Following several promotion denials, appellant, a mechanic, sued the Washington Metropolitan Area Transit Authority, alleging that it had violated Title VII by discriminating against him on the basis of national origin and by retaliating against him for engaging in protected activities. The district court granted summary judgment for WMATA. Because we believe that a reasonable jury could find in appellant's favor with regard to one of his claims, we reverse the summary judgment ruling on that count and remand that portion of the case for further proceedings.

Appellant Luis Salazar, a Peruvian-born Latino, began working for WMATA in 1982 as a bus cleaner. Over the next six years, he worked his way up through several promotions to "Mechanic AA" for Heating, Ventilation, and Air Conditioning. During these and later years, Salazar took numerous night classes in technical skill subjects and pursued a B.A. at the University of the District of Columbia.

As a Mechanic AA, Salazar held the highest nonsupervisory ranking available at WMATA and occasionally served as acting supervisor at his post in Greenbelt, Maryland. Seeking to advance further, he applied five times from 1992 to 1999 for promotions to entry-level supervisory positions. The application process for each position had the same general structure. WMATA first screened out applicants who failed to meet certain minimum requirements. It then selected a panel of supervisors to interview the remaining candidates. The chairperson developed a list of interview questions and assigned each one a weight. At the interviews, some of which were attended by an observer from WMATA's Office of Civil Rights, each panel member graded each candidate on his or her response to each question. After tallying the scores, the panel recommended that the appointment go to the candidate with the highest total points.

Salazar's first four promotion applications met with failure. The fifth time — the one chiefly at issue in this case — Salazar applied for the position of Craft Supervisor in General Equipment at Metro Center. That position required certain mechanical knowledge, including skill at "trouble shooting the electrical/mechanical systems located at bus and rail facilities, and operating/repairing the smoke ventilation fans, drainage pumping stations, sewage ejectors and pneumatic damper sections." When Salazar applied for this position, he contacted Charles Thomas, the Deputy General Manager at Metro, and asked him to ensure that Gary Lewis, the Superintendent for Plant Equipment Maintenance, would not select the members of the interview panel. According to Salazar, Lewis, who had selected the panel members for at least some of Salazar's prior promotion denials, discriminated against Latinos. Indeed, Salazar had filed at least one grievance accusing Lewis of supporting a racially discriminatory supervisor. In his affidavit, Salazar states that he "explained to Mr. Thomas how each time Gary Lewis selected the panel I was ... denied the promotion because Mr. Lewis would `stack' the panel with his friends. These friends, like Mr. Lewis, were discriminatory against Latino people, like myself." Responding sympathetically, Thomas, according to Salazar's affidavit, "selected the panel, which included only three persons," all of whom Salazar acknowledged were not Lewis's friends and not likely to be discriminatory.

Salazar and five other applicants met the minimum qualifications and advanced to the interview round. Of these applicants, Salazar had the most seniority by several years.

Expecting an interview with the three men selected by Thomas, Salazar was surprised to find a fourth man, Buddy Jaggie, serving as chair. Salazar's surprise stemmed not only from Thomas's promise of just three panelists, but also from the fact that "all the panels I had interviewed with in the past years had been made up of three members, not four." More significantly, Jaggie, who held the post of Assistant Superintendent for Plant Maintenance, was Lewis's assistant as well as his close friend. Salazar distrusted Jaggie not just because of Jaggie's relationship with Lewis but also because, while working towards the Mechanic AA position years ago, Salazar had consistently failed a test administered by Jaggie, passing only after filing a grievance to obtain outside review.

Jaggie acknowledged that he was "probably" appointed to the panel by Lewis. He explained that before the interview, he "made up [the] questions" and assigned a point value to each question. He consulted with Lewis in determining these weights. Ultimately, Jaggie developed 13 questions worth a total of 190 points: 12 questions calling for a spoken answer (8 worth 10 points and 4 worth 20 points) and 1 question (worth 30 points) requiring a written answer. Only 2 questions — each worth 10 points and thus amounting to less than a ninth of the total — directly addressed the candidates' experience and education. Other questions posed hypothetical scenarios (3 questions worth 50 total points), inquired about Metro policies and their implementation (3 questions worth 30 total points), called for technical responses (4 questions worth 80 total points), and probed the candidates' motivation levels (1 question worth 10 points). Jaggie also drafted model answers for the panelists to use during the interviews.

