The primary issue on this appeal is whether the creator of a baseball pitching form is entitled to a copyright. The appeal requires us to consider the extent to which the copyright law protects a compiler of information. George L. Kregos appeals from the April 30, 1990, judgment of the District Court for the Southern District of New York (Gerard L. Goettel, Judge) dismissing on motion for summary judgment his copyright and trademark claims against the Associated Press (“AP”) and Sports Features Syndicate, Inc. (“Sports Features”). We affirm dismissal of the trade*702mark claims, but conclude that Kregos is entitled to a trial on his copyright claim, though the available relief may be extremely limited.
Facts
The facts are fully set forth in Judge Goettel’s thorough opinion, 731 F.Supp. 113 (S.D.N.Y.1990). The reader’s attention is particularly called to the appendices to that opinion, which set forth Kregos’ pitching form and the allegedly infringing forms. Id. at 122-24 (Appx. 1-4). Kregos distributes to newspapers a pitching form, discussed in detail below, that displays information concerning the past performances of the opposing pitchers scheduled to start each day’s baseball games. The form at issue in this case, first distributed in 1983, is a redesign of an earlier form developed by Kregos in the 1970’s. Kregos registered his form with the Copyright Office and obtained a copyright. Though the form, as distributed to subscribing newspapers, includes statistics, the controversy in this case concerns only Kregos’ rights to the form without each day’s data, in other words, his rights to the particular selection of categories of statistics appearing on his form.
In 1984, AP began publishing a pitching form provided by Sports Features. The AP’s 1984 form was virtually identical to Kregos’ 1983 form. AP and Sports Features changed their form in 1986 in certain respects, which are discussed in part 1(d) below.
Kregos’ 1983 form lists four items of information about each day’s games — the teams, the starting pitchers, the game time, and the betting odds, and then lists nine items of information about each pitcher’s past performance, grouped into three categories. Since there can be no claim of a protectable interest in the categories of information concerning each day’s game, we confine our attention to the categories of information concerning the pitchers’ past performances. For convenience, we will identify each performance item by a number from 1 to 9 and use that number whenever referring to the same item in someone else’s form.
The first category in Kregos’ 1983 form, performance during the entire season, comprises two items — won/lost record (1) and earned r-un average (2). The second category, performance during the entire season 1 against the opposing team at the site of the game, comprises three items— won/lost record (3), innings pitched (4), and earned run average (5). The third category, performance in the last three starts, comprises four items — won/lost record (6), innings pitched (7), earned run average (8), and men on base average (9). This last item is the average total of hits and walks given up by a pitcher per nine innings of pitching.
It is undisputed that prior to Kregos’ 1983 form, no form had listed the same nine items collected in his form. It is also undisputed that some but not all of the nine items of information had previously appeared in other forms. In the earlier forms, however, the few items common to Kregos’ form were grouped with items different from those in Kregos’ form. Sie-gel’s 1978 form contained won/lost record (1) and earned run average for the season *703(2), but contained no “at site” information (3, 4, 5) and no information for recent starts (6, 7, 8, 9). It contained only two of Kregos’ nine items (1, 2). Fratas’ 1980 form contained “at site” information for the previous season (differing from Kre-gos’ “at site” information for the current season (3, 4, 5)) and contained no information for recent starts (6, 7, 8, 9). It contained only two of Kregos’ nine items (1, 2). Eckstein’s 1981 form, the only prior form to contain information for recent starts (using two of the four items in Kregos’ form (6, 7)) lacked the third and fourth “recent starts” items (earned run average and men on base average (8, 9)), contained “at site” information only for won/lost record and did not report that data for the current season against the opposing team,2 and lacked earned run average for the season (2). It contained only three of Kregos’ nine items (1, 6, 7).
Kregos’ item (9), men on base average in recent starts, had not previously appeared anywhere. However, a supplier for the Associated Press Syndicate, Inc. had distributed on a weekly basis the men on base average for every pitcher for the entire season, rather than for the most recent three starts, as in Kregos’ form.
