with whom
Chief Judge JACOBS, Judge CABRANES, and Judge LIVINGSTON, join,dissenting.
The court today decides not to convene en banc to review Watson v. Geren, 569 F.3d 115 (2d Cir.2009), a decision that declines to apply the remand rule to permit an executive agency to remedy a procedural error of inadequate explanation at its final step of decisionmaking. Instead, upon judicial review of the administrative record, Watson concludes that it would be impossible for the agency to identify any reason with a basis in fact to support its challenged decision. The conclusion is disturbing for many reasons, not least of which are that the “agency” in question is the United States Army and the challenged decision is the denial of a conscientious objector application, an issue on which the applicant bears the burden of proof by clear and convincing evidence and over which judicial review is “the narrowest known to the law.” United States v. Corliss, 280 F.2d 808, 810 (2d Cir.1960) (Friendly, J.) (internal quotation marks omitted).1 To justify its departure from the remand rule — which here results in the affirmance of a district court judgment ordering the Army to classify petitioner as a conscientious objector and to grant him immediate discharge, see Watson v. Geren, 483 F.Supp.2d 226 (E.D.N.Y.2007) — Wafson invokes the doctrine of “futility.” Heretofore, our court has applied “futility” to affirm agency decisions where we could confidently conclude that the agency would reach the same result in the absence of the identified error. See, e.g., Krauss v. Oxford Health Plans, Inc., 517 F.3d 614, 630 (2d Cir.2008) (cited in Watson to support *161application of futility doctrine, see 569 F.3d at 129). Such application, akin to harmless error review, finds support in the Administrative Procedure Act (“APA”), which instructs courts to take “due account ... of the rule of prejudicial error” when reviewing agency action. 5 U.S.C. § 706.2 There is no comparable support for using “futility” as Watson does to reverse a challenged agency action without allowing it to correct an identified procedural omission of explanation. Such an application of the futility doctrine upsets the balance of authority between the executive and judicial branches that underlies the remand rule. See generally SEC v. Chenery Corp., 332 U.S. 194, 196, 67 S.Ct. 1575, 91 L.Ed. 1995 (1947) (cautioning against courts propelling themselves into “domain ... set aside exclusively for the administrative agency”); accord FPC v. Idaho Power Co., 344 U.S. 17, 20, 73 S.Ct. 85, 97 L.Ed. 15 (1952) (observing that “function of the reviewing court ends when an error of law is laid bare. At that point the matter once more goes to the [agency] for reconsideration”). In no area is this balance more carefully calibrated than in matters relating to the nation’s defense, including administration of its armed forces. See generally Orloff v. Willoughby, 345 U.S. 83, 93-94, 73 S.Ct. 534, 97 L.Ed. 842 (1953) (“Orderly government requires that the judiciary be as scrupulous not to interfere with legitimate Army matters as the Army must be scrupulous not to intervene in judicial matters.”). Thus, while I do not, in fact, agree with Watson’s conclusion that it would be impossible as a matter of law for the Army to state any reason with a basis in fact for denying petitioner conscientious objector status, see infra Part II.B, my request for the court to convene en banc was prompted by a larger concern as to the propriety of a court undertaking such review of the record at this stage of the case, see infra Part II.A. Accordingly, I respectfully dissent from the denial of en banc review.
I. Background
This challenge to the Army’s denial of conscientious objector classification came into federal court on a petition for a writ of habeas corpus.3 The background to the petition includes an extensive administrative record developed over several months of proceedings governed by Department of Defense and Army regulations. See generally Department of Defense Instruction *162(“DOD Instr.”) 1300.06; Army Regulation (“AR”) 600-43. I here provide only a brief summary of these proceedings preliminary to discussing my reasons for urging en banc review.