At the interviews, the four panelists asked Jaggie's questions and scored the six candidates. On the two experience-related questions, Salazar scored above all other candidates, but overall he came in fourth. Jaggie and two other panelists gave Salazar mediocre scores, while the remaining panelist scored Salazar better than all other candidates. Had Jaggie's scores not counted, Salazar would still have finished fourth overall. WMATA's observer thought "[t]he interviews were conducted in a fair and nondiscriminatory manner."

According to Salazar — and WMATA offers no evidence to the contrary — WMATA did not assign the successful candidate, Timothy Tucker, to the Metro Center supervisory "position for which he was selected." Instead, Salazar states, "Gary Lewis moved Mr. Tucker to Greenbelt to work as a support equipment supervisor. . . . Working at Greenbelt required less responsibility than in Metro Center."

After exhausting the EEO process, Salazar sued WMATA, alleging discrimination and retaliation in violation of Title VII of the Civil Rights Act of 1964, 42 U.S.C. §§ 2000e-2, 2000e-3, with regard to all five promotion denials. The district court found his claims related to the first three promotion denials to be procedurally barred, and it granted WMATA's motion for summary judgment on the claims related to Salazar's last two promotion denials — the one described above for the position in General Equipment at Metro Center and one earlier that spring for a position in Metro's shop in Alexandria. Specifically, the court held that Salazar could not show that WMATA's asserted reason for refusing to promote him was pretextual. Salazar v. Wash. Metro. Area Transit Auth., No. 99-1631, slip op. at 5 (D.D.C. Oct. 30, 2003). Salazar now appeals.

We review the district court's grant of summary judgment de novo, viewing the evidence in the light most favorable to Salazar and drawing all reasonable inferences accordingly. E.g., Dunaway v. Int'l Bhd. of Teamsters, 310 F.3d 758, 761 (D.C.Cir.2002). We will affirm only if no reasonable jury could find in his favor. Id.

The district court granted summary judgment to WMATA on four counts: Salazar's discrimination claims related to the Alexandria and Metro Center positions and his parallel retaliation claims. As an initial matter, we find that Salazar has preserved only his discrimination claim regarding the Metro Center position. His briefs address none of the other three claims, and while his retaliation claim as to the Metro Center position undoubtedly has much in common with his discrimination claim, he never refutes WMATA's argument that he has failed to prove a prima facie case for retaliation. Salazar has thus waived these three claims. See, e.g., Ark Las Vegas Rest. Corp. v. NLRB, 334 F.3d 99, 108 n. 4 (D.C.Cir.2003) (observing that an argument not raised in briefs is waived).

We view Salazar's remaining claim through the framework established in McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S. Ct. 1817, 36 L. Ed. 2d 668 (1973). To survive summary judgment, Salazar must first show that he satisfies the four elements of the prima facie case for a discrimination claim. WMATA concedes that Salazar has made this showing: as a foreign-born Latino, he is a member of a protected class; he applied for the promotions; he had the minimum qualifications needed; and he lost out to a non-Latino. Next, if the employer "produce[s] admissible evidence that, if believed, would establish that [its] action was motivated by a legitimate, nondiscriminatory reason," the plaintiff must show that a reasonable jury could nonetheless infer discrimination. Teneyck v. Omni Shoreham Hotel, 365 F.3d 1139, 1151 (D.C.Cir.2004). In this case, WMATA claims that it did not promote Salazar because "[m]ore qualified candidates were selected for the position through a fairly administered selection process." Appellee's Br. at 13. For purposes of this appeal, the question then becomes whether a reasonable jury could find in Salazar's favor based on all the evidence, including "(1) the plaintiff's prima facie case; (2) any evidence the plaintiff presents to attack the employer's proffered explanation for its actions; and (3) any further evidence of discrimination that may be available to the plaintiff (such as independent evidence of discriminatory statements or attitudes on the part of the employer)." Waterhouse v. District of Columbia, 298 F.3d 989, 992-93 (D.C.Cir.2002) (quoting Aka v. Wash. Hosp. Ctr., 156 F.3d 1284, 1289 (D.C.Cir.1998) (en banc)) (internal quotation marks omitted).

