This case requires us to determine whether the Supreme Court’s decision in Calbeck v. Travelers Insurance Co., 370 U.S. 114, 82 S.Ct. 1196, 8 L.Ed.2d 368 (1962), which clarified the jurisdictional scope of the Longshoremen’s and Harbor Workers’ Compensation Act of 1927, should be applied to workers’ injuries that antedated the decision.
*82I. Background
On November 21, 1941, Donald Simpson was working as a welder on board a vessel under construction on a marine railway at the Bath Iron Works shipyard. He fell approximately fifty feet from the deck of the vessel, landing on his back on a set of pipes on the marine railway. He was rendered a paraplegic and has been confined to a wheelchair ever since. The employer (BIW) knew of the injury but never filed a First Report of Injury under the Longshoremen’s Act. Simpson sought and received $6,000 from the employer under the provisions of the State of Maine Workers’ Compensation Act between December 18, 1941, and April 8, 1948. See Simpson’s Case, 144 Me. 162, 66 A.2d 417 (1949). On March 22, 1977, Simpson filed a claim for benefits under the federal Act. He seeks $1,500 (the difference between the $6,000 recovered under the Maine Act and the $7,500 that was authorized under the federal Act in 1941) plus reimbursement for thirty-six years’ worth of medical expenses (available under the federal but not the state Act). The claim was not barred by the Act’s statute of limitations, because of the employer’s failure to file a First Report of Injury. 33 U.S.C. §§ 913(a), 930(f).
In proceedings within the Department of Labor, the employer posed a series of jurisdictional challenges to Simpson’s claim. It is clear that Simpson’s injury fell within the coverage of the Act’s jurisdictional rule, 33 U.S.C. § 903(a) (Section 3(a) of the Act) as it was interpreted in Calbeck, supra.1 BIW argued that Calbeck’s interpretation should not apply retroactively to Simpson’s injury; that the claim was barred by the collateral estoppel theory of Shea v. Texas Employer’s Insurance Assoc., 383 F.2d 16 (5th Cir. 1967), as well as the doctrines of election of remedies, res judicata, and laches; and that Simpson’s injury did not meet the federal Act’s situs requirement since he fell from the vessel onto land.
The Administrative Law Judge dismissed Simpson’s claim on the basis of the employer’s second argument, holding She a, supra, controlling. The Benefits Review Board disagreed and held that Shea was inconsistent with Calbeck. Nevertheless, the Board affirmed the dismissal of Simpson’s claim. In a split decision, Administrative Appeals Judge Miller dissenting, it reasoned that Calbeck’s analysis of Section 3(a) was not to be applied retroactively.
We have jurisdiction to review the Board’s decision, pursuant to 33 U.S.C. § 921(c). The Director of the Office of Workers’ Compensation Programs has, as a party-respondent, see Shahady v. Atlas Tile & Marble Co., 673 F.2d 479 (D.C.Cir.1982), filed a brief in support of the petitioner, urging reversal of the Board.
II. The Retroactivity of Calbeck
A. The Calbeck Holding
Calbeck v. Travelers Insurance Co., 370 U.S. 114, 82 S.Ct. 1196, 8 L.Ed.2d 368 (1962), considered two cases, each arising out of facts indistinguishable from those before us here. In each case, a worker had been injured while laboring aboard a new vessel that was still under construction, but already afloat on navigable waters. The insurance company in Calbeck had prevailed before the Fifth Circuit Court of Appeals, Travelers Ins. Co. v. Calbeck, 293 F.2d 52 (5th Cir. 1961); Avondale Shipyards, Inc. v. Donovan, 293 F.2d 51 (5th Cir. 1961) on the basis of the following syllogism:
1. Section 3(a) of the Act declared that compensation was available under federal law only if “recovery ... through workmen’s compensation proceedings may not validly be provided by state law.”2 Fed*83eral and state jurisdiction were therefore mutually exclusive.
2. A line of cases starting with Grant Smith-Porter Ship Co. v. Rohde, 257 U.S. 469, 42 S.Ct. 157, 66 L.Ed. 321 (1922), had established the so-called “maritime-but-local” or “local concern” doctrine. According to this doctrine, when an employee performed tasks of “local concern”, such as constructing a new vessel on navigable waters, a state law could validly provide workers’ compensation in a maritime context.
3. Thus, the Longshoremen’s Act did not permit federal compensation to workers whose situations were covered by the maritime-but-local doctrine.
(Hereafter we will call this argument “the Rohde -3(a) syllogism”.)