A. Watson’s Voluntary Agreement To Serve in the Armed Forces
In 1998, Timothy Watson was enrolled in medical school at George Washington University when he applied to the United States Army for financial assistance with his education. Pursuant to its Health Professions Scholarship program, the Army agreed to pay for Watson’s remaining three years of medical school, in return for which Watson promised that, after graduation, he would serve three years of active duty and five years in the reserves. Although the Army kept its side of the bargain,4 Watson now seeks to avoid his, claiming that he should be excused from even the non-combatant role of an Army doctor because he is a conscientious objector.5 Indeed, Watson contends that his objection to war would require him to refuse to treat any member of the United States armed forces — whether on active duty or in reserve status, whether in a combat zone or in a veterans’ hospital6— because to provide such care would be “the functional equivalent of weaponizing human beings” by making them healthy enough to serve in combat. Application of Timothy Doyle Watson (“Watson App.”) at 5.
B. The Army’s Denial of Watson’s Conscientious Objector Claim
Watson filed for conscientious objector status in 2006, the same year his deferment was to expire. The application was reviewed pursuant to a series of Army procedures, which included (1) Watson’s preliminary interview by an Army chaplain, who expressed reservations as to Watson’s sincerity; (2) a non-adversarial hearing before an investigating officer, Colonel Clinton R. O’Neill III, who heard Watson testify, found him sincere, and recommended that he be classified as a conscientious objector; (3) sequential record review by four army officers, each of whom recommended that Watson be denied conscientious objector classification;7 *163and (4) further record review by a three-member panel of the Department of the Army Conscientious Objector Review Board (“DACORB”), the final Army decisionmaker, which on December 5, 2006, denied Watson’s application for conscientious objector classification by a vote of two to one.
C. Judicial Proceedings
On January 17, 2007, the office of the Army Surgeon General ordered Watson to report for active duty on February 5. One week later, on January 24, Watson challenged this order in a petition for a writ of habeas corpus filed in the United States District Court for the Eastern District of New York. Therein, Watson claimed that the Army had erred in failing to discharge him as a conscientious objector. The district court promptly granted a stay of Watson’s order to report, see Watson v. Harvey, No. 07-CV-0345 (E.D.N.Y. Jan. 25, 2007), and, on April 10, the court granted the writ, identifying procedural error in DACORB’s provision of only conclusory reasons for its denial of Watson’s objector application, see Watson v. Geren, 483 F.Supp.2d at 241.8 The district court further observed that, even if it were to assume that DACORB’s reasons for denial were those expressed in the more detailed recommendations of the four intermediary reviewing officers and the report of the interviewing chaplain, Watson would still be entitled to habeas relief, because none of these reasons had a basis in fact. See id. at 241-42. Accordingly, the court denied the government’s request for remand and ordered the Army “to grant petitioner’s application for CO status and release him through immediate discharge.” Id. at 251.
*164On appeal, the Army did not dispute the district court’s identification of procedural error. It challenged only the refusal to remand the case to afford DACORB an opportunity to provide the required explanation for its denial decision. While Watson acknowledges the general applicability of the remand rule, it declines to apply it in this case, concluding that “remand would be utterly futile” because the court’s own review of the record indicates “no basis in fact for denial of the application on any valid ground.” Watson v. Geren, 569 F.3d at 134-35. Watson affirms the district court judgment, effectively ordering discharge. Id.
II. Discussion
A. The “Futility Doctrine” Does Not Provide an Exception to the Remand Rule in This Case
It is a “fundamental rule of administrative law ... that a reviewing court, in dealing with a determination or judgment which an administrative agency alone is authorized to make, must judge the propriety of such action solely by the grounds invoked by the agency.” SEC v. Chenery Corp., 332 U.S. at 196; accord United States ex rel. Checkman v. Laird, 469 F.2d at 780. A necessary corollary to this rule that “administrative action is to be tested by the basis upon which it purports to rest” is that the basis “must be set forth with such clarity as to be understandable” to the reviewing court. SEC v. Chenery Corp., 332 U.S. at 196. When the basis for a challenged administrative decision is not understandable, “the appropriate course for a reviewing court ordinarily is to remand the case to the agency.” Ward v. Brown, 22 F.3d 516, 522 (2d Cir.1994); see also Florida Power & Light Co. v. Lorion, 470 U.S. 729, 744, 105 S.Ct. 1598, 84 L.Ed.2d 643 (1985) (“[I]f the reviewing court simply cannot evaluate the challenged agency action on the basis of the record before it, the proper course, except in rare circumstances, is to remand to the agency for additional investigation or explanation.”).