In claiming that it rejected Salazar through a fair process, WMATA relies on our decision in Fischbach v. District of Columbia Department of Corrections, 86 F.3d 1180 (D.C.Cir.1996), which found no inherent discrimination in a job application process structured similarly to WMATA's. We held that where the employer had followed its usual procedure, there was "nothing the least bit fishy about the interviewers' giving slightly less emphasis to the applicants' credentials than to the manner in which each candidate proposed to do the job — especially when one considers that they had the benefit of a prior determination that all of the interviewees were qualified." Id. at 1184. We reversed the district court's ruling in favor of the plaintiff, since that court had based its conclusion on "its own opinion that [the plaintiff] was more qualified than the [successful candidate]." Id.

Unlike the plaintiff in Fischbach, Salazar does not allege that WMATA's selection process is discriminatory simply because it may not always end up with the best-qualified candidate. Indeed, Salazar never claims that WMATA's general process is discriminatory. His challenge turns instead on the specific process used by WMATA in selecting a candidate for the Metro Center position. Although we find the fact that Salazar had a four- rather than three-member panel not all that probative — particularly since WMATA had deviated from its normal appointment process in response to Salazar's concerns — we are more troubled by Jaggie's role as chair. Salazar alleges that after he told Thomas that Lewis and his panel appointees discriminated against Latinos, Thomas promised Salazar a panel that Lewis would have no hand in selecting. Yet Lewis ended up appointing Jaggie as the panel's chair and even helped determine the weights of the questions. We agree with Salazar that a jury could infer something "fishy" from the fact that Lewis placed himself squarely at the center of a process designed to exclude him. Specifically, a jury could conclude that WMATA failed to provide a "fairly administered selection process" and that its claim to the contrary is pretextual. Cf. Lathram v. Snow, 336 F.3d 1085, 1093-94 (D.C.Cir.2003) (holding that a jury could draw an inference of discrimination where an agency departed from its normal process without justification); Johnson v. Lehman, 679 F.2d 918, 922 (D.C.Cir.1982) (noting that although "a finding of a failure on the part of the prospective employer to follow its own regulations and procedures, alone, may not be sufficient to support a finding of . . . discrimination," such a failure "is a factor that the trier of fact may deem probative . . . in determining the true motivation behind the hiring decision of the prospective employer").

To be sure, the scores Jaggie gave Salazar differed little from those given by two other panelists whom Salazar himself acknowledges would be unlikely to discriminate. It is also true that Salazar would have finished fourth even had Jaggie's scores not counted, and no evidence suggests that Jaggie attempted to influence the other panelists' scores during the interview. As Salazar points out, however, Jaggie's role in the process went beyond the specific scores he gave Salazar. Jaggie "not only developed the questions but assigned the relative value for each," Appellant's Br. at 20, and had apparently unfettered discretion in doing so. Jaggie consulted with Lewis as to the weights of the questions — thus extending Lewis's influence beyond Jaggie's appointment — and they developed an interview format that by design assigned only marginal value to candidates' experience and education, which were Salazar's particular strengths.

The possibility that the interview process for the Metro Center job may not have been fairly designed increases in light of the fact that Tucker, the successful candidate, never held that job. Instead, Lewis transferred Tucker to what Salazar described as a less difficult job in Greenbelt — a characterization not contested by WMATA. From this, we think a reasonable jury could infer that Tucker was unsuited for the Metro Center job and that the selection process was geared not to finding the best person for the position, but rather to keeping Salazar from advancing.

