concurring in part and dissenting in part:
While I concur in the majority’s conclusion that Kregos has displayed sufficient creativity to satisfy the Feist Publications standard for copyrightability, I would affirm the district court’s grant of summary judgment because I conclude that Kregos’ idea here has merged into his expression.
1. Kregos’ Idea
I agree with the majority that Kregos displayed at least some creativity in selecting his nine statistics for his pitching form, and thereby satisfied the creativity standards of Feist Publications. However, I believe that this selection was not simply an expression of the idea that this information might be of interest to sports fans, but instead represented his very specific idea that these nine statistics were the most significant ones to consider when attempting to predict the outcome of a baseball game. I respectfully disagree with the majority’s statement that Kregos’ idea was the abstract “general idea that statistics can be used to assess pitching performance,” 1 because while I agree that the application of the merger doctrine will always depend on the level at which the idea is formulated, I do not believe that the majority has set forth convincing grounds for its determination as to the idea at issue here.
If few people had previously considered using pitching statistics to handicap baseball games, I would agree that Kregos’ particular selection of statistics might indeed be entitled to protection as a detailed expression of his idea. Here, however, the format and the arrangement of data existed prior to his choice of particular items to report, and thus his “creation” was nothing more than that choice, which as an expression seems to me inseparable from its idea.
I have difficulty distinguishing between Kregos’ work and the example of a doctor’s identification of the symptoms to use in diagnosing a disease, which the majority suggests would be unprotectible because of merger. In making this suggestion, the majority eschews the notion that the doctor’s idea is the “general idea that the disease in question can be identified by observable symptoms” in favor of the more precise formulation that the idea is that “his selection of symptoms is a useful identifier of the disease.” Majority Opinion at 707. This reasoning would appear to apply equally well to Kregos’ case. In both cases, the creators have conceived very precise “ideas” concerning the significant data which ought to be considered in predicting a given result, and those ideas can be expressed only by identifying the relevant data. Of course, there is the obvious distinction that a system for making medical diagnoses is more socially beneficial than system for estimating sports odds, but such a distinction does not offer a basis for denying copyright protection to one while granting it to the other.2
*712The seminal articulation of the basic tension between idea and expression is found in Baker v. Selden, 101 U.S. 99, 25 L.Ed. 841 (1879). Although the case is most often cited for its formulation of the “blank form” doctrine, in fact that issue was secondary to the decision. Selden had devised a new bookkeeping system and had written a book explaining its use and published ledger forms to be used when practicing the system. When the defendant published his own forms, slightly different from Sel-den’s but similar enough that they could be used to practice Selden’s system, Selden sued, claiming that Baker’s forms had infringed his copyright in the book.
In denying relief, the Court relied primarily on the proposition that “[t]he use of the art is a totally different thing from a publication of the book explaining it.” 101 U.S. at 104. In discussing this proposition, the Court commented
The copyright of a work on mathematical science cannot give to the author an exclusive right to the methods of operation which he propounds, or to the diagrams which he employs to explain them, so as to prevent an engineer from using them whenever occasion requires. The very object of publishing a book on science or the useful arts is to communicate to the world the useful knowledge which it contains. But this object would be frustrated if the knowledge could not be used without incurring the guilt of piracy of the book. And where the art it teaches cannot be used without employing the methods and diagrams used to illustrate the book, or such as are similar to them, such methods and diagrams are to be considered as necessary incidents to the art, and given therewith to the public; not given for the purpose of publication in other works explanatory of the art, but for the purposes of practical application.
Id. 101 U.S. at 103 (emphasis added). The Court therefore concluded that Selden’s valid copyright in his book could not be extended to prevent others from using his system or the forms.