Watson acknowledges the broad scope of this remand rule but declines to apply it in this case, concluding that it would be “futile” to ask DACORB to explain its conclusory denial of conscientious objector classification because the court’s own review of the record reveals “no basis in fact to support” any possible explanation. Watson v. Geren, 569 F.3d at 134. In support, Watson cites Krauss v. Oxford Health Plans, Inc., 517 F.3d 614. See Watson v. Geren, 569 F.3d at 129. Krauss does not, in fact, support Watson’s use of the futility doctrine to reverse an administrative decision before the agency has had an opportunity to provide the required explanation.
In Krauss — an ERISA case — we determined that remand was futile because the plan’s decision to deny benefits was substantively correct, notwithstanding procedural error in reaching that decision. We explained that because “the relevant information has been finally disclosed [to plaintiffs], we are confident that administrative remand would be futile.... [The plan]’s benefits determination, even if not properly explained at the time of denial and during administrative review, was, as a substantive matter, an appropriate implementation of ... the Plan.” 517 F.3d at 630. In short, we used the futility doctrine in Krauss to affirm a challenged agency decision, in much the same way we use the rule of harmless error.
In other cases in which our court has invoked “futility” as an exception to the remand rule, our analysis has similarly been akin to harmless error review. Notably, in immigration cases, our determination that remand would be futile has invar*165iably turned on “how confidently we can predict that the agency would reach the same decision absent the errors that were made.” Li v. Mukasey, 529 F.3d 141, 150 (2d Cir.2008) (internal quotation marks omitted); see also, e.g., Niang v. Mukasey, 511 F.3d 138, 149-50 (2d Cir.2007); Manzur v. U.S. Dep’t of Homeland Sec., 494 F.3d 281, 295-96 (2d Cir.2007); Siewe v. Gonzales, 480 F.3d 160, 166-67 (2d Cir.2007). We have taken this same approach to futility in reviewing challenges to other administrative rulings, for example, by the Treasury Department, see Karpova v. Snow, 497 F.3d 262, 269 (2d Cir.2007) (declining to remand to Treasury’s Office of Foreign Assets Control on ground that we were “confident that the Agency would reach the same conclusion absent [the alleged] error”); by the Environmental Protection Agency, see Natural Res. Def. Council, Inc. v. U.S. EPA 494 F.2d 519, 525 (2d Cir.1974) (declining to remand to EPA because “[t]here is no reason” to make “the Administrator busy himself disproving a thousand negatives to prove a single positive”); and, as in Krauss, by ERISA plan administrators, see Giordano v. Thomson, 564 F.3d 163, 168 & n. 3 (2d Cir.2009) (characterizing remand as futile because plan did not err in denying benefits). In all these cases, we invoked futility to uphold agency decisions because we had no doubt that the agency would reach the same conclusion in the absence of the identified error.
This use of “futility” to affirm agency rulings sensibly expedites resolution of the underlying disputes in circumstances where the losing party was not prejudiced by the error. Such application is specifically provided for in the APA, see 5 U.S.C. § 706 (instructing reviewing courts that “due account shall be taken of the rule of prejudicial error”), and in no way diminishes the deference courts owe agencies’ substantive expertise or decisionmaking authority.