In sum, though it is a close call, when we view the evidence in the light most favorable to Salazar and make all reasonable inferences in his favor, as we must at this stage of the litigation, see Dunaway, 310 F.3d at 761, we think a reasonable jury could find pretextual WMATA's assertion that it employed a fairly administered selection process with regard to the Metro Center job. The jury could base this determination on Lewis's unexplained participation — despite Thomas's assurances — that in turn led to the appointment of Jaggie and the development of the interview agenda, including the weights of the questions, along with Tucker's subsequent move to Greenbelt.

The dissent disagrees with our conclusion that a reasonable jury could infer that WMATA's claim of a fairly administered selection process was pretextual. In reaching its own conclusion to the contrary, the dissent both ignores this court's obligation to draw reasonable inferences in Salazar's favor and relies on arguments never made by WMATA.

First, the dissent doubts that a reasonable jury could find anything suspect about the fact that Tucker moved to a less rigorous job rather than ending up in the Metro Center position, speculating that "the innocent reasons why a winning competitor might take another job are legion — personal convenience, a better fit with his skills, a better match with fellow workers, etc." Dissenting op. at 513. Although we agree that a jury may draw such inferences, nothing in the record suggests it must, and at this stage in the process we are required to view the evidence in Salazar's favor, not WMATA's, see Dunaway, 310 F.3d at 761. The dissent's list of "innocent reasons" has another problem: WMATA never offers any, much less the ones suggested by the dissent. Indeed, WMATA's brief never mentions Tucker's transfer at all. Attempting to neutralize the implications of this transfer, the dissent abandons our general rule of requiring parties to make their own arguments. See, e.g., Williams v. United States, 396 F.3d 412, 415 (D.C.Cir.2005) (noting that we find forfeited any arguments not raised in briefs).

Second, the dissent points out that a "departure from [a] departure" from existing practice could be unsuspicious. See dissenting op. at 513. Once again, we see no reason to assume that a reasonable jury must make that inference; a jury could instead see this fact as undermining WMATA's assertion that it provided a fair and neutral selection process. This is particularly true since WMATA neither disputes the substance of Thomas's promise to Salazar nor explains why, in light of that promise, Lewis nonetheless ended up appointing the chairperson and helping him determine weights.

Lastly, the dissent reasons that WMATA's given explanation — that it had a fairly administered selection process — must be true because, as the dissent sees it, Salazar would have lost under virtually any circumstances. See id. at 513-15. We are not as certain. The parties' arguments on this front are piecemeal. Salazar claims that "because Mr. Jaggie assigned the weights to the questions asked by the panel, he was able to minimize Mr. Salazar's strengths, which included his vastly superior experience and training." Appellant's Br. at 20. Attempting to neutralize Salazar's claim that Jaggie tainted the process, WMATA points out — correctly — that Salazar would not have won even had Jaggie's scores and the technical questions been eliminated. Appellee's Br. at 17. Salazar responds less specifically, but also accurately, that by "weighing less significant questions more and by minimizing the value of the important questions regarding experience, education and training, the selectee came up with a higher score," and he calls into question "the validity of those values, how significant [sic] items were weighted heavily and significant ones given less weight and the reasonableness of the method of obtaining them." Reply Br. at 2. Although a reasonable jury could well embrace WMATA's position that Jaggie arranged a fair interview process based on his conduct at the interview and the content of the interview questions, we think it also possible that a jury could infer, in line with Salazar's arguments, that Jaggie and Lewis selected an interview agenda which, though facially acceptable, was designed to downplay Salazar's strengths.