District Court decision. The District Court granted summary judgment for the defendants on both Kregos’ copyright and trademark claims. On the copyright side of the case, the Court ruled that Kregos lacked a copyrightable interest in his pitching form on three grounds. First, the Court concluded that Kregos’ pitching form was insufficiently original in its selection of statistics to warrant a copyright as a compilation. Second, the Court concluded that, in view of the limited space available for displaying pitching forms in newspapers, the possible variations in selections of pitching statistics were so limited that the idea of a pitching form had merged into its expression. Third, the Court ruled that Kregos’ pitching form was not entitled to a copyright because of the so-called “blank form” doctrine. On the trademark side of the case, the Court granted summary judgment for the defendants on the ground that Kregos’ trademark claims encountered, as a matter of law, a functionality defense.
Discussion
I. Copyright Claim
A. Copyright for a Compilation of Facts. The basic principles concerning copyright protection for compilations of facts are clear and have recently been authoritatively restated in the Supreme Court’s decision rejecting copyright protection for telephone book white pages. Feist Publications, Inc. v. Rural Telephone Service Co., Inc., — U.S.-, 111 S.Ct. 1282, 113 L.Ed.2d 358 (1991) (“Feist ”). “A factual compilation is eligible for copyright if it features an original selection or arrangement of facts, but the copyright is limited to the particular selection or arrangement. In no event may copyright extend to the facts themselves.” Id. at-, 111 S.Ct. at 1290. “Original, as the term is used in copyright, means only that the work was independently created by the author (as opposed to copied from other works), and that it possesses at least some minimal degree of creativity.” Id. at-, 111 S.Ct. at 1287 (citation omitted).
In principle, this “independent creation” sense of the word “original” applies to the originality required to entitle a selection of factual information to a copyright. “A compiler may settle upon a selection or arrangement that others have used; novelty is not required. Originality requires only that the author make the selection or arrangement independently (i.e., without copying that selection or arrangement from another work)_” Id. at-, 111 S.Ct. at 1294. In practice, however, as the Court in Feist made clear, application of the “originality” standard to compilations of facts is narrowed somewhat by the requirement of “some minimal level of creativity.” Id. In Feist, for example, it would have *704availed the compiler nothing to have persuaded a fact-finder that it came upon the idea of “selecting” all the telephone subscribers in town and “arranging” their names in alphabetical order without copying these features from any preexisting work. The widespread prior existence of white page lists of all telephone subscribers in a community, arranged alphabetically, precluded, as a matter of law, a finding of the requisite creativity. The white pages were deemed “entirely typical” and “garden-variety.” Id. at ——, 111 S.Ct. at 1296. The selection of listings “could not be more obvious.” Id. And “there is nothing remotely creative about arranging names alphabetically in a white pages directory. ... It is not only unoriginal, it is practically inevitable.” Id.
Thus, as to compilations of facts, independent creation as to selection and arrangement will not assure copyright protection; the requirement of minimal creativity becomes an important ingredient of the test for copyright entitlement. “[T]he selection and arrangement of facts cannot be so mechanical or routine as to require no creativity whatsoever.” Id.
Prior to Feist, we had applied these principles to require some minimal level of creativity in two fairly recent cases that illustrate compilations of facts one of which is and one of which is not entitled to a copyright, Eckes v. Card Prices Update, 736 F.2d 859 (2d Cir.1984), and Financial Information, Inc. v. Moody’s Investors Service, 808 F.2d 204 (2d Cir.1986) (“FII”), cert. denied, 484 U.S. 820, 108 S.Ct. 79, 98 L.Ed.2d 42 (1987). In Eckes we upheld a District Court’s finding, made after trial, that a selection of 5,000 out of 18,000 baseball cards to be considered “premium” was entitled to a copyright. Eckes, 736 F.2d at 863. In FII we upheld a District Court’s finding, also made after trial, that the listing of five items of information concerning municipal bond calls lacked sufficient selection to warrant a copyright; in almost all instances, the five items for the various bond issues had all appeared in “tombstone” ads, and only “minor additional research” was needed to complete the listings. FFI, 808 F.2d at 208.