Watson, however, invokes futility for a quite different purpose. Rather than invoke futility to affirm an agency decision presenting only harmless error, Watson invokes futility to reverse an agency decision on the ground that the error is purportedly so fundamental and pervasive as to be uncorrectable as a matter of law. I respectfully suggest that our authority to take such action in lieu of remand is extremely narrow9 and not properly exercised in this case, where the identified error is one of procedural omission, i.e., a lack of explanation, and where the matter in dispute, i.e., conscientious objector classification, is one whose substance is committed exclusively to military expertise. See generally McGee v. United States, 402 U.S. 479, 486, 91 S.Ct. 1565, 29 L.Ed.2d 47 (1971) (noting that, unlike “dispute[s] about statutory interpretation,” military classification decisions “depend[] on the application of expertise by administrative bodies in resolving underlying issues of fact”).10 In such circumstances, it is inl*166perative to remand for the correction of the administrative procedural error before undertaking judicial review of the substantive decision because a court must “know what a decision means” before it can “say whether it is right or wrong.” United States v. Chicago, M., St. P. & P.R. Co., 294 U.S. 499, 511, 55 S.Ct. 462, 79 L.Ed. 1023 (1935) (Cardozo, J.).
That conclusion is only reinforced by the fact that our review of Army classification decisions is limited to “basis in fact.” While such review is not “toothless,” Hanna v. Sec’y of the Army, 513 F.3d 4, 12 (1st Cir.2008), it is narrower even than the substantial evidence review we apply to the vast majority of administrative challenges, see United States ex rel. Checkman v. Laird, 469 F.2d at 787. Precisely because our review of military classification decisions is so narrow, we have emphasized that “[t]he proper focus of a reviewing court is on the reasons given” by DACORB, “not on reasons that may come to light if and when a court rummages throughout the record in an effort to reconstruct on what basis the board might have decided the matter.” Id. at 783. In Checkman we made that point in rejecting a government argument that a court’s own identification of a basis in fact for an Army classification could support denial of an objector’s habeas challenge to the classification. See id. at 784-85. But the same reasoning surely applies to the argument that a court may reverse an unexplained Army classification on its own attempted identification of the reasons for an agency decision. A classification decision “must stand or fall on the basis of the reasons stated.” Id. at 780 (emphasis added).
In Checkman, the Army had provided actual reasons for its challenged classification decision, which our court found to lack a basis in fact. It was in the face of these “errors of commission” that Checkman refused to remand to afford the Army an opportunity to amplify its reasons. Id. at 787-88. This is consistent with other decisions in which we have entered judgments for conscientious objectors after finding that articulated reasons for decision lacked a basis in fact. See, e.g., United States ex rel. Foster v. Schlesinger, 520 F.2d 751 (2d Cir.1975); Ferrand v. Seamans, 488 F.2d 1386 (2d Cir.1973). In each of these cases, we reviewed the actual reasons advanced by the final decisionmaker; we made no assumptions as to what reasons might have informed an unexplained decision. Indeed, Watson cites to no case in which a court, upon finding only an omission error of explanation, simply bypassed the remand rule and, after identifying what it thought were all possible reasons for the decision and finding them lacking a basis in fact, entered a judgment that effectively reversed DACORB and ordered the objector’s immediate discharge. Such action hardly comports with Checkman’s cautionary observation that “[t]he court is not a hostile stranger to the official or agency” making a final administrative decision; *167“rather the two act, in conjunction,” each with distinct “roles in furtherance of justice and the public interest.” 469 F.2d at 781.11
I respectfully submit that the court exceeds its limited role in the review of military classification decisions when it conducts basis-in-fact review before affording the military an opportunity to remedy an omission error of explanation. The Fourth Circuit recognized as much in United States ex rel. Coates v. Laird, 494 F.2d 709 (4th Cir.1974). In that case, a district court, confronted with a military failure of explanation, “reviewed the full military record in order to glean from it possible reasons for the denial” and then ordered discharge. Id. at 711. The Fourth Circuit reversed, holding that such review of the record was “unnecessary.” Id. “[T]he proper procedure” was “to remand the proceedings to the Marine Corps” with instructions to “follow scrupulously the regulation” requiring that the agency state its reasons for denying the application. Id. at 713 (internal quotation marks omitted). The same conclusion applies here.12
In sum, I think the remand rule should have been applied in this case to afford the Army an opportunity to remedy its explanation omission, and I do not think the futility doctrine provides an exception to that rule that permits a court to reverse an Army denial of conscientious objector classification because the court itself could not identify a reason with a basis in fact for the challenged decision.