To reach its conclusion that no reasonable jury could think that Jaggie's interview schema disadvantaged Salazar, the dissent substitutes its own arguments for those made by WMATA. The dissent relies on the fact that "[i]f all questions had been weighted equally, Salazar would still have finished in fourth place," dissenting op. at 514 — a point never made by WMATA. Moreover, not only does WMATA never suggest that equal weights are its norm, but the dissent gives no reason, nor can we think of one, to assume that it should be. The dissent further argues that we must discount the value of a "scoring error" in Salazar's favor, id., that WMATA itself neither claimed was an error nor sought to change on the "Corrected Scoring Matrix" it submitted to us. In relying on these points the dissent must not only ignore our doctrine of waiver, see, e.g., Williams, 396 F.3d at 415, but also assume that a reasonable jury must reach the same conclusions without WMATA even raising the issues. Cf. Carducci v. Regan, 714 F.2d 171, 177 (D.C.Cir.1983) (noting that the "premise of our adversarial system is that appellate courts do not sit as self-directed boards of legal inquiry and research, but essentially as arbiters of legal questions presented and argued by the parties before them").

We further note that while the dissent is willing to make unraised arguments favoring WMATA, it makes no similar effort on Salazar's behalf. The dissent never considers, for example, whether Jaggie's unexpected presence at the interview may have caused Salazar to underperform, although the record contains evidence that might support this. Nor does the dissent point out that Salazar's argument that Jaggie weighted the questions to disadvantage him finds support even under the dissent's evenly weighted scheme. Excluding Jaggie's scores but using Jaggie's and Lewis's weights, Salazar's average score was 129, only 86.6% of Tucker's 149, 87.8% of the second-best candidate's 147, and 91.1% of the third-best candidate's 142. With even weights, Salazar comes significantly closer: with an average score of 94, he has 92.1% of Tucker's 102, 96.6% of the second-best candidate's 98, and 99.1% of the third-best candidate's 95. And by tweaking the equal weights, there are many possible scenarios under which Salazar can win. He would win by tripling the two experience-and-education-related questions, by doubling those scores and the scores of two other questions, or by many other permutations. Even assuming, as the dissent does — but not WMATA — that a jury must conclude that one scorer overscored Salazar and that it must therefore discount this score, Salazar can still win by less drastic methods than the dissent implies, see dissenting op. at 514-15: he could win by halving two questions and tripling two others, by tripling four questions, by dropping three questions, or by various other scenarios. We pose these possibilities only to illustrate the problems implicit in abandoning our "salutary rule," see id. at 515, of sticking to the parties' arguments rather than striking out on our own.

Having failed — through unraised arguments and inferences that favor WMATA — to find anything probative in any of Salazar's arguments, the dissent has no need to consider their cumulative effect. But we do, and considering Salazar's evidence all together and giving him the benefit of every reasonable inference, we think a reasonable jury could find pretext.

Finally, we must consider whether a jury which found that WMATA rigged the process to keep Salazar from getting the job could further conclude that WMATA did so because of Salazar's national origin. In Reeves v. Sanderson Plumbing Products, Inc., 530 U.S. 133, 120 S. Ct. 2097, 147 L. Ed. 2d 105 (2000), the Supreme Court held that "[i]n appropriate circumstances, the trier of fact can reasonably infer from the falsity of the explanation that the employer is dissembling to cover up a discriminatory purpose," particularly since "the employer is in the best position to put forth the actual reason for its decision." Id. at 147, 120 S. Ct. 2097; see also Aka, 156 F.3d at 1293 (noting that "[t]he jury can conclude that an employer who fabricates a false explanation has something to hide; that `something' may well be discriminatory intent"). A reasonable jury may not infer discrimination, however, "if the record conclusively revealed some other, nondiscriminatory reason for the employer's decision, or if the plaintiff created only a weak issue of fact as to whether the employer's reason was untrue and there was abundant and uncontroverted independent evidence that no discrimination had occurred." Reeves, 530 U.S. at 148, 120 S. Ct. 2097.

Here, WMATA offers no explanation for its hiring decision other than its "fair process" argument; indeed, WMATA gives no explanation for how, given Thomas's promise to Salazar, Lewis came to appoint the panel chair. And while the presence of the other panelists and the near parity between their scoring and Jaggie's scoring, see supra at 509, weakens the likelihood that a jury could find pretext, it does not amount to "abundant and uncontroverted independent evidence" that Lewis and Jaggie did not discriminate. Under Reeves, then, we believe a jury might infer discrimination if it concluded that WMATA's proffered reason was pretextual. See 530 U.S. at 147-49, 120 S. Ct. 2097; Aka, 156 F.3d at 1290-92; cf. Desert Palace, Inc. v. Costa, 539 U.S. 90, 99-100, 123 S. Ct. 2148, 156 L. Ed. 2d 84 (2003).