Kregos’ pitching form presents a compilation of facts that falls between the extremes illustrated by Eckes and FFI. Kregos has selected nine items of information concerning a pitcher’s performance. The universe of known facts available only from inspection of box scores of prior games is considerably greater than nine, though perhaps not as great as the quantity of 18,000 cards in Eckes. For example, Kregos could have selected past performances from any number of recent starts, instead of using the three most recent starts. And he could have chosen to include strikeouts, walks, balks, or hit batters. By consulting play-by-play accounts of games, instead of box scores, he could have counted various items such as the number of innings in which the side was retired in order, or in which no runner advanced as far as second base. Or he could have focused on performance under pressure by computing the percentage of innings in which a runner scored out of total innings in which a runner reached second base, and he could have chosen to calculate this statistic for any number of recent starts. In short, there are at least scores of available statistics about pitching performance available to be calculated from the underlying data and therefore thousands of combinations of data that a selector can choose to include in a pitching form.3
It cannot be said as a matter of law that in selecting the nine items for his pitching form out of the universe of available data, Kregos has failed to display enough selectivity to satisfy the requirement of originality. Whether in selecting his combination of nine items he has displayed the requisite degree of creativity is a somewhat closer question. Plainly he has done better than the compiler in FFI who “selected” only the five facts about bond calls already grouped together in nearly all tombstone ads. Judge Goettel *705was persuaded to rule against Kregos, at least in part, because “most of the statistics ... had been established in previously existing forms.” 731 F.Supp. at 118. But that observation is largely irrelevant to the issue of whether Kregos’ selection of statistics displays sufficient creativity to warrant a copyright. Nearly all copyrighted compilations of facts convey facts that have been published elsewhere. Each of the cards selected for the “premium” category in Eckes had previously been published. To hold a valid copyright, a compiler of facts need not be a discoverer of facts. Indeed, any discovered fact, or, in Kregos’ case, any newly devised statistic, would not, in and of itself, be eligible for copyright protection.
The prior publication of some of the statistics on Kregos’ form might indicate, however, that his selection is not sufficiently different from those grouped in earlier publications to satisfy minimal creativity. That conclusion cannot rest on just the prior appearance of the statistics Kregos chose to use, but might rest on the prior appearance of a selection of those statistics that either is identical to Kregos’ selection or varies from his only to a trivial degree. In that event, the first issue would be whether Kregos could demonstrate originality, i.e., persuade the trier that he had not copied from the similar selection. Even if Kregos could satisfy that burden, the issue would then arise as to whether the previously published selections of statistics had reached the point where it could be said that Kregos’ selection was insufficiently creative, or in the words of Feist, “entirely typical,” “garden-variety,” or “obvious.” In view of the variety of pitching forms disclosed in the record, it is unlikely that such a conclusion could be reached and certainly could not be reached as a matter of law.
But these issues are not even likely to arise because the record discloses no prior pitching form with more than three of the pitching performance statistics that are included in Kregos’ selection of nine statistics. There is no prior form that is identical to his nor one from which his varies in only a trivial degree. The validity of his copyright in a compilation of facts cannot be rejected as a matter of law for lack of the requisite originality and creativity.
B. Idea/Expression Merger. . The fundamental copyright principle that only the expression of an idea and not the idea itself is protectable, see Mazer v. Stein, 347 U.S. 201, 217, 74 S.Ct. 460, 470, 98 L.Ed. 630 (1954), has produced a corollary maxim that even expression is not protected in those instances where there is only one or so few ways of expressing an idea that protection of the expression would effectively accord protection to the idea itself. See Educational Testing Services v. Katzman, 793 F.2d 533, 539 (3d Cir.1986); Toro Co. v. R & R Products Co., 787 F.2d 1208, 1212 (8th Cir.1986). Our Circuit has considered this so-called “merger” doctrine in determining whether actionable infringement has occurred, rather than whether a copyright is valid, see Durham Industries, Inc. v. Tomy Corp., 630 F.2d 905, 916 (2d Cir.1980), an approach the Nimmer treatise regards as the “better view.” See 3 Nimmer on Copyright § 13.03[B][3] at 13-58 (1990). Assessing merger in the context of alleged infringement will normally provide a more detailed and realistic basis for evaluating the claim that protection of expression would inevitably accord protection to an idea.