Only after it had reached this conclusion did the Court find it necessary to consider whether Selden’s forms were themselves copyrightable. After reviewing several English authorities dealing with the limits of copyrightability, the Court concluded that “blank accountbooks are not the subject of copyright,” and therefore dismissed Selden’s claims. 101 U.S. at 107. It is this second holding, based on a rather abbreviated analysis, which has become known as the blank form doctrine. Due in large part to the limited explanation for this holding in Baker, the blank form doctrine has generated numerous questions and critiques, see, e.g., 1 Nimmer on Copyright § 2.18[C] (1990), and not a few outright inconsistent decisions. Compare Harcourt, Brace & World, Inc. v. Graphic Controls Corp., 329 F.Supp. 517 (S.D.N.Y.1971) and Norton Printing Co. v. Augustana Hospital, 155 U.S.P.Q. 133 (N.D.Ill.1967) with Bibbero Systems, Inc. v. Colwell Systems, Inc., 893 F.2d 1104 (9th Cir.1990) and John H. Harland Co. v. Clarke Checks, Inc., 711 F.2d 966 (11th Cir.1983).
On the other hand, the Baker Court’s initial holding, distinguishing between an “art” and an explanation of that art, has not been seriously challenged, and has steadily evolved and been refined into what is commonly referred to today as the idea/expression dichotomy, a concept which includes within its bounds the special case of merger. Of course, this doctrine has also generated extensive commentary and has produced its share of contradictory decisions, but in contrast to the blank form doctrine the major concern with respect to the idea/expression dichotomy has been how to apply it in any particular case, i.e., how to select the appropriate level of abstraction, and not whether the doctrine is itself fundamentally sound.
Despite the numerous developments in this area since Baker v. Selden and the myriad complexities which have been graft*713ed onto the basic doctrine, I see the present case as falling directly within the scope of Baker’s original discussion of the explanation/use distinction. In my opinion, Kre-gos’ form constitutes an explanation of his preferred system of handicapping baseball games, and he seeks to use his copyright here to prevent others from practicing that system. This seems to present a clear example of a case where “the art [the work] teaches cannot be used without employing the methods and diagrams” which make it up. Of course, one difference between this case and Baker is that Kregos’ copyright is not in a separate work explaining how to use the pitching form, but rather covers the form itself. However, the pertinent issue in Baker was not that Sel-den’s book explaining the system was separable from the forms for using the system (in fact the forms were part of the book), but rather was that he sought to use his copyright in the book to prevent others from using the system. It was this use of the copyright which the Baker court condemned, and regardless of any intervening developments concerning the idea/expression issue, the reasoning remains as sound today as the day it was written.
Both on the summary judgment motion in the district court and on this appeal, Kregos has asserted that his form should not be held uncopyrightable as a blank form because it “convey[s] to his subscribers the categories of information which Kregos himself subjectively considers to be important for evaluating pitching performance.” Appellants’ Brief at 30; see also 731 F.Supp. at 120. Thus Kregos admits that his form incorporates an explanation of his outcome-predicting system and that in distributing the form he intends to communicate that system to his customers. Once he has done voluntarily so, Baker prevents him from relying on the copyright to prevent others from using the system on their own.
The majority characterizes Kregos’ work as dealing with “matters of taste and opinion,” and therefore compares it to the list of baseball card prices in Eckes v. Card Prices Update, 736 F.2d 859 (2d Cir.1984) or to a listing of socially prominent families rather than to the hypothetical doctor’s diagnostic chart. In my view, both the pitching form and the diagnostic chart are expressions intended to assist in predicting particular outcomes, with the data intended to be used as a basis for that prediction. In contrast, neither the card price list nor the social register is associated with any defined event or result, and the information reported — the card prices, the names of the families — is itself the primary feature or attraction.3
I find this situation closer to that in Matthew Bender & Co. v. Kluwer Law Book Publishers, Inc., 672 F.Supp. 107 (S.D.N.Y.1987), in which the plaintiff sought protection only for its method of arranging information concerning personal injury awards. There, as here, there was no allegation that the defendant had copied the plaintiff’s data, but merely that it had arranged its own data in a format which was similar to that used by the plaintiff. In denying relief to the plaintiff, Judge Conner held that the idea of the plaintiff’s work was to provide useful information in a functional format, and that the copyright could not be used to preclude others from adopting that format. 672 F.Supp. at 110.4
*714In contrast, the issue in Educational Testing Services v. Katzman, 793 F.2d 533 (3d Cir.1986), was not whether the defendant could arrange its own original questions in the plaintiffs format, but whether it could copy the plaintiffs copyrighted questions for its own use. The court held that the proper level of abstraction for the ideas which underlay the plaintiffs questions was “the rules of punctuation, analogies, vocabulary,” and concluded that “[ojther persons, similarly resourceful, have ample latitude and opportunity to frame noninfringing questions testing the same subjects.” 793 F.2d at 540. This conclusion at least implies that the plaintiff would not have been entitled to protect the categories or the distribution of its questions, but only the specific questions themselves.