Because given this evidence we would not disturb a jury verdict in Salazar's favor, we cannot uphold a summary judgment order concluding otherwise. Reeves, 530 U.S. at 150, 120 S. Ct. 2097 (noting that "the standard for granting summary judgment mirrors the standard for judgment as a matter of law, such that the inquiry under each is the same") (internal quotation marks omitted). We reverse and remand with regard to Salazar's discrimination claim relating to the Metro Center job. In all other respects, we affirm.

So ordered.

STEPHEN F. WILLIAMS, Senior Circuit Judge, dissenting.

The majority concludes that a reasonable jury could find in appellant's favor with regard to one of his claims. Maj. Op. at 505-06. Like the majority, I view the evidence as a "close call," Maj. Op. at 509, but I believe no reasonable jury could make the required finding.

In Reeves v. Sanderson Plumbing Products, Inc., 530 U.S. 133, 120 S. Ct. 2097, 147 L. Ed. 2d 105 (2000), the Supreme Court stated that

an employer would be entitled to judgment as a matter of law if the record conclusively revealed some other, nondiscriminatory reason for the employer's decision, or if the plaintiff created only a weak issue of fact as to whether the employer's reason was untrue and there was abundant and uncontroverted independent evidence that no discrimination had occurred.

Id. at 148, 120 S. Ct. 2097; see also Aka v. Washington Hospital Center, 156 F.3d 1284, 1291-92 (D.C.Cir.1998) (en banc). This is just a linguistic variant of the standard rule that the party with the burden of persuasion cannot defeat summary judgment by offering only a "scintilla" of proof. See, e.g., Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 252, 106 S. Ct. 2505, 91 L. Ed. 2d 202 (1986); Taylor v. Small, 350 F.3d 1286, 1295 (D.C.Cir.2003). In my view, Salazar's evidence, viewed in the light most favorable to him, yields just such a "weak issue of fact," no more than a scintilla.

Salazar came in fourth in a six-person competition for promotion to the position of Craft Supervisor in General Equipment at WMATA's Metro Center station. If I understand the majority, it believes that three features of the selection process could in the aggregate enable a reasonable jury to find that WMATA had discriminated: first, the fact that Timothy Tucker, number one in the competition, ended up not taking the Metro Center position, but instead took what Salazar said was a "less difficult" one in Greenbelt; second, the role of two WMATA higher-ups, Gary Lewis and Buddy Jaggie, in the competitive process; third, and related to the second, the nature of the questions posed in the competition and the weighting of questions.

The majority regards the fact that the selected applicant (Tucker) apparently never held the Metro Center job as increasing "the possibility that the interview process ... may not have been fairly designed." Maj. Op. at 509. There is in fact an evidentiary vacuum as to what happened after the competition (other than Tucker's going to Greenbelt). Technically, this is a gap in the plaintiff's prima facie case, in which the fourth element is evidence that "the position remained open [after the plaintiff lost out] and the employer continued to seek applicants from persons of [the plaintiff's] qualifications." McDonnell Douglas Corporation v. Green, 411 U.S. 792, 802, 93 S. Ct. 1817, 36 L. Ed. 2d 668 (1973). In Teneyck v. Omni Shoreham Hotel, 365 F.3d 1139 (D.C.Cir.2004), we held that plaintiff's failure to show the sequel to her rejection was fatal to her prima facie case. Id. at 1151-53. WMATA here waived any argument as to inadequacies in the prima facie case, but it is still ironic that Salazar has managed to transform a similar gap into a special point in his favor. The transformation is especially odd because the innocent reasons why a winning competitor might take another job are legion — personal convenience, a better fit with his skills, a better match with fellow workers, etc. Even winners change their minds.