Determining when the idea and its expression have merged is a task requiring considerable care: if the merger doctrine is applied too readily, arguably available alternative forms of expression will be precluded; if applied too sparingly, protection will be accorded to ideas. Recognizing this tension, courts have been cautious in applying the merger doctrine to selections of factual information, see Educational Testing Services, 793 F.2d at 540 (doctrine inapplicable to selection of test questions); Toro Co., 787 F.2d at 1212 (doctrine inapplicable to selection of data for numbering parts), though the doctrine has been applied on occasion to selections of categories of data, see, e.g., Matthew Bender & Co. v. Kluwer Law Book Publishers, Inc., 672 F.Supp. 107, 110 (S.D.N.Y.1987) (categories of data concerning personal injury awards).
*706In one sense, every compilation of facts can be considered to represent a merger of an idea with its expression. Every compiler of facts has the idea that his particular selection of facts is useful. If the compiler’s idea is identified at that low level of abstraction, then the idea would always merge into the compiler’s expression of it. Under that approach, there could never be a copyrightable compilation of facts. However, if the idea is formulated at a level of abstraction above the particular selection of facts the compiler has made, then merger of idea and expression is not automatic. Even with an idea formulated at a somewhat high level of abstraction, circumstances might occur where the realistic availability of differing expressions is so drastically limited that the idea can be said to have merged in its expression.
In this case, Judge Goettel understood Kregos’ idea to be “to publish an outcome predictive pitching form.” 731 F.Supp. at 119. In dissent, Judge Sweet contends that Kregos’ idea is that the nine statistics he has selected are the most significant ones to consider when attempting to predict the outcome of a baseball game. Unquestionably, if that is the idea for purposes of merger analysis, then merger of that idea and its expression has occurred — by definition.
Though there is room for fair debate as to the identification of the pertinent idea whenever merger analysis is applied to a compilation of facts, we think the “idea” in this case is the one as formulated by Judge Goettel. Kregos has not devised a system that he seeks to withdraw from the public domain by virtue of copyright. He does not present his selection of nine statistics as a method of predicting the outcome of baseball games. His idea is that of “an outcome predictive pitching form” in the general sense that it selects the facts that he thinks newspaper readers should consider in making their own predictions of outcomes. He does not purport to weight the nine statistics, much less provide a method for comparing the aggregate value of one pitcher’s statistics against that of the opposing pitcher in order to predict an outcome or even its probability of occurring. He has not devised a system, as had the deviser of a bookkeeping system in Baker v. Selden, 101 U.S. (11 Otto) 99, 25 L.Ed. 841 (1879). He has compiled facts, or at least categories of facts.
Though formulating the idea as “an outcome predictive pitching form,” Judge Goettel applied the merger doctrine, concluding that the idea of selecting outcome predictive statistics to rate pitching performance was capable of expression in only a very limited number of ways. He thought the ease more attracted by Matthew Bender, involving categories of data for personal, injury awards, than by the cases involving categories of statistics for performances of race horses, see Wabash Publishing Co. v. Flanagan, No. 89-C-1923 (N.D.Ill. Apr. 3, 1989) (1989 WL 32939); Triangle Publications, Inc. v. New England Newspaper Publishing Co., 46 F.Supp. 198, 201-02 (D.Mass.1942); see also Triangle Publications, Inc. v. Sports Eye, Inc., 415 F.Supp. 682, 684 (E.D.Pa.1976). In Matthew Bender, Judge Conner concluded that the categories in the plaintiff’s chart (amount, case, plaintiff, event, injury, and relevant data) were “the only sensible ones which could have been used to compile the data.” 672 F.Supp. at 112. That observation may well have been so as to categories of data for personal injury awards, but we think the same cannot be said of categories of pitching statistics.
As the various pitching forms in the record indicate, the past performances of baseball pitchers can be measured by a variety of statistics, as can the past performances of race horses, Kregos’ selection of categories includes three statistics for the pitcher’s current season performance against the day’s opponent at the site of the day’s game; other charts select “at site” performance against the opponent during the prior season, and some select performance against the opponent over the pitcher’s career, both home and away. Some charts include average men on base per nine innings; others do not. The data for most recent starts could include whatever number of games the compiler *707thought pertinent. These variations alone (and there are others) abundantly indicate that there are a sufficient number of ways of expressing the idea of rating pitchers’ performances to preclude a ruling that the idea has merged into its expression.