Despite the neatness of the Rohde-3(a) syllogism, the Supreme Court reversed. The Court made several points in its meticulous analysis of § 3(a)’s legislative history. Earlier drafts of section 3(a) had explicitly incorporated the “local concern” doctrine. A House Committee had eliminated the language completely; a Senate Committee had responded to employer complaints about the language by replacing it with the “may not validly be provided by State law” language. The Senate language was adopted, not to incorporate “local concern” obliquely, but rather to eliminate a parliamentary obstacle on an unrelated issue (the exclusion of seamen from coverage). Calbeck, supra, 370 U.S. at 122-24 & n.13, 82 S.Ct. at 1200-02 & n.13.
In addition, the purpose of the Act, as stated in the Senate Report, S.Rep.No.973, 69th Cong., 1st Sess., at 16, was to provide certain compensation for a class of employees who had been denied the certainty offered by state compensation statutes. The Senate Report named three Supreme Court cases that had created the need for such an Act: Southern Pacific Co. v. Jensen, 244 U.S. 205, 37 S.Ct. 524, 61 L.Ed. 1086 (1917); Knickerbocker Ice Co. v. Stewart, 253 U.S. 149, 40 S.Ct. 438, 64 L.Ed. 834 (1920); and Washington v. Dawson & Co., 264 U.S. 219, 44 S.Ct. 302, 68 L.Ed. 646 (1924). (Jensen had declared unconstitutional New York’s efforts to apply its state compensation statute to an injury sustained on a gangplank between a vessel and a wharf; Stewart and Washington had struck down federal statutes attempting to authorize the states to use their own compensation statutes in such cases.) This Senatorial statement of purpose makes no mention of Rohde or the local concern doctrine. See Calbeck, supra, 370 U.S. at 126 n.14, 82 S.Ct. at 1203 n.14. Moreover, the Rohde-3(a) syllogism would undermine the stated goal. It would eliminate certainty because “the contours of the ‘local concern’ concept were and have remained necessarily vague.” Id. at 124, 82 S.Ct. at 1202. It would perpetuate the risk that some maritime employees would be covered by neither federal nor state laws, wherever a state chose not to push its coverage to the Constitutional limit. Id. at 125, 82 S.Ct. at 1202. And it would make “federal coverage expand and recede in harness with developments in constitutional interpretation as to the scope of state power, [meaning] that every litigation raising an issue of federal coverage would raise an issue of constitutional dimension, with all that that implies; and that each and every award of federal compensation would equally be a constitutionally premised denial of state competence in a like situation.” Id. at 126, 82 S.Ct. at 1203.3
B. The Benefits Review Board Decision
In its opinion below, the Benefits Review Board concluded that Calbeck’s interpretation should not be applied to all claims arising out of injuries since 1927. It anchored its analysis in the following language from Chevron Oil Co. v. Huson, 404 U.S. 97, 106-07, 92 S.Ct. 349, 355, 30 L.Ed.2d 296 (1971):
*84“In our cases dealing with the nonretroactivity question, we have generally considered three separate factors. First, the decision to be applied nonretroactively must establish a new principle of law, either by overruling clear past precedent on which litigants may have relied, see, e.g., Hanover Shoe v. United Shoe Machinery Corp., [392 U.S. 481,] 496 [88 S.Ct. 2224, 2233, 20 L.Ed.2d 1231], or by deciding an issue of first impression whose resolution was not clearly foreshadowed, see, e.g., Allen v. State Board of Elections, [393 U.S. 544,] 572 [89 S.Ct. 817, 835, 22 L.Ed.2d 1]. Second, it has been stressed that ‘we must . .. weigh the merits and demerits in each case by looking to the prior history of the rule in question, its purpose and effect, and whether retrospective operation will further or retard its operation.’ Linkletter v. Walker, [381 U.S. 618,] 629 [85 S.Ct. 1731, 1738, 14 L.Ed.2d 601], Finally, we have weighed the inequity imposed by retroactive application for, ‘[w]here a decision of this Court could produce substantial inequitable results if applied retroactively, there is ample basis in our cases for avoiding the “injustice or hardship” by a holding of nonretroactivity.’ Cipriano v. City of Houma, [395 U.S. 701,] 706 [89 S.Ct. 1897, 1900, 23 L.Ed.2d 647].”
The Board majority, impervious to the protestations of its own dissenting number, embraced the views of the Calbeck dissent to support its belief that Calbeck established a new principle of law. It then declared that Calbeck’s purposes would not be promoted by applying it to Simpson’s claim, but purported to reserve judgment on whether those purposes might be promoted by retroactive application to other, different claims. Finally, it concluded that the balance of equities favored BIW over Simpson. It held that the controversy was governed by “the law in effect during the time when claimant’s state claim was pending”— “the maritime but local rule, at that time incorporated by Section 3(a)”. Since we disagree with the manner in which the board applied Chevron Oil to this case, we must vacate its judgment.