B. The Application of Basis-irir-Fact Review to This Case
Because I do not think it is appropriate for this court to undertake any basis-in-fact review of Watson’s classification chal*168lenge in the absence of a proper statement of reasons for the final Army decision, I am reluctant to comment on what valid reasons might be advanced. Nevertheless, because I am far from convinced by Watson’s analysis that it would be impossible as a matter of law for the Army to advance any such reasons, I offer a few observations, by no means exhaustive, about basis-in-fact review generally and as applied to this case.
First, as already noted, basis-in-fact review is the narrowest standard known to law, see United States v. Corliss, 280 F.2d at 810, specifically intended to limit judicial intrusion into military decisionmaking, see United States ex rel. Checkman v. Laird, 469 F.2d at 781. Thus, once a court is satisfied that procedural fairness was afforded, its basis-in fact review is limited to ensuring the rationality of the challenged military decision. As Judge Friendly described the task: a court “in effect must determine, as best it can, whether the Local Board or the hearing officer and, ultimately, the Appeal Board were rational and sincere in disbelieving the sincerity of registrant’s belief [even] in the absence of conduct inconsistent with the registrant’s assertion, and this on a record always cold and often thin.” United States v. Corliss, 280 F.2d at 814-15 (emphasis added).
Second, the fact that a person states a prima facie case for conscientious objection classification does not require military authorities to accept it as sincere. See Witmer v. United States, 348 U.S. 375, 381-82, 75 S.Ct. 392, 99 L.Ed. 428 (1955); United States ex rel. Checkman v. Laird, 469 F.2d at 778; United States v. Corliss, 280 F.2d at 814. When the hearing officer who interacts directly with the applicant finds him to be sincere, that finding is entitled to “great weight.” Ferrand v. Seamans, 488 F.2d at 1390. Nevertheless, even in such circumstances, other officers in the review chain may recommend denial of objector classification provided there is “objective evidence affording a rational basis” for the “refusal to accept the validity of the applicant’s claims.” Lovallo v. Resor, 443 F.2d 1262, 1264 (2d Cir.1971); see also United States ex rel. Checkman v. Laird, 469 F.2d at 778; United States v. Corliss, 280 F.2d at 814. On judicial review, such objective evidence need not be “preponderant” or even “substantial” to constitute a basis in fact for decision. United States ex rel. Foster v. Schlesinger, 520 F.2d at 755. Indeed, courts conducting basis-in-fact review of objector classification do well to bear in mind Judge Friendly’s observation that “though the state of a man’s mind is as much a fact as the state of his digestion, it is a lot less susceptible of objective determination.” United States v. Corliss, 280 F.2d at 810 (internal quotation marks and internal citation omitted). Moreover, courts must recall that the burden remains always on the applicant clearly and convincingly to prove his claim of conscientious objection. See DOD Instr. 1300.06 § 5.3; AR 600-43, ¶ 1-5c.
Third, while objective evidence is more than mere speculation, it does not equate to direct evidence. See Witmer v. United States, 348 U.S. at 383, 75 S.Ct. 392 (observing that it would be “pure naivety” to expect “an outright admission of deception”; thus, the most “competent evidence” will generally be “inference[s]” drawn from the applicant’s testimony and conduct). Nor is it evidence admitting only a single inference. See id. (recognizing Congress’s intent to make Selective Service review final “in all cases where there was conflicting evidence or where two inferences could be drawn from the same testimony”). Indeed, basis-in-fact review does not contemplate any judicial weighing of the evidence. See United States ex rel. Checkman v. Laird, 469 F.2d at 779. Rather, courts *169must view the record in the light most favorable to the decisionmaker in determining whether a challenged classification was rational.
