances of going public successfully, or of securing favorable terms from licensees.18
The only evidence before the court that speaks to any of these possibilities is a statement in the Supplemental Declaration of Steven W. Moulton, Vice President and General Counsel of JumpSport. After describing the role that the report was to play in applying for the patent enforcement insurance policy, Mr. Moulton avers that “JumpSport intended to engage Deloitte to provide a separate valuation report that JumpSport could use for the business purpose of obtaining potential *352investors to help fund the patent enforcement insurance policy, and the report would be based upon the same information Deloitte had obtained during its analysis for the patent enforcement litigation. Ultimately JumpSport and Deloitte did not agree on the scope and use of any such report,19 and JumpSport did not obtain a report for this purpose.” Supplemental Moulton Declaration, ¶ 16.
While this evidence suggests that JumpS-port contemplated the possibility of using the Deloitte report for more than one purpose, and while it is possible that JumpSport also understood that acquiring patent enforcement insurance might redound to its business benefit even if no litigation ensued, there is no evidentiary basis for a finding that any of these kinds of considerations was anything but peripheral. In other words, the record does not support a finding that any consideration or purpose other than positioning the company to acquire the insurance coverage would have been sufficient to cause JumpS-port to commission the report that is in issue here. Nor does the record support a finding that JumpSport’s interest in acquiring this kind of insurance was inspired to any appreciable extent by any motivation other than funding patent enforcement litigation.
In sum, we conclude, on the record before us, that the prospect of litigation to enforce its patent rights was the only significant factor underlying JumpSport’s pursuit of the insurance — and that trying to meet the informational requirements of the prospective insurer was at least the primary factor that led JumpSport to commission the Deloitte report. In these circumstances, a conclusion that the prospect of litigation was “a substantial factor in the mix of considerations” that led to the preparation of the document is not foreclosed either by the fact that the connection between the report and the litigation was not immediate and direct (the report would support the acquisition of the insurance, then the insurance would support the litigation), or by the fact that one could characterize the interest that most directly inspired the company’s pursuit of the coverage as a “business” or “financial” interest. Much of litigation, after all, is business — or about business. Stated differently, a “factor” can be “substantial” (as that term is used in this setting) even if more than one label might be attached to it (business and litigation), even if the causal connection between it and the prospect of litigation is in some measure indirect, and even if more than one interest motivated the preparation of the document.
Similarly, the fact that JumpSport ultimately decided not to purchase the insurance does not foreclose a finding that the prospect of litigation was a substantial factor leading to the preparation of the report. The time on which courts primarily should focus when addressing this issue is the time the party was making the decision to have the document prepared. It is the party’s prospective mind set at that juncture that matters — for we are trying to determine what role the “prospect” of litigation played in the preparation of the document. Had JumpSport abandoned pursuit of the insurance because it abandoned the idea of initiating litigation we would have no occasion to struggle with this matter — but that is not the reason JumpS-port elected not to buy the insurance. Moreover, the litigation that JumpSport had in mind when it commissioned the report in fact occurred.
Our conclusion that the prospect of litigation was not only a substantial factor, but the dominant consideration driving the decision to commission the Deloitte report is buttressed by the fact that no one has suggested that in the normal course of business, or for some other reason independent of the prospect of litigation, JumpSport would have had a report of this character prepared during this period. While not essential to a conclusion that the prospect of litigation was a substantial factor, this kind of finding is likely to be sufficient to support that conclusion when it is clear that the prospect of litigation was real and relevant to the decision to have the document prepared.
*353Given all these considerations, we hold that JumpSport has satisfied the first prong of the test for determining whether a document can fall within the reach of Rule 26(b)(3).
(ii) The Second Stage of the Test: How Much Harm Would be Done to the Policy Objectives of the Work Product Doctrine if Protection is Not Available to the Deloitte Report?
We cannot conclude that the Deloitte report was “prepared in anticipation of litigation” unless JumpSport also satisfies the second prong of the test (something partially akin to the “proximate cause” or “legal cause” inquiry in the torts field). In the second stage of the test we focus on the policy objectives that the work product doctrine is intended to serve — then we determine how much (if any) harm they would suffer if the document was not protected (or how much those objectives would be advanced by a determination that the document falls within the reach of the Rule). In addressing these issues we again consider the nature of the document and the particular factual setting in which it was created.