In reaching this conclusion, we confess to some unease because of the risk that protection of selections of data, or, as in this case, categories of data, have the potential for according protection to ideas. Our concern may be illustrated by an example of a doctor who publishes a list of symptoms that he believes provides a helpful diagnosis of a disease. There might be many combinations of symptoms that others could select for the same purpose, but a substantial question would nonetheless arise as to whether that doctor could obtain a copyright in his list, based on the originality of his selection. If the idea that the doctor is deemed to be expressing is the general idea that the disease in question can be identified by observable symptoms, then the idea might not merge into the doctor’s particular expression of that idea by his selection of symptoms. That general idea might remain capable of many other expressions. But it is arguable that the doctor has conceived a more precise idea— namely, the idea that his selection of symptoms is a useful identifier of the disease. That more limited idea can be expressed only by his selection of symptoms, and therefore might be said to have merged into his expression. Thus, as with the idea/expression dichotomy itself, see Nichols v. Universal Pictures Corp., 45 F.2d 119, 121 (2d Cir.1930) (Judge Learned Hand’s formulation of the “abstractions test”), cert. denied, 282 U.S. 902, 51 S.Ct. 216, 75 L.Ed. 795 (1931), application of the doctrine of an idea merging with its expression depends on the level of abstraction at which the idea is formulated.
As long as selections of facts involve matters of taste and personal opinion, there is no serious risk that withholding the merger doctrine will extend protection to an idea. That was surely the case with the selection of premium baseball cards in Eckes. It is also true of a selection of prominent families for inclusion in a social directory. See Social Register Ass’n v. Murphy, 128 F. 116 (C.C.D.R.I.1904); see also New York Times Co. v. Roxbury Data Interface, Inc., 434 F.Supp. 217, 222 n. 2 (D.N.J.1977). However, where a selection of data is the first step in an analysis that yields a precise result or even a better-than-average probability of some result, protecting the “expression” of the selection would clearly risk protecting the idea of the analysis.
Kregos’ pitching form is part way along the continuum spanning matters of pure taste to matters of predictive analysis. He is doing more than simply saying that he holds the opinion that his nine performance characteristics are the most pertinent. He implies that his selections have some utility in predicting outcomes. On the other hand, he has not gone so far as to provide a system for weighing the combined value of the nine characteristics for each of two opposing pitchers and determining a probability as to which is more likely to win. Like the compilers of horse racing statistics, Kregos has been content to select categories of data that he obviously believes have some predictive power, but has left it to all sports page readers to make their own judgments as to the likely outcomes from the sets of data he has selected. His “idea,” for purposes of the merger doctrine, remains the general idea that statistics can be used to assess pitching performance rather than the precise idea that his selection yields a determinable probability of outcome.4 Since there are various ways of expressing that general idea, the merger doctrine need not be applied to assure that the idea will remain in the public domain.
*708C. “Blank Form” Doctrine. The District Court also ruled that Kregos could not obtain a valid copyright in his pitching form because of the so-called “blank form” doctrine. The doctrine derives from the Supreme Court’s decision in Baker v. Selden, supra. The Court there denied copyright protection to blank`forms contained in a book explaining a system of double-entry bookkeeping. The forms displayed an arrangement of columns and headings that permitted entries for a day, a week, or a month to be recorded on one page or two facing pages. The Court made clear that the author could not obtain copyright protection for an “art” that “might or might not have been patented” and reasoned that since the “art” was available to the public, “the ruled lines and headings of accounts must necessarily be used as incident to it.” Id. at 104. Then, in a concluding statement that is susceptible to overreading, the Court said that “blank account-books are not the subject of copyright.” Id. at 107.