C. Preliminary Observations on Retroac-tivity
A fundamental perspective that must inform our analysis is that the retroactive applicability of judicial decisions of federal courts is the rule, not the exception. This century4 has seen the Supreme Court establish that both state courts, see Great Northern Railway v. Sunburst Co., 287 U.S. 358, 53 S.Ct. 145, 77 L.Ed. 360 (1927), and federal courts, see Linkletter v. Walker, 381 U.S. 618, 85 S.Ct. 1731, 14 L.Ed.2d 601 (1965), may legitimately conclude that justice requires a decision to be applied non-retroactively. But, at least in the federal courts, the strong presumption has remained in favor of retroactivity. Robinson v. Neil, 409 U.S. 505, 507-08, 93 S.Ct. 876, 877-78, 35 L.Ed.2d 29 (1973); Miro Martinez v. Compania Trasatlantica Espanola, 643 F.2d 897 (1st Cir. 1981). This presumption follows in part from the fact that every federal decision applies “retroactively” to a set of facts that has already taken place.5 The presumption promotes uniform treatment of litigants, since the reasons that lead the court to make such a “retroactive” *85decision should apply with equal force to other similarly situated litigants, no matter when their cases arise. Most court-made law is interstitial, constrained by previously enacted principles of positive constitutional and statutory law, and merely clarifies points that were previously open to debate. It is applied retroactively, since that is usually the fairest, least arbitrary way to proceed. Cf. Bradley v. Richmond School Board, 416 U.S. 696, 94 S.Ct. 2006, 40 L.Ed.2d 476 (1974) (federal statute modified between trial decision and appellate decision; new version must control); Vanden-bark v. Owens-Illinois Glass Co., 311 U.S. 538, 61 S.Ct. 347, 85 L.Ed. 327 (1941) (state Supreme Court reversed prior holdings between federal trial and federal appeal; new interpretation must control).
Yet sometimes retroactivity would be excessively unsettling. At times a decision not only makes law, it effects a volteface. Even in such cases, retroactivity would still provide for the most uniform application of the new, correct legal standards, but that interest in uniformity is outweighed by the need to protect justifiable reliance on a contrary, seemingly authoritative view of the law. Thus, the Supreme Court has given only prospective effect (other than application to the case before it) to some cases that overruled prior doctrine. E.g., Daniel v. Louisiana, 420 U.S. 31, 95 S.Ct. 704, 42 L.Ed.2d 790 (1975); Mackey v. United States, 401 U.S. 667, 91 S.Ct. 1160, 28 L.Ed.2d 404 (1971); Williams v. United States, 401 U.S. 646, 91 S.Ct. 1148, 28 L.Ed.2d 388 (1971); Fuller v. Alaska, 393 U.S. 80, 89 S.Ct. 61, 21 L.Ed.2d 212 (1968); Johnson v. New Jersey, 384 U.S. 719, 86 S.Ct. 1772, 16 L.Ed.2d 882 (1966); Linklet-ter, supra. And, in the area of criminal procedure, where society through its law-enforcement agents depends daily on whatever the established rules of legitimate police conduct may be, the Court has given prospective effect to decisions creating new rules unrelated to the truth-finding process. E.g., United States v. Peltier, 422 U.S. 531, 95 S.Ct. 2313, 45 L.Ed.2d 374 (1975); Hill v. California, 401 U.S. 797, 91 S.Ct. 1106, 28 L.Ed.2d 484 (1971); Williams, supra; Jenkins v. Delaware, 395 U.S. 213, 89 S.Ct. 1677, 23 L.Ed.2d 253 (1969); Desist v. United States, 394 U.S. 244, 89 S.Ct. 1030, 22 L.Ed.2d 248 (1969); Fuller, supra; Stovall v. Denno, 388 U.S. 293, 87 S.Ct. 1967, 18 L.Ed.2d 1199 (1967); Johnson, supra; Linkletter, supra. Chevron Oil’s factors are intended to help courts bring this form of analysis to bear in other legal areas. The factors are not discrete, disembodied tests, but rather offer three perspectives on the central question of retroactivity: was reliance on a contrary rule so justified and the frustration of expectation so detrimental as to require deviation from the traditional presumption of retroactivity?
Before applying Chevron Oil’s factors to this case, we address a question on which the Supreme Court has not yet given explicit guidance: should the factors (and the issue of reliance) be analyzed in light of the particular litigants in a given case, or in light of general patterns throughout society? We address this question, first, because respondent BIW contends that Calbeck should not be applied retroactively in this case for the reason that Simpson was not denied a remedy but in fact received all that the state compensation system offered. We address it for the additional reason that language in Chevron Oil arguably supports a highly particularistic, if not atomistic, application of retroactivity, such a reading having been adopted in the recent case, Wachovia Bank and Trust Co. v. National Student Marketing Corp., 650 F.2d 342 (D.C.Cir.1980). We believe that the more general perspective should control.