In this case, it is the nature of the document that undermines most the contention that it should be deemed to fall within the ambit of the work product rule. As noted above, the Deloitte report contains no legal analysis, no discussion of legal strategies, no observations at all about the substantive or procedural matters that litigation might entail. Its authors expressly disclaim any “responsibility to address any legal matters or questions of law.” They also expressly articulate, twice, their understanding of the limited purposes the document was intended to serve: “management will use the results of our analysis for internal financial reporting and planning purposes in connection with its business and risk management activities. No other use of this document is intended or should be inferred.” JumpSport Report cover letter (emphasis added).
The primary objective of the report is to ascribe dollar value to JumpSport’s invested capital and its intellectual property (which consisted, at the time, of one issued patent and some related pending applications). The vast majority of the verbiage in the document addresses business and financial matters — virtually all discoverable. There are instances in which the authors make assumptions (clearly ascribed to JumpSport’s management) that have law-based predicates, e.g., that the patents are enforceable, and that competitors will recognize this fact and pay substantial sums to settle lawsuits that JumpSport might bring to enforce its rights. But these assumptions are neither explained nor defended. They are simply made.
The report’s brief description of the one issued patent is taken directly from the language of the patent itself — a public document. That description includes no analysis of claims, no discussion of prior art or other bases for challenging validity, and no comparison of the claims with the products that might be accused of infringing. One paragraph identifies two “design around” possibilities, both suggested by JumpSport’s management, but there is no discussion of why, under the law or the arguable scope of the claims, the alternative designs would not infringe. JumpSport Report, IV. Company Overview, at 10. Nor do the authors identify or discuss evidence, or sources of evidence that might come into play in litigation. In fact, they say nothing about what might be involved in future litigation or how it might be prosecuted or defended. They simply assume, for purposes of valuing the company’s assets, that JumpSport would be victorious in any such litigation. These superficial assumptions could not possibly reveal anything of real litigation value to any prospective party opponent.
Given the content of the report, would a refusal to extend protection to this document invade that “zone of privacy” that courts have assumed that lawyers and clients need if they are to think reliably, critically, and creatively about their litigation? Because the document reveals no legal reasoning and no thoughts about what might occur in litigation or how litigation might be handled, the answer must be no.
Alternatively, would a refusal to extend protection to this document damage the sets of incentives that are said to be essential to the effective functioning of the adversary system? Answering this question is a more *354complicated undertaking because the Deloitte report includes some information that might be relevant to litigating damages issues in an enforcement action. The report contains information (and some opinions) about the income, assets, and value of the company (as of late 2000), as well as about the value of its patent and what level of royalty rates the patent might support at that-time.
Several observations about this information are in order. First, as noted above, virtually (perhaps literally) all of the data on which the report was based would be discoverable (in fact, according to plaintiffs counsel, this information already has been discovered in this case by defendant’s counsel). Declaration of Eileen Taglang In Support of Plaintiff JumpSport, Inc.’s Reply Brief In Support of Motion to Compel Return of Inadvertently Produced Document (“Taglang Declaration”), ¶¶ 5-11. And there is nothing in the way the Deloitte authors (non-lawyers) organized or packaged this material, or in the topics they decided to address and the categories of information they decided to describe, that would tend to reveal anything about litigation strategies or theories that JumpSport’s lawyers might later decide to adopt. Because a party could foresee that at least the vast majority of the underlying data would be discoverable, and would know that a document like this (that was prepared by non-lawyers) would not disclose litigation strategies, the fact that the draft report might not be protected by the work product doctrine would not serve as a consequential deterrent to commissioning its preparation.
Second, by the time this case goes to trial, much of the information in the report (which was gathered in the latter half of 2000) will be at least somewhat dated. While some of the information in the report might remain relevant to some parts of the damages questions, the primary focus of damages litigation likely would be on information covering the intervening period. These facts also would be foreseeable at the time JumpSport decided to commission the report — and would help reduce the likelihood that JumpSport would be inhibited from having the report prepared by the possibility that it might in the future get into the hands of an adversary in litigation.
We acknowledge that Jumpking’s damages expert has submitted a declaration in which he asserts that he relied to some extent on information in the Deloitte draft when he developed his rebuttal report. In some circumstances, how an adversary later uses a document in the litigation, or how he proposes to use it, could be relevant to the issues we address here — if that use was foreseeable at the time the decision was originally made to direct the preparation of the document. If, for example, it was foreseeable that a document of the kind in issue could be used tellingly by an adversary to gain litigation leverage on an important disputed issue, and if it was unlikely that the adversary could gain that same leverage through evidence, data, or analyses with other (accessible) sources, then a decision that the document could receive no protection under Rule 26(b)(3) might well inhibit a party from deciding to have such a document prepared in the first place.