Though there are some statements suggesting broadly that no blank forms are copyrightable, see Bibbero Systems, Inc. v. Colwell Systems, Inc., 893 F.2d 1104, 1106-07 (9th Cir.1990), amended, reh’g denied, 1990 WL 1285, 1990 U.S.App.LEXIS 2562; M.M. Business Forms Corp. v. UARCO, Inc., 472 F.2d 1137, 1139 (6th Cir.1973), many courts have recognized that there can be protectable elements of forms that include considerable blank space.5 See, e.g., Harcourt, Brace & World, Inc. v. Graphic Controls Corp., 329 F.Supp. 517, 524 (S.D.N.Y.1971) (answer sheets); Norton Printing Co. v. Augustana Hospital, 155 U.S.P.Q. 133 (N.D.Ill.1967) (laboratory test forms). See generally Whelan Associates, Inc. v. Jaslow Dental Laboratory, Inc., 797 F.2d 1222, 1242-43 (3d Cir.1986), cert. denied, 479 U.S. 1031, 107 S.Ct. 877, 93 L.Ed.2d 831 (1987).
The regulations of the Copyright Office are careful to preclude copyright registration to:
Blank forms, such as ... account books, diaries, bank checks, scorecards, address books, report forms, order forms and the like, which are designed for recording information and do not in themselves convey information;
37 C.F.R. § 202.1(c) (1990) (emphasis added).
Of course, a form that conveys no information and serves only to provide blank space for recording information contains no expression or selection of information that could possibly warrant copyright protection. See, e.g., John H. Harland Co. v. Clarke Checks, Inc., 711 F.2d 966, 971-72 (11th Cir.1983) (check stubs). At the same time, it should be equally obvious that a writing that does contain a selection of categories of information worth recording, sufficiently original and creative to deserve a copyright as a compilation of facts, cannot lose that protection simply because the work also contains blank space for recording the information. When the Copyright Office denies a copyright to scorecards or diaries that “do not in themselves convey information,” it must be contemplating works with headings so obvious that their selection cannot be said to satisfy even minimal creativity (a baseball scorecard with columns headed “innings” and lines headed “players”; a travel diary with headings for “cities” “hotels,” and “restaurants”). Such a work conveys no information, not just because it contains blanks, but because its selection of headings is totally uninformative. On the other hand, if a scorecard or diary contained a group of headings whose selection (or possibly arrangement) displayed cognizable creativity, the author’s choice of those headings would convey to users the information that this group of categories was something out of *709the ordinary. See 1 Nimmer on Copyright § 2.18[C][1] at 2-201 (1990) (“Thus books intended to record the events of baby’s first year, or a record of a European trip, or any one of a number of other subjects, may evince considerable originality in suggestions of specific items of information which are to be recorded, and in the arrangement of such items.”) (emphasis added; footnote omitted).
The Ninth Circuit has rejected this approach, believing that the rule of cases like Norton Printing “is potentially limitless.” Bibbero Systems, 893 F.2d at 1107. In the Ninth Circuit’s view, “[a]ll forms seek only certain information, and, by their selection, convey that the information sought is important.” Id. With deference, we suggest that this critique of Norton Printing and other cases recognizing a copyright in the selection of categories of information for forms is not well taken. All forms may convey that the information called for is important (or at least worth recording), but the form-maker does not necessarily display even minimal creativity by selecting categories of “important” information. The principle of cases like Norton Printing is not limitless because courts are obliged to determine as to forms, as with all compilations of information, whether the author’s selection of categories of data to be recorded displays at least minimal creativity. The check stub in Clarke Checks was plainly deficient in this regard. Whether the forms in Bibbero Systems and Norton Printing displayed the requisite creativity are matters of fair dispute. But all forms need not be denied protection simply because many of them fail to display sufficient creativity.
In the pending case, once it is determined that Kregos’ selection of categories of statistics displays sufficient creativity to preclude a ruling as a matter of law that it is not a copyrightable compilation of information, that same conclusion precludes rejecting his copyright as a “blank form.” 6
D. Extent of Protection. Our ruling that Kregos’ copyright claim survives defendants’ motion for summary judgment does not, of course, mean that he will necessarily obtain much of a victory. “Even if a work qualifies as a copyrightable compilation, it receives only limited protection .... [Cjopyright protects only the elements that owe their origin to the compiler — the selection, coordination, and arrangement of facts.” Feist, — U.S. at --, 111 S.Ct. at 1294. If Kregos prevails at trial on the factual issues of originality and creativity, he will be entitled to protection only against infringement of the pro-tectable features of his form. Only the selection of statistics might be entitled to protection. We agree entirely with Judge Goettel that nothing in Kregos’ arrangement of the selected statistics displays the requisite creativity. As to the arrangement, Kregos’ form is surely a “garden-variety” pitching form. The statistics are organized into the “obvious” arrangement of columns, and the form follows the pattern of most other forms: the statistics are organized into three groups, first the statistics about each pitcher’s performance for the season, then the statistics about the pitcher’s performance against the day’s opponent, and finally the statistics concerning the pitcher’s recent starts.