The question is whether a new decision’s retroactivity is decided just once, or is decided anew in each case raising the issue. We know of no cases that have been applied prospectively to some litigants and retroactively to others. This is undoubtedly because such a rule would be unworkable. An individualized inquiry into reliance in each case would force courts to spend energy in each case determining how the law on an issue appeared at any given moment in history. It would also require detailed investigation into the facts bearing on the *86quality of particular parties’ reliance. Thus, once the Chevron Oil analysis is applied to a particular decision of law, the result — whether “retroactive” or “prospective” — should govern all subsequent cases.6 Such, of course, was the result when, for example, the ruling in Miranda v. Arizona, 384 U.S. 436, 86 S.Ct. 1602, 16 L.Ed.2d 694 (1966), was made retroactive, Johnson v. New Jersey, 384 U.S. 719, 733-34, 86 S.Ct. 1772, 1780-81, 16 L.Ed.2d 882 (1966); courts were not expected to search whether, in fact, suspects in custody had lawyers, or wanted to consult them, and whether, in fact, police officers had relied on cases such as People v. Hartgraves, 31 Ill.2d 375, 202 N.E.2d 33 (1964) that had narrowly construed Escobedo v. Illinois, 378 U.S. 478, 84 S.Ct. 1758, 12 L.Ed.2d 977 (1964).
Of course, that conclusion must influence the manner in which the inquiry is conducted; a general decision on retroactivity should not be controlled by specific, possibly aberrational facts. Although the circumstances of particular litigants may be important when they typify general patterns of reliance,7 case facts per se are not the relevant data in applying Chevron Oil. In terms of the values described above, the presumption of retroactivity (and the interest in uniformity) is outweighed when there are sufficient indications of justifiable societal reliance on a different rule. We now apply that analysis to Calbeck.
D. Chevron Oil’s First Factor
Chevron Oil’s first factor focuses on how much a decision changes the law. To overcome the presumption of retroactivity, the decision must establish “a new principle of law.” 404 U.S. at 106, 92 S.Ct. at 355. It can do so “either by overruling clear past precedent on which litigants may have relied ... or by deciding an issue of first impression whose resolution was not clearly foreshadowed.” Id. This requires an inquiry into the state of the law when the new decision was handed down.8 As BIW concedes, Calbeck was not an overruling decision. The opinion was explicit in reconciling its conclusion with prior Supreme Court decisions. Calbeck, supra, 370 U.S. at 127, 82 S.Ct. at 1203. BIW contends, however, and the Benefits Review Board held, that Calbeck “decid[ed] an issue of first impression” in a surprising manner that “was not clearly foreshadowed.”
We note at the outset the difficulty of arguing that a decision that does nothing but interpret a statute in light of its legislative history was not “clearly foreshadowed”. Nevertheless, we will not blink the fact that Calbeck’s line of analysis stunned most legal commentators;9 that considera*87tion alone leads us to assume, for purposes of careful analysis, that any “threshold” aspect of Chevron Oil’s first factor was met.10 Yet this is not a particularly significant step. Almost every plenary decision of the Supreme Court resolves an issue of first impression, narrowly defined, and hardly any is clearly foreshadowed. Analysis of Chevron Oil’s first factor requires more. It remains for us to determine whether it weighs heavily, moderately, or only slightly against retroactivity. The question is whether preexisting doctrine was so clear as to be likely to induce extensive commitments based on a particular different understanding of the law. Our review of the history leads us to conclude that the legal waters were so muddy before 1962 that it would have been quite imprudent to rely on the Rohde -3(a) syllogism.
We note first that the language of § 3(a) was too awkward to inspire much reliance. Many commentators did read it to imply that Rohde’s maritime-but-local rule had been incorporated in section 3(a). E.g., G. Gilmore & C. Black, The Law of Admiralty 346 (1957). But that reading was not problem-free, and the Calbeck construction was a linguistically defensible alternative.11 Additionally, the legislative history cut sharply against the Rohde -3(a) syllogism, see Calbeck, supra, 370 U.S. at 122-26, 82 S.Ct. at 1200-03; Note, Tulane L.Rev. 843, 845-56 & nn.10-12.