For reasons explained below, however, we are not persuaded that any part of the De-loitte draft is likely to serve as a unique source of significant leverage on any issue of real consequence in this case. We also note that Jumpking’s damages expert claims to have cited the draft only twice in a rebuttal report that is at least 26 pages long. And while part of the report would shed some light on some of the information that JumpS-port’s management had in late 2000 about arguably relevant royalty rates, there is no reason to believe management did not simultaneously have additional information in the same subject area.
Moreover, the public sources from which the Deloitte authors drew the information they used to estimate “comparable” royalty rates were and are equally available to defendant and its expert. Taglang Declaration, K11. Similarly, defendant has access (through documents, its own witnesses, and through depositions of others) to at least most of the limited information in the draft about negotiations between Jumpking and JumpSport and about how a merger of the companies might be expected to affect prod*355uct distribution expenses. It also is significant that while these matters are not wholly peripheral to the litigation, they hardly can be characterized as pivotal — and it is unlikely that the sparse information about them in the Deloitte draft could determine their outcome. For these reasons, as well as the considerations detailed in the paragraphs that follow, it is not at all clear that a party who was trying to decide whether to have a document like the Deloitte draft prepared would either foresee the uses that Jumpk-ing’s expert says he is making of the document or, if she foresaw those uses, would be deterred thereby from commissioning the work.
We turn here to a broader consideration of the quality and depth of the draft report— and thus to its potential probative utility in this litigation. The authors admittedly relied extensively on data already compiled by the company and on assumptions that originated in its management. At least for the most part, Deloitte did not challenge or investigate those assumptions. And it seems clear that Deloitte’s effort to gather and analyze financial data from the company was not as careful, probing, or complete as it would have been had Deloitte been preparing a report for a less limited or more public purpose, e.g., for submission to a regulatory body or in connection "with a public offering.
These inferences, suggested by the content of the document, are reinforced by the vigorous disclaimers that accompany it. As we mentioned above, the second paragraph of the letter delivering this draft to JumpS-port, and a preliminary section of the body of the report, emphasize the authors’ understanding that “management will use the results of our analysis for internal financial reporting and planning purposes in connection with its business and risk management activities. No other use of this document is intended or should be inferred.” JumpSport Report cover letter; JumpSport Report, I. Engagement Overview, at 1. The cover letter further explicitly sets forth the report’s limitations by emphasizing that in “conjunction with this assignment, we were provided with compiled financial statements, internal operating statements, and detailed three year forecasts. We have accepted this data as fairly reflecting the results of business operations. We have not audited this data as part of our valuation analysis and, therefore, we express no opinion or other form of assurance regarding their accuracy or fairness of presentation.” JumpSport Report cover letter (emphasis added.) Apparently unsatisfied that these assertive disclaimers were sufficient to protect them, the Deloitte authors went on to insist that JumpSport not distribute the report to anyone but its intended users (presumably limited to LRM and Lloyds of London) and not, without De-loitte’s written consent in advance, “reference our name or report, in whole or in part, in any document distributed to third parties unrelated to this assignment.” JumpSport Report cover letter.
Similarly restrictive notions are set forth in the “General Business Terms and Conditions” under which Deloitte insisted on carrying out this assignment. In those Conditions, Deloitte disclaimed, again, any responsibility for data or assumptions provided by the client (JumpSport), as well as any “responsibility to address any legal matters or questions of law.” The Conditions also declared that Deloitte would “not issue any opinion, report or other form of assurance with respect to any financial information in connection with its services hereunder.” JumpSport Report, Appendix, Terms & Conditions, 1(c).
One thing made clear by all these disclaimers is that Deloitte wanted no one to believe that this report would satisfy the substantially more demanding standards that would have to be met for reports that were to be submitted to regulatory agencies or that were intended to satisfy securities laws.
From the content of the report, from the disclaimers and express limitations that accompanied it, as well as from the fact that the broker, LRM, seems to have regularly used Deloitte, and only Deloitte, for this kind of assignment, we infer that when it prepared this draft report Deloitte expected JumpSport to use it for only one purpose: to support an application for patent enforcement insurance. We further infer that it was that very limited anticipated use of the re*356port that permitted Deloitte, professionally, to prepare a document whose underlying data it had neither gathered nor verified and whose critical assumptions it had not tested or analyzed.