Even as to the selection of statistics, if Kregos establishes entitlement to protection, he will prevail only against other forms that can be said to copy his selection. That would appear to be true of the AP’s 1984 form, which, as Judge Goettel noted, is “identical in virtually every sense to plaintiff’s form.” 731 F.Supp. at 115. Whether it is also true of the AP’s current form, revised in 1986, is far less certain. That form contains six of Kregos’ nine items (1, 2, 6, 7, 8, 9). It also includes four items that Kregos does not have. Three of these items concern performance against the day’s opposing team — won-lost record, innings pitched, and earned run average; though these three statistics appear on Kregos’ form, the AP’s 1986 form shows data for the current season both home and away, whereas Kregos’ form shows data *710for the pitcher’s current season at the site of that day’s game. The fourth item on the AP’s 1986 form and not on Kregos’ form shows the team’s record in games started by that day’s pitcher during the season.
The reason for doubting that the AP’s 1986 form infringes Kregos’ form arises from the same consideration that supports Kregos’ claim to a copyright. Kregos can obtain a copyright by displaying the requisite creativity in his selection of statistics. But if someone else displays the requisite creativity by making a selection that differs in more than a trivial degree, Kregos cannot complain. Kregos contends that the AP’s 1986 form makes insignificant changes from its 1984 form.7 But Kregos cannot have it both ways. If his decision to select, in the category of performance against the opposing team, statistics for the pitcher’s current season at the site of today’s game displays, in combination with his other selections, enough creativity to merit copyright protection, then a competitor’s decision to select in that same category performance statistics for the pitcher’s season performance both home and away may well insulate the competitor from a claim of infringement. Thus, though issues remain to be explored before any determination can be made, it may well be that Kregos will have a valid claim only as to the AP’s 1984 form.8
II. Trademark Claims
A. Infringement. In his complaint Kregos alleged that his form “has become recognized as an indication of source” and further alleged that the defendants’ forms are likely to cause confusion as to origin, in violation of section 43(a) of the Lanham Act, 15 U.S.C. § 1125(a) (1988). As the District Court recognized, Kregos appears to be claiming infringement of trade dress, 731 F.Supp. at 121 n. 17. The Court rejected the claim on the ground that a functionality defense was established as a matter of law. We prefer to affirm on the alternate ground suggested by the District Court, id. at 122 n. 18, that Kregos has totally failed to create a triable issue of fact as to the existence of secondary meaning in the look or appearance of his form. See Stormy Clime Ltd. v. Progroup, Inc., 809 F.2d 971, 974 (2d Cir.1987) (secondary meaning required for trade dress infringement claim).
B. False Copyright Notice. The District Court also entertained Kregos’ claim of a Lanham Act violation against Sports Features by reason of a copyright notice affixed to its form supplied to AP. Though the claim was not included in the complaint, Judge Goettel indicated in a supplemental oral ruling that he would have permitted an amendment to set forth the claim, but he rejected it on the ground that Kregos lacked an interest in his form pro-tectable under the Lanham Act and for that reason could have no valid claim under the Act. This analysis, though not necessarily the ultimate conclusion, was incorrect. The fact that a proprietor fails to show sufficient secondary meaning in a mark to establish an infringement claim does not preclude his assertion of some other Lanham Act claim such as false designation of origin or false description. Kregos could complain, for example, if Sports Features unfairly competed with him by falsely claiming that its form was originated by some well-known baseball player.