Nor can it be said that Supreme Court case law would have induced widespread reliance on the syllogism. After the passage of the Act, the Court continued to develop the maritime-but-local doctrine as a limitation on state jurisdiction, but it never held that the doctrine affected federal jurisdiction under § 3(a). The only conceivable Supreme Court support for such an inference came in a passing remark in Cro-well v. Benson, 285 U.S. 22, 41 & n.3, 52 S.Ct. 285, 288 & n.3, 76 L.Ed. 598 (1932). Yet holdings in subsequent cases made it clear that the maritime-but-local doctrine would not be used to oust federal jurisdiction. For example, Parker v. Motor Boat Sales, Inc., 314 U.S. 244, 62 S.Ct. 221, 86 L.Ed. 184, rev’g, 116 F.2d 789 (4th Cir. 1941), set forth a paradigmatic maritime-but-local fact situation: a janitor in a boat store “had been repeatedly and specifically cautioned and instructed that he was not to get on boats, ride in boats, or perform any services on navigable waters.” 116 F.2d at 791. After helping a customer mount an outboard motor, the janitor went out for a ride and drowned. The court of appeals *88held the federal claim barred by Rohde and § 3(a). The Supreme Court reversed, saying that, whether or not maritime-but-local governed these facts, Congress had intended in § 3(a) to provide compensation in the area that certain “specific decisions” — Jensen, Knickerbocker, and Dawson — had placed beyond the states’ reach. 314 U.S. at 248. And the following year, in Davis v. Department of Labor and Industries, 317 U.S. 249, 63 S.Ct. 225, 87 L.Ed. 246 (1942), the Court declared the existence of a “twilight zone" of injuries that could be compensated under either state or federal law. This holding was clearly incompatible with the first step in the Rohde-3(a) syllogism.
Nor was there much lower-court precedent that would have justified an employer in relying on the Rohde -3(a) syllogism until 1962. The only glimmer of encouragement for such a view came in the form of dictum. Royal Indemnity Co. v. Puerto Rico Cement Corp., 142 F.2d 237 (1st Cir. 1944), cert. denied, 323 U.S. 756, 65 S.Ct. 89, 89 L.Ed. 605 (1944), held § 3(a) inapplicable because the case was covered by the Defense Bases Act, but suggested that if § 3(a) had been applicable, it would be construed as limiting federal jurisdiction under the Act to areas in which federal power was exclusive. Id. at 238-39. Similar dictum was found in Union Oil Co. v. Pillsbury, 63 F.2d 925 (9th Cir. 1933).
But although such dicta existed, “[d]evi-ant rulings by circuit courts of appeals, particularly in apparent dictum, cannot generally provide the ‘justified reliance’ necessary to warrant withholding retroactive application of a decision construing a statute as Congress intended it.” United States v. Estate of Donnelly, 397 U.S. 286, 295, 90 S.Ct. 1033, 1038, 25 L.Ed.2d 312 (1970). Moreover, the holdings in cases that turned on the issue were uniformly to the contrary. In Continental Casualty Co. v. Lawson, 64 F.2d 802, 805 (5th Cir. 1933), the Fifth Circuit held that “Congress intended to exercise to the fullest extent all the power and jurisdiction it had over the subject matter.” And in DeBardeleben Coal Corp. v. Henderson, 142 F.2d 481, 483-84 (5th Cir. 1944), the same court waxed fervent:
“In the application of the act, .. . the broadest ground it permits of should be taken.... [T]he Federal Compensation Law now in effect, the judicial tergiver-sations which went on before its passage no longer have point. This is what we held in the Lawson case, what the Supreme Court held in the Parker case, supra. We adhere to that holding. The judgment was right. It is affirmed.”
This was precisely the Calbeck analysis, eighteen years before Calbeck.12 Similarly in Travelers Ins. Co. v. McManigal, 139 F.2d 949 (4th Cir. 1944), the court held that, in interpreting § 3(a) of the federal act, “there is no room for the application of the rule announced in cases of the Rohde class.” Id. at 953. See also Travelers Ins. Co. v. Branham, 136 F.2d 873 (4th Cir. 1943).
Thus, Chevron Oil’s first factor does nothing to rebut Calbeck’s presumptive retroactivity. Indeed, it supports it.
E. Chevron Oil’s Second Factor
Chevron Oil’s second factor draws us to consider whether retroactivity would be inconsistent with the purposes of Calbeck. The case sought to implement Congress’s desire that the Longshoremen’s Act provide the federal scheme of compensation to a certain, clearly identified class of workers. Retroactive application would provide that certainty equally to all employees injured since the passage of the Act. Nonretroac-tivity would leave workers injured before 1962 without the certain remedy that the 1927 Congress supposedly intended to provide. Thus, retroactivity would be consistent with Calbeck’s purpose, and nonretro-activity inconsistent.
In its opinion below, the Benefits Review Board majority found nonretroactivity wholly consistent with Calbeck’s purposes. *89It reasoned, first, that Calbeck could be given limited retroactive effect, applying only to litigants who (unlike Simpson) had never accepted benefits under a state scheme. Second, it suggested that general retroactivity would lose “a flood of litigation” on the courts, thereby promoting the uncertainty and lawsuits that Calbeck sought to eliminate.
As we have noted above, it is error as a general proposition to assume that a single rule can be applied retroactively as to some litigants but not as to others. This point has special strength in analyzing Calbeck. In Calbeck itself relief was unquestionably available to the two workers under state law, and benefits had in fact been paid under state law in one of the two cases. The Court nevertheless reasoned broadly, and supported its construction of section 3(a) by noting that in other, more typical cases, such a remedy might not be assured. 370 U.S. at 125, 82 S.Ct. at 1202. It would be peculiar indeed if the analysis of retroac-tivity should be bound by particular case facts when the underlying decision was not so constrained.