Because of the infirmities in the report’s underpinnings, and the apparent lack of depth in its analyses, it is unlikely that a party would rely in litigation on this kind of document to provide substantial support for a damages claim. More importantly, for present purposes, it is not likely that when a party was deciding to commission a report like this (and worrying about keeping its cost as low as possible) the party would anticipate being able to put it to telling use in future litigation.
Nor is it likely that JumpSport’s principals expected the preparation of this report to play any significant role in their decisions about whether to initiate litigation, or how much money to invest in enforcement actions. We draw this inference in part because it appears that the leaders of JumpS-port had decided, months before commissioning the report, that, in order for JumpSport to survive, the company would have to sue Jumpking if that competitor would not agree to an acceptable business solution.20 This inference also is supported by the fact that the principals of JumpSport were the source of most of the information and opinions in the report that would bear on such decisions, e.g., the opinion that the patent was valid, that competing products infringed, that there were no practical design-around solutions, and that competitors would recognize these realities and either quit the field, sign licensing or joint venture agreements, or pay substantial sums to settle any lawsuits that JumpSport was forced to file. Finally, it appears that the content of the draft report had little or nothing to do with JumpSport’s decision, a few months later, not to buy the insurance. Rather, that decision turned on the “prohibitively high” cost of the insurance (in premiums and anticipated co-payments). Supplemental Moulton Declaration, ¶ 23.
The report’s limitations also reduce, considerably and foreseeably, its potential value to a litigation adversary. JumpSport’s lawyers would know that disposition of the major issues in any anticipated litigation would turn on other, more clearly reliable evidence — and that Rule 403 might well prevent admission of the report at trial. For these reasons, it seems unlikely that a party who was considering commissioning a study such as this would be inhibited from doing so to any appreciable extent by apprehension that someday it might be acquired through discovery by an opponent.
These same considerations suggest that knowing that she could have access to a document like this would not likely reduce an opposing attorney’s motivation to conduct energized and independent investigations of the matters addressed in the report, or to develop independent analyses (for litigation-related uses) of the issues to which some parts of the report might be relevant. Given the report’s limitations, and its authors’ disclaimers, an opponent who was permitted to “borrow” this document (through discovery) would be borrowing precious few of his adversary’s “wits.” Given that fact, it is difficult to see how refusing to extend work product protection to this kind of report would tend to “demoralize” or “de-energize” to any appreciable extent either a party who might need such a report or a party opponent. In other words, it is difficult to see how a conclusion that this kind of report does not fall within the reach of Rule 26(b)(3) could do any real harm to the sets of incentives on which the adversary system relies.
Having considered the content of the document, we turn to examine “the factual situation in [this] particular case”21 to see if anything in that situation would indicate that exposing the Deloitte draft to discovery would damage (more than minimally) any of the several policy objectives that inform the work product doctrine. We note first that we have been able to find nothing in the factual circumstances surrounding the decision to commission this report that triggers concern that permitting Jumpking to retain *357the Deloitte draft might threaten the vitality of the incentives that are said to undergird the adversary system. As indicated, the report was prepared in order to meet the requirements of an insurance application. And while the purpose of the insurance in question would have been to help fund litigation, JumpSport eventually decided not to buy the insurance — but to proceed with the litigation anyway. Given that fact, we clearly cannot say that acquiring the insurance was an essential prerequisite to proceeding with the litigation.
Nor have we been persuaded, at a more generic level, that there is a significant risk that an entity that wanted to acquire patent prosecution insurance would be deterred from seeking coverage because of the possibility that this kind of report might be discoverable. To repeat, virtually all of the underlying data on which the draft is based is accessible, foreseeably, through discovery. So the commissioning of a report such as this would create little risk that an adversary would gain access to evidence, or learn important facts, which it would not otherwise be able to obtain. Moreover, because this type of report is prepared for a very limited and specific purpose, and because parties always will want to keep the cost of this kind of document as low as possible, it will be foreseeable, at least in most instances, that the people charged with preparing the report will not conduct thorough investigations, will not gather and verify independently all the pertinent data, will not commit huge resources to developing sophisticated analyses, and, generally, will not attempt to form or express legal opinions.