Similarly, in Toro Co. v. R & R Products Co., 787 F.2d 1208 (8th Cir.1986), the question was not whether the defendant could use the plaintiff’s format for a parts numbering list, but whether it could copy the data contained in that format. Because there was no preexisting relationship between the parts and their numbers, the court held that there were many other possible “expressions” of the “idea” of a numbering system, and thus concluded that there was no merger. 787 F.2d at 1212.
While I agree with the majority’s statement that “[ejvery compiler of facts has the idea that his particular selection of facts is useful,” I believe that there is a difference between the idea that a selection is useful and the more precise idea that it is the most useful selection possible.
The less precise idea is found in cases of compilations which are based on aesthetic choices, such as a list of socially-prominent families, or those in which the value of the compilation is due more to the fact of compilation than to the intrinsic value of the data being compiled: this would include the parts numbering list at issue in Toro Co. v. R & R Products, where the numbers were not significant prior to being associated with the parts, the test questions in Educational Testing Services, where the questions were valuable to the defendants precisely because they had been used by the plaintiff, and, to the extent the selection was in fact subjective, the premium card list in Card Prices, where the defendants had no independent knowledge of which cards it would be profitable to update.
The more precise idea is found where the compiler’s main purpose is to compile data which will allow a reader of the compilation to reach a clearly-defined goal, such as the case reports in Matthew Bender, where the goal was to estimate the possible settlement value of a case, or a doctor’s diagnostic form, where the goal would be to diagnose the given disease, or Kregos’ pitching form, where the goal is to predict the outcome of the game.
Finally, in light of the majority’s agreement with the district court that Kregos’ arrangement of the statistics was not itself creative or original, and therefore that his particular ordering is not protected, it is difficult to grasp exactly what “expression” the majority intends to protect, if not the fundamental expression that these nine items are valuable in predicting games. This difficulty becomes apparent as the majority speculates about the extent of protection to be given to Kregos’ form.
2. The Application of the Merger Doctrine
As a secondary matter, I disagree with the majority’s characterization of how the merger doctrine is applied in this Circuit. In Durham Industries v. Tomy Corp., 630 F.2d 905 (2d Cir.1980), the court did not face the question of merger at all: in that case there were two dissimilar expressions which clearly shared only an underlying idea, and the court refused to extend the plaintiff’s copyright on its expression to the idea, holding that there was no infringement of the expression, whether or not the defendant had copied the idea. The *715only aspect of the case which touched upon the merger doctrine was the initial finding that there were two dissimilar expressions of the same idea, which implicitly rejected any claim that the idea and the expression had not merged.5
On the other hand, those cases which actually address the question of merger are those in which the copyrighted work and the allegedly infringing one appear at first glance to be similar or identical. See, e.g., Kern River Gas Transmission Co. v. Coastal Corp., 899 F.2d 1458, 1463 (5th Cir.), cert. denied, — U.S.-, 111 S.Ct. 374, 112 L.Ed.2d 336 (1990); Toro Co. v. R & R Products Co., 787 F.2d 1208, 1212 (8th Cir.1986); Apple Computer, Inc. v. Franklin Computer Corp., 714 F.2d 1240, 1253 (3d Cir.1983), cert. dismissed, 464 U.S. 1033, 104 S.Ct. 690, 79 L.Ed.2d 158 (1984); Herbert Rosenthal Jewelry Corp. v. Kalpakian, 446 F.2d 738, 742 (9th Cir.1971); Lotus Development Corp. v. Paperback Software International, 740 F.Supp. 37, 61 (D.Mass.1990); Matthew Bender & Co., supra, 672 F.Supp. at 109; Freedman v. Grolier Enterprises, Inc., 179 U.S.P.Q. 476, 478 (S.D.N.Y.1973). Only in such a situation does it make sense to inquire whether the apparent similarity is required by the use of the plaintiff’s unprotectible expression, or whether there are enough alternative expressions of that idea that the plaintiff’s choice deserves protection.