Similarly, we are unmoved by the Board’s fear of a flood of litigation. The majority expressed its fear as a bald assertion, unsupported by evidence or expert predictions. Moreover, that fear is illogical. At the time Calbeck was decided, the theory of federal nonretroactivity was still inchoate. Any flood that the decision might have stimulated would certainly have come early in the 1960’s, not in the 1980’s.
F. Chevron Oil’s Third Factor
Chevron Oil’s third factor requires us to consider the equities of retroactivity. Claimants in Simpson’s position assert that they should receive the benefits that the 1927 Congress said they were entitled to. The interests of employers in BIW’s position are those of reliance — allegedly an employer of a Rohde -type employee would have chosen to forgo insurance against liability under the federal act. The equitable strength of these reliance interests varies with both their justification and their extent. As we have seen, an employer would have had only slight justification for relying on the Rohde -3(a) syllogism.
Furthermore, the extent of that reliance — though not easily gauged — does not seem to have been substantial. One possible approach to measuring the extent of reliance might be to determine the potential gross liability that employers in BIW’s position assumed by not carrying federal liability insurance before Calbeck. The record contains no evidence of how many employers forwent federal liability insurance because of the Rohde-3(a) syllogism, how many of those employers’ workers would actually have been declared maritime-but-local, how many of those employees actually became injured, and how much the injured employees were entitled to receive in federal benefits. Given the ceiling on federal compensation13 (even before subtracting any benefits already recovered under state law), we doubt the final figure would be very large.
An alternate approach to measuring the extent of reliance might be to compute the amount of money employers in BIW’s position would have saved by limiting their insurance coverage to state liability. The record again gives no guidance, but a widely cited secondary source reports that, at *90the time Calbeck was decided, there was “a standard workmen’s compensation policy used by all carriers on a nationwide basis, which may be extended by endorsements, to cover operations subject to the Longshoremen’s Act ... [T]he only extra expense to the employer for coverage under both acts is a small fee that the carriers charge for the endorsement to the second act.” Note, 50 Cal.L.Rev. 342, 347 (1962). Accord, Comment, NACCA L.J. 200, 203, 206 (1964); Gardner, Remedies for Personal Injuries to Seamen, Railroadmen, and Longshoremen, 71 Harv.L.Rev. 438, 449-50 & n.34 (1958). No matter how one approaches the measurement of employer reliance, we have no reason to think it was particularly extensive.
We think guidance in balancing the equities of employers and employees may be taken from Usery v. Turner Elkhorn Mining Co., 428 U.S. 1, 96 S.Ct. 2882, 49 L.Ed.2d 752 (1976). That case upheld the retroactive application of the Black Lung Benefits Act of 1972. The statute required coal mine operators to compensate former employees who terminated their work in the industry before the Act was passed and who developed pneumoconiosis after leaving their jobs. Even though it was possible “that the Operators may not have known of the danger of their employees’ contracting pneumoconiosis, and that even if they did know of the danger their conduct may have been taken in reliance upon the current state of the law, which imposed no liability on them for disabling pneumoconiosis”, the statute was upheld in order “to spread the costs of the employees’ disabilities to those who had profited from the fruits of their labor — the operators and the coal consumers.” Id. at 17-18, 96 S.Ct. at 2893-2894. Even though the presumptive validity of a statute may be greater than the presumptive retroactivity of a judicial decision, Usery is relevant here as a clear evaluation of the balance of equities between disabled employees and employers who claim reliance on prior law.
When the situation is analyzed in light of Chevron Oil’s three factors, we are not persuaded that Calbeck was decided in the face of such widespread justifiable reliance on the Rohde -3(a) syllogism as to overcome the general presumption of retroactivity. Calbeck’s holding should therefore govern all decisions under the 1927 Act.
III. Other Asserted Grounds for Affirmance
The respondent BIW urges that even if Calbeck was retroactive, Simpson’s claim should have been dismissed in accord with the Fifth Circuit’s holding in Shea v. Texas Employers Insurance Assoc., 383 F.2d 16 (5th Cir. 1967). This argument was accepted by the Administrative Law Judge, but rejected by the Benefits Review Board.
In She a, the Fifth Circuit held that a person who had already recovered through state court litigation was collaterally es-topped from arguing in federal court that he satisfied the “may not validly be provided” clause in § 3(a). The court understood Calbeck to be a statement that an accident sustained on navigable waters was presumptively beyond the scope of valid state coverage, but where a state court decision explicitly rebutted that presumption, collateral estoppel would not allow it to be revived later in federal court. Thus, the Fifth Circuit would apply Calbeck only in the absence of state litigation.