These facts, in turn, will make it foreseeable that the quality and character of the document will impose significant limits on how tellingly it might be used in litigation. And while a document like this would be of interest to an adverse party as it decided how to approach settlement negotiations, even in that arena its role, realistically, is not likely to be substantial. For these reasons, it appears unlikely that a party who wanted to acquire patent litigation insurance would be deterred from proceeding with an application out of fear that this kind of report might someday fall into the hands of a litigation adversary. So, even if we assume that work product policies would be implicated by court rulings that deterred parties from pursuing litigation insurance, and thus, indirectly, from litigating, we are not persuaded, given the facts of this particular case, that permitting Jumpking to keep the draft Deloitte report would tend to inhibit, to any significant extent, a party in JumpSport’s situation from seeking litigation insurance.
Accordingly, we find that refusal to extend protection to this document does not damage the sets of incentives essential to the effective functioning of the adversary system. Because we find that disclosure of the De-loitte Report does not threaten the parties’ “zone of privacy,” or enable one side to borrow its adversaries wits to any significant degree, we hold that JumpSport has not satisfied the second stage of the test.
VIII. CONCLUSION
Although JumpSport satisfied the requirements of the first stage of the test by persuading us that the prospect of litigation was a “substantial factor” leading to the preparation of the document, JumpSport has failed to demonstrate that extending protection to this draft report would advance appreciably any of the policy objectives that the work product doctrine has been developed to serve. In sum, the content of the Deloitte report, and the circumstances in which it was prepared, remove it too far from the heartland of work product sensitivity to justify ruling that it warrants protection under this doctrine. It follows that we cannot conclude that “the document can fairly be said to have been prepared or obtained because of the prospect of litigation.”
Because we conclude that the draft De-loitte report does not fall within the ambit of Rule 26(b)(3) we DENY JumpSport’s motion for an order compelling Jumpking to return the document. We need not address the other grounds that Jumpking advances for reaching the same conclusion.
IT IS SO ORDERED.
. Solely for purposes of resolving the instant discovery dispute we assume (but do not decide) that the facts set forth in the cited portions of the declarations submitted by JumpSport are true.
. No business relationship was ever formed. Supplemental Moulton Declaration, ¶ 8.
. Deloitte submitted a second draft of its report in March of 2001. This draft, unlike the first iteration of this work, apparently reflected legal analysis by the independent patent lawyer. Supplemental Moulton Declaration, ¶ 22. This second draft is not in issue in this motion.
. Of course, some parts of a document might constitute "opinion” work product while other parts of the same document might constitute “non-opinion” work product. And some parts might not constitute work product at all.
. Charles Alan Wright, Arthur R. Miller, and Richard L. Marcus, 8 Federal Practice & Procedure, § 2024 (1994).
. We explore in a subsequent section what this phrase might mean and the analytical pitfalls it may conceal.
. In Chevron Texaco Corp., we intimated no discomfort with the general notion that Rule 26(b)(3) would not cover a document that was prepared in the ordinary course of business. We had no occasion in that litigation to undertake the kind of inquiry into the difficulties this notion can spawn that we have made in the case at bar.
. JumpSport also seems to urge us to collapse our inquiry into one question by using a "but for” definition of "because of.” To provide evidentiary support for its position, JumpSport’s general counsel declares that he directed Deloitte to prepare the draft report "only because of the anticipated litigation against defendants, to assist JumpSport in malting business decisions relating to the prospective litigation. The Deloitte report would not have been prepared in substantially similar form but for the prospect of patent infringement litigation against defendants." Moulton Declaration, ¶ 4.
All this may be true enough, at least at some level of abstraction. But should it follow that the draft report is entitled to be deemed "work product” — even though that report contains nothing even remotely like the “detailed legal analysis” that was in issue in Adlman, and even though it is not at all clear that disclosing the report would harm the policy objectives that the work product doctrine is designed to promote? Because we prefer not to be trapped into a corner that has the patina of logical tidiness but that leaves no room to do justice to the purposes of work product doctrine, we feel compelled to examine Adl-man much more closely.
. Other than the policies that the work product doctrine has been fashioned to promote, it is not at all clear what concepts, values, or policies should be used to give meaning to the word "fairly” in this setting. We fear that the test for determining whether a document was “prepared in anticipation of litigation” would become hopelessly muddled and rootless if courts were to inject into it, via the phrase "fairly be said,” an effort to determine, by values unconnected with work product policy, whether it would be “fair” to extend Rule 26(b)(3)’s coverage to the document.
There is a place in work product analysis where courts apply factors that implicate some general or conventional notions of "fairness," but courts arrive at that place only if they first have decided, on grounds independent of general notions of what is “fair” under the circumstances, that a particular document was “prepared in anticipation of litigation.” The point at which Rule 26(b)(3) calls upon courts to consider matters of “fairness,” as that phrase is commonly understood, is after a court has concluded that a document was prepared in anticipation of litigation and that it falls in the category of "non-opinion” work product. At that juncture the Rule directs courts to consider whether the party trying to discover the document "has substantial need of the materials in the preparation of the party’s case and that the party is unable without undue hardship to obtain the substantial equivalent of the materials by other means.”