The majority thinks that a reasonable jury "could infer something `fishy'" from Lewis's appointment of Jaggie, his assistant, as panel chair. See Maj. Op. at 509. The majority sees this as problematic first as a breach of Thomas's promise that the search process would be Lewis-free, and second as leading to interview characteristics that the majority believes might reasonably be thought discriminatory. I am unable to identify a scintilla of evidence that Jaggie's participation disadvantaged Salazar.

To give weight to the broken promise, the majority cites a line of cases holding that departures from normal hiring processes may justify inferences of discrimination. See Maj. Op. at 509 (citing Lathram v. Snow, 336 F.3d 1085, 1093-94 (D.C.Cir.2003); Johnson v. Lehman, 679 F.2d 918, 922 (D.C.Cir.1982)). But an employer's departure from standard procedures does not automatically support such an inference; the plaintiff must still establish discriminatory motive. Johnson, 679 F.2d at 922. Here, in fact, WMATA's effort to devise a Lewis-free hiring process was itself a departure from existing practice; a departure from the departure merely returned things to the status quo. See Maj. Op. at 509. So it is far from obvious that the cases involving departure from existing practice apply. Even assuming they do, such cases always challenge a departure that actually disadvantaged the applicant in some way. See, e.g., Lathram, 336 F.3d at 1093 (decision to expand search outside agency thereby allowing second candidate to receive a veteran's preference and thus to outscore internal candidate); Pratt v. City of Houston, 247 F.3d 601, 605 (5th Cir.2001) (allegation that interviewer sprung a computer skills test on applicant, having given no notice that such a test would be part of the process). Here there is simply no evidence that the alleged departure from practice hurt Salazar's chances, much less an affirmative showing that the "departure" was discriminatory. See Risher v. Aldridge, 889 F.2d 592, 597 (5th Cir.1989).

As the majority points out, the record evidence clearly shows that "the scores Jaggie gave Salazar differed little from those given by two other panelists whom Salazar himself acknowledges would be unlikely to discriminate," Maj. Op. at 509, that "Salazar would have finished fourth even had Jaggie's scores not counted," id., and that "no evidence suggests that Jaggie attempted to influence the other panelists' scores during the interview," id. Because Jaggie's involvement had no measurable impact on the hiring outcome, the majority seeks to shore up the case for a plausible inference by emphasizing that Jaggie developed the interview questions and assigned the relative value for each, giving "only marginal value to candidates' experience and education, which were Salazar's particular strengths." Maj. Op. at 509.

But Salazar nowhere asserts that any of the questions themselves was unfair or inappropriate, or that the distribution was in any way unusual. Indeed, they would seem to cover precisely the range of scenarios relevant to the selection of a qualified supervisor. Compare Maj. Op. at 509 ("Jaggie not only developed the questions but assigned the relative value for each, and had apparently unfettered discretion in doing so.") (emphasis added) (internal quotation marks and citations omitted). And lest there be any confusion about the matter, it was not an odd or idiosyncratic weighting scheme that cost Salazar the post. The majority notes that out of 13 questions, 11 (with a weight of nearly 90%) related to hypothetical scenarios, Metro policies, technical responses, and candidate motivation, and two (with a weight of slightly over 10%) to the candidate's experience and education. Maj. Op. at 504-505, 509. The majority disparages the former by characterizing the latter two questions as the "only" ones that "directly addressed the candidates' experience and education." Id. at 505. But interview procedures that give more weight to direct evidence of a candidate's ability to handle problems than to the length of his resume are hardly evidence of discrimination — at least in the absence of evidence that the weighting was unusual.

Moreover, Jaggie's weighting decisions had no impact on Salazar. If all questions had been weighted equally, Salazar would still have finished in fourth place. More striking, in light of Salazar's and the court's focus on "experience," Salazar still would not have been selected even if the scorers had counted only the two resume questions — unless a scoring error that awarded Salazar 20 points for a question with a maximum allowable score of 10 points went uncorrected (itself, one hopes, a deviation from standard practice). See J.A. 111. Thus, in a contest focused exclusively on Salazar's "particular strengths," he could still have won only by means of an scoring error.