We agree with the Benefits Review Board that She a misapprehended the import of Calbeck. Calbeck was not just another “twilight zone” case establishing a presumption of federal jurisdiction. Rather, Calbeck held that Congress had actually (albeit inartfully) authorized “federal compensation for all injuries to employees on navigable waters.” 370 U.S. at 120, 82 S.Ct. at 1200. That authorization was to be accepted regardless of what a particular claimant had been able to recover under state law. Thus, although a state court opinion could collaterally estop the litigant from debating the scope of state court jurisdiction, the question of state court jurisdiction was simply not relevant in cases under the federal act. See generally 4 Larson’s Workmen’s Compensation Law § 89.53(b) (1981).
The respondent’s remaining arguments were not passed upon by the Benefits Re*91view Board below. We believe that its arguments based upon the situs of the injury and the theories of election of remedies and res judicata are without merit. The argument that Simpson’s injury is not covered because he fell onto the ground is inconsistent with the reasoning in Minnie v. Port Huron Terminal Co., 295 U.S. 647, 55 S.Ct. 884, 79 L.Ed. 1631 (1935). The election of remedies argument was explicitly rejected by the Supreme Court in Calbeck, supra, 370 U.S. at 131-32, 82 S.Ct. at 1205-06, and again in Sun Ship, Inc. v. Pennsylvania, 447 U.S. 715, 724, 100 S.Ct. 2432, 2438, 65 L.Ed.2d 458 (1980). And any res judicata argument (to the extent distinguishable from Shea’s collateral estoppel argument) is refuted by the fact that claims under the Longshoremen’s and Harbor Workers’ Compensation Act may not be pressed in state court. The procedures for pressing claims under the Act are set forth in 33 U.S.C. § 919 et seq. We do not pass on the issue of laches which was not considered by the Board, or briefed or argued by either party in this appeal.
The decision of the Benefits Review Board is vacated and the cause is remanded for further proceedings consistent with this opinion.
. In 1972, Section 3(a) was amended. The parties agree that those amendments had no effect on Simpson’s already existing claim.
. The relevant text of 33 U.S.C. § 903(a) before the 1972 amendments was:
“(a) Compensation shall be payable under this Act in respect of disability or death of an employee, but only if the disability or death results from an injury occurring upon the navigable waters of the United States (including any drydock) and if recovery for the disability or death through workmens’ compensation proceedings may not validly be provided by State law.”
. Justice Stewart and Justice Harlan dissented. 370 U.S. at 132-38, 82 S.Ct. at 1206-09. They dubbed the majority approach an “exercise in judicial legerdemain”, arguing that the clear language of the statute supported the Rohde -3(a) syllogism. They offered an alternative explanation for the legislative history, and they accused the majority of departing from precedent.
. The early history of the theory of prospective judicial decisions is reviewed in detail in Linkletter v. Walker, 381 U.S. 618, 85 S.Ct. 1731, 14 L.Ed.2d 601 (1965). The associated theoretical problems have long been a source of interest to scholarly commentators. See, e.g., Mishkin, Foreword: The High Court, The Great Writ, and the Due Process of Time and Law, 79 Harv.L.Rev. 56, 58-69 (1965); Schaefer, The Control of ‘Sunbrusts’: Techniques of Prospective Overruling, 42 N.Y.U.L.Rev. 631 (1967); R. Keeton, Venturing To Do Justice 25-53 (1969); Beytagh, Ten Years of Non-retroactivity: A Critique and a Proposal, 61 Va.L.Rev. 1557 (1975); R. Aldisert, The Judicial Process 877-938 (1976).
. In theory, at least, this is a result of Article Ill’s requirement that the federal courts resolve only cases and controversies. See Stovall v. Denno, 388 U.S. 293, 301, 87 S.Ct. 1967, 1972, 18 L.Ed.2d 1199 (1967). Statements of principle that do not apply to the case that spawns them are technically dictum, even though they may in practice be accorded precedential force. Cf. Morrissey v. Brewer, 408 U.S. 471, 490, 92 S.Ct. 2593, 2604, 33 L.Ed.2d 484 (1972).
. We therefore must withdraw our dictum in Jackson v. Justices of the Superior Court of Mass., 549 F.2d 215, 218 n.2 (1st Cir. 1977), where we implied that our holding of non-retro-activity might be limited to certain would-be beneficiaries of the new rule.
. This has been the case in several Supreme Court decisions on retroactivity. E.g., Chevron Oil v. Huson, 404 U.S. 97, 92 S.Ct. 349, 30 L.Ed.2d 296 (1971); Allen v. State Board of Elections, 393 U.S. 544, 572, 89 S.Ct. 817, 835, 22 L.Ed.2d 1 (1969); England v. State Board of Medical Examiners, 375 U.S. 411, 84 S.Ct. 461, 11 L.Ed.2d 440 (1964). In each case the Court buttressed its analysis by making reference to the facts of the parties before it and then phrased its finding on retroactivity in general terms.