These requirements are imposed, at least in part, to help a court determine whether it would be "unfair” to continue to shield (or to order disclosure of) the particular work product material in issue. In this context, courts consider (among other things) whether shielding the material would give one party an informationally unfair advantage over the other, i.e., would deprive a party of an opportunity to have the important issues decided fairly (meaning on the basis of all the available evidence that has real probative significance.)
. For a very clear and instructive discussion of the relationship in the development of tort law doctrine between "cause in fact” and "legal cause” ("proximate cause”), see W. Keeton, D. Dobbs, R. Keeton & D. Owen, Prosser and Kee-ton on Law of Torts, § 41 (5th ed.1984). I have relied extensively on this source in developing my understanding of the meanings (in torts) of "cause in fact" and "legal cause” (a.k.a. "proximate cause").
. We do not mean to imply that there is no connection at all, or that, there would never be any overlap, between the inquiry that would be made in the first stage and the inquiries that would be made in the second. In the torts arena, for example, there are cases that look in part to the physical and temporal proximity between the act on which liability would be predicated and the harm that occurred to help decide whether the former was the "proximate cause” of the latter. See, e.g., Mark F. Grady, Proximate Cause Decoded, 50 UCLA L.Rev. 293 (Dec.2002) (comparing cases that use the “reasonable fore-seeablity” and the "direct consequences” formulations of the standard for identifying "proximate cause” and attempting to resolve apparent doctrinal tensions between these camps). While the conceptual basis for doing so in the work product arena may he less clear, we would not be surprised if some courts take similar kinds of proximity considerations into account when trying to decide whether, given all the factual circumstances, it would be fair to say that a document was created "because of the prospect of litigation.”
What we do intend to suggest, however, is that the "cause in fact” and the "legal cause” inquiries are not coterminous — either in tort theory or, as adapted and re-cast here, in the work product setting. Satisfying the requirements of one by no means guarantees satisfaction of the requirements of the other. This fact is well illustrated when, in the latter sections of this opinion, we apply the test that we have developed.
. Advisory Committee Notes are not binding legal authority — at least until incorporated into law by authoritative judicial opinions. See, e.g., United States v. Hayes, 983 F.2d 78, 82 (7th Cir.1992); see also Tome v. United States, 513 U.S. 150, 167-68, 115 S.Ct. 696, 130 L.Ed.2d 574 (1995) (Scalia, J., concurring in part and concurring in the judgment.) Many courts, however, have looked to the Notes as important sources of guidance when trying to construe and apply the rules.
. Goosman v. A. Dule Pyle, Inc., 320 F.2d 45 (4th Cir.1963) and United States v. New York Foreign Trade Zone Operators, Inc., 304 F.2d 792 (2d. Cir.1962).
. See Supplemental Moulton Declaration, ¶ 21, where he states: "To the best of my knowledge, the Deloitte Report does not contain any independent legal analysis or conclusions by Deloitte employees.”
. The Court relied on the following articles to develop a familiarity with patent litigation insurance: William N. Hulsey III, Patent Insurance Can Guard Intellectual Capital, Austin Bus. Journal, June 12, 1998; Joby A. Hughes & Kate L. Birenbaum, Insuring Intellectual Property Risks: Creative Solutions on the Cutting Edge, 568 PLI/ Pat 203; and Bruce E. Burdick, Patent Insurance, Is It Worth It? (© 2002), available at http://members.tripod.com/burdicklawfirm/insu-rart.htm.
. While the record does not make it clear that carriers would not issue this kind of insurance without such a valuation report, we assume, for the limited purpose of resolving the issues before us, that a report of this kind was a necessary prerequisite to the issuance of a policy.
. See Hulsey, supra note 15.
. Burdick, supra note 15; Hughes & Birenb-aum, supra note 15; Hulsey, supra note 15.
. The fact that JumpSport and Deloitte could not agree about "the scope and use” of a report that JumpSport would utilize for the more sensitive purpose of attracting investors heightens our concern about the informational underpinnings of the draft Deloitte report.
. See Supplemental Moulton Declaration, HH 6, 8, 9 and 10.
. Wright, Miller, and Marcus, supra note 5, § 2024 at 343.