Lastly, the majority is "not as certain" as I that Salazar would have lost under virtually any plausible circumstances, because Salazar argues that "by weighing less significant questions more and by minimizing the value of the important questions regarding experience, education and training, the selectee came up with a higher score." Maj. Op. at 510-11 (internal quotation marks omitted). Salazar did not explain how, but it is possible (as the majority indicates) to reverse-engineer weighting schemes that allow Salazar to emerge the winner (correcting errors and excluding Jaggie's scores). One obvious scheme would simply ignore all questions except three — the two resume questions noted above, and a third question on WMATA's Five-Point Pledge. (The pledge reads: "(a) Maintain safe, clean, and attractive facilities and services, (b) Always be courteous, helpful and informative, (c) Strive to provide on time service, (d) Listen and respond to our customers, (e) Be innovative, resourceful, market driven, and entrepreneurial.") Candidates were asked to state the pledge and explain their role in its implementation. Inclusion of this question (and exclusion of all others) would put Salazar over the top because of the surprising coincidence that the candidate who did better than he on the experience and education questions managed to do worse on the pledge. The majority suggests other schemes in which Salazar could have won, for example, by "halving two questions and tripling two others, by tripling four questions," or "by dropping three questions." Maj. Op. at 511. Indeed because question weights are continuous variables, it is theoretically possible to identify an infinite number of weighting schemes from which Salazar would have emerged victorious, and from that trivial epiphenomenon the majority believes the jury could infer discrimination. Thus, it argues, a jury could infer that WMATA's purpose in giving material weight to questions about technical proficiency, safety, or management of other employees was to "get" Salazar. If such an inference is "reasonable," judges really are potted plants.

The court correctly points out our salutary rule against considering arguments not raised in the parties' briefs. Maj. Op. at 510. But WMATA did argue that if Jaggie's scores and the technical questions were eliminated Salazar would still lose. Appellee's Br. at 17. That argument is in fact correct (though narrower than the one I've made), and Salazar never responded to it except with vague generalities. See Appellant's Reply Br. at 2-4. The court instead takes up the Reply Brief's assertion that "significant [sic] [meaning, evidently, insignificant] items were weighted heavily." Id. at 2; see Maj. Op. at 510. Thus the court implicitly endorses Salazar's characterization of virtually all the questions relating to technical proficiency, safety and employee management as "[in]significant." Why? In any event, as Salazar presented the argument only in his Reply Brief, I'm unsure why WMATA should be faulted for not responding to it. More generally, when appellate judges confront conflicting claims about the meaning of the record, I had thought it kosher for them — actually, their duty — to look at the record.

Because no single factor cited by Salazar could justify a finding of discrimination, the majority turns to its own somewhat "distorted" weighting scheme — emphasizing the "cumulative effect" of Salazar's arguments. Maj. Op. at 511. Yet whether considered alone or in sum, the evidence presented is simply not sufficient for a reasonable jury to infer discrimination. The question is "not whether there is literally no evidence, but whether there is any upon which a jury could properly proceed to find a verdict for the party producing it, upon whom the onus of proof is imposed." Anderson, 477 U.S. at 251, 106 S. Ct. 2505 (quoting Pennsylvania R. Co. v. Chamberlain, 288 U.S. 333, 343, 53 S. Ct. 391, 77 L. Ed. 819 (1933) (emphasis in original)). The evidence must be such that a jury could find not merely that the point in question (here, discrimination) is conceivable; the proof must be such that the jury could — reasonably — find it more probable than not. The majority's method, effectively requiring WMATA to offer proof absolutely excluding any possibility of discrimination, however remote, is simply not the standard for summary judgment.

In the end, the case boils down to the presence in the hiring process of Jaggie, and thus by extension Lewis. Believing this is too slender a reed to support a jury verdict in Salazar's favor, I would affirm the decision of the district court.