With these principles in mind, I proceed to explain why I cannot agree with Watson’s conclusion that it would be impossible for DACORB to point to any basis in fact for its denial decision. It is useful to begin by considering exactly how Watson’s professed beliefs conflict with the service commitment the Army asks him to honor.
Watson submits that he opposes war because he believes that (1) all human life is sacred, (2) war is a futile means for resolving human disputes, and (3) war inevitably causes a large number of noncombatant casualties. As a result, Watson unequivocally refuses to kill other human beings, and indicates that he would die himself rather than do so. Up to this point, there is no problem. The Army does not ask Watson to serve in a combat role. Quite the contrary, the Army asks Watson to honor his commitment by serving as a non-combatant in a role where his sole responsibility would be to save human lives. Thus, we come to the crux of Watson’s objection claim: he refuses to play even the life-saving role of an Army doctor because he deems it immoral to participate in any way in “weaponizing human beings.” Watson App. at 5.
In considering the sincerity of this claim, the Army might reasonably consider that Watson’s involvement in any such “weaponization” would necessarily be indirect. Watson is not, after all, a surgeon, an orthopedist, or a neurologist. He is a radiologist — in short, he reads films — and, as such, would not likely make any final treatment decisions for American soldiers, much less would he be likely to provide that treatment.13 Nevertheless, Watson insists that it would be immoral to play even a tangential role in providing any medical care for American soldiers.
The Army might reasonably identify an inconsistency between this opposition and Watson’s professed belief in the inviolable sanctity of every human life, the core of his personal moral code. Watson submits that there is no inconsistency because he is not suggesting that wounded American soldiers should be left to die; he is simply refusing himself to play any part in their treatment. He expects that other doctors — presumably with less refined moral codes — would provide the treatment necessary to save soldiers’ lives.
It would hardly be irrational for DACORB to reject this tortured reasoning. If a person sincerely believes an act is immoral, then the person might reasonably be expected to believe that the act is immoral no matter who commits it. I am hardly suggesting that Watson should believe it immoral for any doctor to treat wounded American soldiers, a position that *170necessarily leads to the breathtaking conclusion that it is morally preferable for such wounded Americans to die for lack of medical care than for any doctor to play a part in their “weaponization.” I am noting simply that Watson’s effort to have it both ways, refusing to treat wounded soldiers himself but not opposing treatment by others — thereby denying the possibility of, and avoiding responsibility for, the loss of soldiers’ lives — exposes the shallow moral foundation of his claim to conscientious objector status. DACORB could reach the same determination from the stated facts and rationally conclude that Watson had not clearly and convincingly established that his moral opposition to providing any medical care for American soldiers is a belief “held with the strength and devotion of traditional religious conviction.” AR 60CM3, Glossary.14
Indeed, the shallowness of Watson’s moral opposition to treating soldiers is further revealed by another objective fact: his failure to take any steps in his civilian medical practice to ensure that he did not treat, and thereby “weaponize,” any United States military personnel, for example, reservists or persons on leave from active duty. Watson apparently failed even to consider such action.15 The oversight is noteworthy because the likelihood of Watson’s encountering some reservists, and even active duty soldiers, as a hospital resident was hardly remote, and the very essence of his claim is a moral objection to playing any role, however tangential, in weaponizing soldiers.16 Moreover, DOD Instr. § 5.2.2 specifically states that “[s]incerity is to be determined by an impartial evaluation of the applicant’s thinking and living in its totality, past and present,” and § 5.2.2.1 requires careful examination of an applicant’s “outward manifestation of the beliefs asserted.” 17
*171In addition to troubling facts inherent in Watson’s claim, there is the objective chronological fact that Watson did not file for conscientious objector classification until 2006, the same year his service deferment was scheduled to expire. Watson notes that Army regulations preclude belated or coincidental filing from being the sole reason for denying conscientious objector classification. See Watson v. Geren, 569 F.3d at 132. True enough, see AR 600-43, ¶ 1-5a(5)(c) (stating that “timing of an application alone ... is never enough to furnish a basis in fact to support a disapproval”), but the same regulation specifically recognizes that coincidental timing may indicate “that further inquiry as to the person’s sincerity is warranted,” id. The reason is grounded in common sense. While “[sjudden accessions of belief may be utterly sincere, as the memorable one on the Damascus road,” United States v. Corliss, 280 F.2d at 812, “[a]ny parent whose child claims to be sick on the morning of the final exam knows better than to take such a claim at face value,” Hanna v. Sec’y of the Army, 513 F.3d at 18 (Boudin, J., dissenting). Thus, the Army is not foreclosed from relying on timing together with other facts in providing a rational explanation for its denial decision. See Witmer v. United States, 348 U.S. at 383, 75 S.Ct. 392 (recognizing that facts “possibly insignificant standing alone” may in context of totality of circumstances “help support the finding of insincerity”); Lobis v. Sec’y of the U.S. Air Force, 519 F.2d 304, 307 (1st Cir.1975) (observing that timing might warrant “substantial weight” where “reinforced by other evidence”).