I believe the proper approach requires the court first to decide whether the copyrighted work satisfies the primary requirement of creativity, then to determine whether there is merger before extending copyright protection. This is based on the wording of § 102(b) of the Copyright Act, 17 U.S.C. § 102(b), which provides
In no case does copyright protection for an original work of authorship extend to any idea, procedure, process, system, method of operation, concept, principle, or discovery, regardless of the form in which it is described, explained, illustrated, or embodied in such work.
I interpret this language as indicating that protection cannot be given to a work which is inseparable from its underlying idea.
The Nimmer treatise supports the majority’s approach, suggesting that merger must be considered in the context of determining whether infringement has occurred rather than in deciding the issue of copy-rightability. Under this approach, a court which finds that merger exists should hold that the two works in question are not “substantially similar,” even where they are in fact identical, a result which I view as a not useful variety of doublespeak.6
Nimmer notwithstanding, the majority of cases have rejected this approach and instead followed the method in which merger becomes an issue only when the two works in question — the copyrighted one and the alleged infringement- — appear on the surface to be similar, and under which merger is used as a reason for denying all copyright protection to the plaintiff and thereby excusing the defendant’s use of a similar or even identical expression. See Kern River Gas Transmission Co. v. Coastal Corp., 899 F.2d 1458, 1463 (5th Cir.), cert. denied, — U.S. -, 111 S.Ct. 374, 112 L.Ed.2d 336 (1990); Toro Co., 787 F.2d at 1212; Herbert Rosenthal Jewelry, supra, 446 F.2d at 742; Matthew Bender & Co., 672 F.Supp. at 109 (work which does not “exceed[ ] the boundaries of ‘idea’ and enter[ ] the realm of ‘expression’ ” is not protective); Freedman v. Grolier Enterprises, supra, 179 U.S.P.Q. at 478 (“When an idea is so restrictive that it necessarily requires a particular form of expression, that is, when the idea and its expression are functionally inseparable, to permit the copyrighting of the expression would be to *716grant the copyright owner a monopoly of the idea”).
The difference in applying these two approaches is not insignificant. Nimmer’s method lends itself much more readily to the erroneous conclusion that merger is only available where the defendant has independently created an expression which happens coincidentally to be similar to the plaintiffs work. Merger is then viewed as a means of explaining the unintentional similarity between the two works — thus Nimmer’s characterization of it as a means of negating substantial similarity. In other words, if a defendant has actually copied the plaintiff's work, it is unlikely to be allowed to rely on merger to avoid liability. See, e.g., NEC Corp. v. Intel Corp., 1989 Copyright Law Decisions (CCH) 11 26,379 at p. 22,390, 1989 WL 67434 (N.D.Cal.1989) (defendant demonstrated independent “clean room” creation of software similar to plaintiff’s to support its claim that similarities were required by external factors). This approach owes little if anything to the strictures of § 102(b), and instead depends on the fundamental principle of copyright law that independent creation is never infringement.7
The more common approach, in which merger is considered as part of the determination of copyrightability, absolves even a defendant who has directly copied the plaintiff’s work if the idea of that work is merged into the expression. See Kern River Gas, supra, 899 F.2d at 1463; Herbert Rosenthal Jewelry, supra, 446 F.2d at 742 (“When the ‘idea’ and its ‘expression’ are thus inseparable, copying the ‘expression’ will not be barred, since protecting the ‘expression’ in such circumstances would confer a monopoly of the ‘idea’) (emphasis added; citations omitted). I believe this approach accords more fully with both the language and the purpose of § 102(b), and serves to focus consideration on the proper definition of the idea at the outset of the inquiry.