. In Wachovia Bank and Trust Co. v. National Student Marketing Corp., 650 F.2d 342 (D.C.Cir.1980), the court concluded that one should analyze the state of the law at the time a particular claim arose rather than at the time of the lawmaking decision. We decline to follow Wachovia since it implicitly assumes that Chevron Oil’s factors should be analyzed in light of particular litigants’ reliance interests. See three previous paragraphs in text supra.
. For commentators surprised by the breadth of Calbeck, see G. Gilmore & C. Black, The Law of Admiralty (2d ed. 1976); H. Baer, Admiralty Law of the Supreme Court § 6-4 (2d ed. 1969); Note, 17 Southwestern L.J. 345, 350 (1963); The Supreme Court, 1961 Term, 76 Harv.L.Rev. 54, 95-98 (1962); Note, 15 S.Cal.L.Rev. 982 (1963). But see Note, 1963 Duke L.J. 327, 331 (“Calbeck was forecast by past decisions”); Note, 35 Tulane L.Rev. 841 (1962) (suggesting that the Fifth Circuit’s opinion in Calbeck was inconsistent with the legislátive history); Note, 50 Cal.L.Rev. 342, 345 (1962) (suggesting that the Fifth Circuit’s opinion in Calbeck was inconsistent with Davis ).
It bears mention that Calbeck was not a case where the Court based its decision on a new *87theory not imagined by the parties. Solicitor General Cox and Assistant Attorney General Orrick had presented the argument in full detail in their brief for appellant.
. It has frequently been said that an issue of retroactivity “is not even presented” unless a decision satisfies a threshold condition of novelty. This condition has usually been termed a “sharp break” requirement. See, e.g., Hanover Shoe, Inc. v. United Shoe Machinery Corp., 392 U.S. 481, 499, 88 S.Ct. 2224, 2234, 20 L.Ed.2d 1231 (1968). See also United States v. Peltier, 422 U.S. 531, 544 & n.1, 95 S.Ct. 2313, 2321 & n.1, 45 L.Ed.2d 374 (1974) (Brennan, J., dissenting) (suggesting that the condition has been abandoned); Milton v. Wainwright, 407 U.S. 371, 381 n.2, 92 S.Ct. 2174, 2180 n.2, 33 L.Ed.2d 1 (1972) (Stewart, J., dissenting). We see no advantage in attempting to specify where on the continuum of sharpness that threshold might be, if it still exists. Instead we consider the manner in which Calbeck changed the law as one of several factors bearing on the justification for reliance on the Rohde -3(a) syllogism.
. The relevant statutory language is, “if recovery ... may not validly be provided by State law." The Rohde-3(a) syllogism interprets the word “may” in a constitutional sense, reading the statute to mean:
“If it would be unconstitutional for state law to provide compensation, then there is federal jurisdiction.”
As has been noted in D. Robertson, Admiralty and Federalism 311-18 (1970), this interpretation makes the statutory word “validly” quite superfluous. An alternative interpretation, consistent with Calbeck, would interpret the word “may” in a different sense, reading the statute to mean:
“If it is possible that State law might fail to provide constitutionally valid compensation, then there is federal jurisdiction.”
Id. Although we do not suggest that we find this latter interpretation more natural in the absence of any legislative history, we are content to note that — on a purely linguistic level— the statute is not unambiguous.
. Indeed, the Fifth Circuit opinion that was reversed by the Supreme Court in Calbeck had sought to overrule DeBardeleben and Lawson. See Travelers Ins. Co. v. Calbeck, supra, 293 F.2d at 58.
. Between 1927 and 1948, compensation under the federal act was limited to $7500. Longshoremen’s and Harbor Workers’ Compensation Act of 1927, ch. 509, § 14(m), 44 Stat. 1432. Between 1948 and 1956, the ceiling was removed for cases of permanent total disability or death, and was raised to $11,000 (sometimes $10,000) for other cases. Act of June 24, 1948, ch. 623, § 5, 62 Stat. 603. Between 1956 and 1961 the ceiling was raised to $17,280. Act of July 26, 1956, ch. 735, § 5, 70 Stat. 655. Between 1961 and Calbeck’s decision in 1962, the ceiling was raised to $24,000. Act of July 14, 1961, P.L. 87-87, § 3, 75 Stat. 203.
Of course “compensation” does not encompass reimbursement for medical expenses. Thus, § 14(m)’s ceiling only limited employers’ potential retroactive liability for the other (less easily measured) aspects of maritime-but-local workers’ damages. We mention the ceiling only as one factor supporting our conclusion that the employers’ potential gross liability is not so extraordinary as to deserve special weight.