The Army might identify further objective facts relevant to its denial decision in the findings in Col. O’Neill’s investigation report. To explain, it is useful to begin by listing the subjects expected to be addressed in such a report:
(a) The underlying basis of the person’s professed conscientious objection (what applicant believes, and why).
(b) The time period (being as specific as possible) in which the person’s belief became fixed.
(c) Whether the belief constitutes conscientious objection (1-0 or 1-A-0) under this regulation.
(d) The sincerity of the person, including reasons for such conclusions.
AR 600-43, ¶ 2-5k(5).
In response to the first inquiry, the investigation report states simply that “Watson no longer believes in serving in the military in any capacity and he feels very strongly that treating wounded soldiers and sending them back to fight results ‘in the functional equivalent of weaponizing human beings.’ ” Investigation Report, July 12, 2006, at 1. It is hardly apparent that these views, particularly the first one, equate to the required opposition to all war, a point that might reasonably inform DACORB’s consideration of other aspects of the report.
In response to the second inquiry, the report identifies the terrorist attacks of September 11, 2001, and the United States’ “pre-emptive strikes in Afghanistan and especially Iraq” as the events prompting Watson to study various writers and philosophies addressing “violence, the causes of violence and alternatives to violence.” Id. The report makes no express finding, however, that this study in fact expanded Watson’s opposition from the wars in Afghanistan and Iraq to wars generally, nor does it indicate when any of Watson’s views became “fixed.” Id. The silence is significant. DACORB — which has much more experience than this court in reviewing such reports — might conclude therefrom that the four reviewing officers, each of whom determined that Watson had convincingly demonstrated opposition only *172to the Afghanistan and Iraq wars and not to all wars, were not rejecting a specific finding by the investigation officer but, rather, identifying a concern not addressed by that officer. The Army review process “relies heavily on the considered opinions of a number of officers in addition to those who serve on the DACORB.” Aguayo v. Harvey, 476 F.3d 971, 979 (D.C.Cir.2007). When they are all in agreement that conscientious objector classification should be granted, see, e.g., Hanna v. Sec’y of the Army, 513 F.3d at 11, DACORB confronts a greater challenge in identifying a basis in fact for a denial decision. But where, as here, the investigation officer’s recommendation to grant conscientious objector classification is rejected by all four reviewing officers on a ground not specifically addressed in the investigation report, surely a court should proceed cautiously in concluding that these reviewing officers all acted irrationally and without a basis in fact, and that DACORB would necessarily do likewise if it were asked to provide a fuller explanation of its reasons for denial.