As a final note, I have no quarrel with the majority’s statement that “[assessing merger in the context of alleged infringement will normally provide a more detailed and realistic basis for evaluating” a claim of merger: the analysis is greatly simplified by the example of the defendant’s work and by the defendant’s argument as to how the plaintiff’s idea has merged with its expression. This does not alter my conclusion that the inquiry must involve the copyrightability of the plaintiff’s work rather than the defendant’s infringement.
Therefore, because I believe that the district court was correct in holding that Kre-gos’ form is uncopyrightable because his idea has merged with his expression of that idea, I respectfully dissent.
. While this statement may seem more consistent with Judge Goettel’s statement that Kregos’ idea was "to publish an outcome predictive pitching form,” 731 F.Supp. at 119, I believe that Judge Goettel’s analysis of the merger issue and his finding that there was merger here indicate that he in fact considered Kregos' idea to be more specific than his initial statement suggests.
. Nor is it appropriate to distinguish based on a claim that diagnosing a disease is more “scien-*712tifie” and less "creative” than handicapping a baseball game. No doubt many people, including both doctors and those who make their living, either legally or illegally, in the sports gambling profession, would dispute such a claim.
. I would also suggest that, to the extent that Card Prices turned on the plaintiffs’ “selection, creativity and judgment” in selecting the 5,000 premium cards from the universe of 18,000 cards in all, see 736 F.2d at 863, the court appears to have overlooked the fact that the cards identified as "premium” were basically those with the highest market prices. Given that the market price data itself was within the public domain, it is not clear what creativity could have been involved in the mere identification of the high priced cards. In fact, the attention given to the amount of labor invested by the plaintiffs in developing their list, see 736 F.2d at 860, and the nature of the alleged infringement — copying the plaintiffs’ premium list but using it to report updated price data — suggests that the decision was at least in some part based on the "sweat of the brow” test, notwithstanding the court’s explicit recognition that “sweat of the brow” was an inappropriate measure of copyrightability. Id. at 862.
. Judge Conner’s comment that the plaintiff’s categories were "the only sensible ones which could have been used to compile the data,” 672 F.Supp. at 112, cited by the majority as grounds upon which to distinguish that case from the *714present one, related not to the merger question but rather to the additional determination that the plaintiffs form did not demonstrate sufficient creativity to warrant a copyright. Thus while Bender may be distinguishable, the distinction does not affect the applicability of Judge Conner’s merger analysis to the present case.
. The question of merger is fundamentally disjunctive: if there are only a limited number of ways in which to express an idea, there is merger, otherwise there is not. There can be no "partial merger” of an idea and an expression of that idea. Thus the determination that numerous distinct expressions exist for a given idea is equivalent to a determination that no merger has occurred.
. Thus Nimmer places the discussion of merger under the heading of "Negating Substantial Similarity." See 3 Nimmer on Copyright § 13.03[B] at 13-48 (1990).
. In NEC v. Intel, this method of applying merger was justified on the grounds that "[t]he Register of Copyrights will not know about the presence or absence of constraints that limit ways to express an idea.” 1989 Copyright Law Decisions at 22,381. This statement misunderstands the majority approach: the Register of Copyrights is not expected to determine whether to deny registration because of merger, as the fact of registration offers no more than prima facie evidence of copyrightability. E.g., Weissman v. Freeman, 868 F.2d 1313, 1320 (2d Cir.1989). Just as a plaintiff with a registered copyright may nevertheless be denied protection if the work lacks creativity, e.g., Feist Publications, supra, the presumption of validity arising out of registration may be overcome by proof to the court that the expression merges with the idea.