It is not clear from Kregos’ papers in what respect he contends that the Sports *711Features copyright notice is false. The notice makes no claim that the form originated from any source other than Sports Features. Nor does the notice contain any representation that is false, cf. Eden Toys, Inc. v. Florelee Undergarment Co., 697 F.2d 27, 37 (2d Cir.1982) (copyright notice falsely stating “Original” violates Lanham Act). If Kregos is simply saying that the notice is false in claiming creation by Sports Features in that the form is, under the copyright law, an infringement of Kre-gos’ form, we reject his attempt to convert all copyright claims into Lanham Act violations. The Lanham Act claims were properly rejected.
Conclusion
The judgment of the District Court is reversed and remanded with respect to the copyright claim, and affirmed with respect to the trademark claims.
. There is some uncertainty as to the time period to which the data in this category pertain. In his deposition, Kregos says the data are "year to date,” (J.A. 298), and the numbers in the won/lost column are too low to be career figures (it is not likely that by 1984 Steve Carlton had pitched at Philadelphia only 8 innings against Cincinnati in his career, see 731 F.Supp. at 122, appx. 1; the numbers more likely reflect one season). Yet the numbers may reflect season performance for the previous season, at least for games played early in the season, before current season data is significant. For example, Larry Gura is shown as having a won/lost record of 4-0 against all opponents in t984 and a won/lost record of 0-2 against the day's opponent at the site of the day’s game. Obviously, his "at site” numbers cannot be for the 1984 season, but might pertain to the previous season. An example of Kregos’ earlier form, printed on June 11, 1980, shows "1979" as the heading, for "at site” data. (J.A. 76). Judge Goettel characterized Kregos’ “at site” data as pertaining to the pitcher's performance "during his career against the scheduled opponent.” 731 F.Supp. at 114. Since the matter has not been resolved by fact-finding at trial, we will accept, for purposes of this appeal, Kregos’ deposition statement that the "at site” data pertain to the current season.
. It is not clear from the record whether Eck-stein's "at site" data reflects won/lost record against the day’s opponent for the previous season or against all opponents during the current season. In any event the won/lost records Eck-stein includes are not the "at site” results against the day’s opponent during the current season, which is the data Kregos includes.
. If the universe of available data included even 20 items and a selector was limited to 9 items, there would be 167,960 combinations of items available.
. The merger doctrine might be applicable even in some situations where a precise outcome, or the precise probability of an outcome, has not been determined. The doctor whose list of symptoms provides a reliable basis for predicting that a disease is more likely to be present than in the population generally might well be precluded from copyright protection in his selection, even though he makes no estimate as to the degree of predictive power of his list of symptoms.
. We are concerned with protectable elements in the selection (and perhaps arrangement) of the categories of information to be recorded on the forms. There is widespread agreement that a work containing a blank form may be copyrightable because of the protectable elements of the textual matter accompanying the form. See Bibbero Systems, 893 F.2d at 1107; Edwin K. Williams & Co. v. Edwin K. Williams & Co.—East, 542 F.2d 1053, 1061 (9th Cir.1976), cert. denied, 433 U.S. 908, 97 S.Ct. 2973, 53 L.Ed.2d 1092 (1977); Continental Casualty Co. v. Beardsley, 253 F.2d 702, 704 (2d Cir.), cert. denied, 358 U.S. 816, 79 S.Ct. 25, 3 L.Ed.2d 58 (1958).
. We agree with Judge Goettel that whether the form is copyrightable is unaffected by the fact that the form is to be used by newspapers, rather than by newspaper purchasers.
. The District Court described the AP’s 1986 form as “the same basic format as its earlier form with some modifications." 731 F.Supp. at 115. In a footnote, the Court referred to minor changes in headings as "cosmetic distinctions." Id. at 116 n. 8. As an example, the Court referred to the change in a column heading from "pitchers" to “probable pitchers.” Though Kre-gos contends that the Court considered all of the changes from the AP’s 1984 form to its 1986 form to be "cosmetic,” Brief for Appellant at 10, we think Judge Goettel was careful to limit this adjective to the inconsequential variations in headings, and was not undertaking to assess whether the substantive changes in the statistics selected for inclusion in the 1986 form would render it non-infringing as to Kregos’ form, in the event that Kregos’ copyright was valid.
. At this stage of the case, we make no determination concerning the possible effect of the statute of limitations.