Indeed, such caution is further signaled by the investigation report’s response to the third inquiry, which states at the outset that Watson presently fits somewhere in “the middle ground” of the definition of a conscientious objector. Investigation Report, July 12, 2006, at 1. While the report does not explain this characterization, context suggests that it is not meant to signal that Watson fit squarely within the definition but, rather, that the evidence did not all point in one direction on the issue. The report states that Watson has “deeply held moral and ethical beliefs that fit in with the definition” of a conscientious objector, noting particularly that “[h]e is against the U.S. Army and all Armies, and would not care for injured service members.” Id. At the same time, it notes that Watson is “not a strict pacifist” because he is willing to use force as necessary to “defend his loved ones and himself,” but not “to inflict harm or exact retribution.” Id. Observing that these circumstances required “a judgment call on where this soldier fits in the definition” of a conscientious objector, the investigation report recommends 1-0 classification because of Watson’s “strong anti war bias, and refusal to treat combatant Soldiers.” Id. at 1-2. In response to the fourth inquiry, the report states that Watson’s views are sincerely held and not likely to change. See id.
In considering these responses to the last two inquiries, one is struck by the finding that what Watson’s beliefs prompted was opposition to “all Armies,” a curious formulation for the investigation officer to employ given his undoubted awareness that the regulatory requirement is opposition to “all wars.” See DOD Instr. 1300.06 § 3.5.1; AR 600-43, Glossary, Section II; see also Gillette v. United States, 401 U.S. 437, 443, 91 S.Ct. 828, 28 L.Ed.2d 168 (1971). To the extent the report indicates that Watson has a “strong anti war bias,” it makes no specific finding as to whether this bias is directed at all wars or at the wars presently being fought by the United States in Afghanistan and Iraq, in which Watson might be expected to serve. Thus, even if one were to assume that this reference was to all wars, the failure of explanation, particularly on the point of concern noted by four reviewing officers, is itself an objective fact that DACORB might reasonably consider in making its own final decision.
The reviewing officers’ concern — and possibly DACORB’s — on this point is further reasonably informed by Watson’s own testimony at the investigation hearing. While this testimony was not formally transcribed, the investigation officer prepared a memorandum of the questions *173asked and answers given that is usefully contrasted to Watson’s statements in his counseled written submissions.18 Three questions and answers suffice to make the point:
Q: What exactly are your beliefs?
A: Basically being as concise as I can be concerning my beliefs they are that human life is a gift and we are charged with the responsibility of honoring that gift. And to do so we must make every effort to ensure we save, protect, and uphold human life.
Q: What basic values besides objecting to violence do you have?
A: My basic values are complex. Human beings are rational ] and should be able to decipher between good and bad and this is a core value. In addition, our gift is to think and reason and we have an obligation to do so. To shirk this responsibility is to deny who we are. I believe in the humanistic view point. I believe in the goodness of us as a species animal, and respect we are basically an animal, but we have the ability to reason and can deny any animalistic instincts, and be more responsible and reach gracious conclusions. Dealing on a day-to-day basis, I honor each individual as an individual with their own understanding of the world. I help them with decisions concerning their health and well being to the best of my ability while remaining non-judgmental and practical, but using a meaningful approach.
Q: When did these beliefs become fixed in your mind, and you realized you were a CO?
A: Sometime in late 2004 my beliefs became fixed when our efforts in Fallujah came to light and I had an increasing concern that our approach in Iraq and Afghanistan was unfounded. I remember the picture of a soldier shooting an unarmed Iraqi at close range crying out. This really shook me and prompted me to investigate what we were doing and what got us into this situation and quest. My personal awakening of this and future missions peculated [sic] in early 2005, which *174led to on-line organizations that made more sense to me. Sometime in the late or early summer of 2005, I was uncomfortable with our part in the current situation. I discussed this with friends who were supportive and family members who were unhappy with this decision. My friends suggested fraudulent approaches such as using homosexuality or using my position as a physician to process a disability to get out of the Army. That made no sense. I broke out my contract specifically Section 21 concerning resignation as a CO. I also went online and connected with the Organization Center for the Counsel of COs which was most helpful to me.
Investigation Hearing Memorandum, July 12, 2006, at 2-8.
Notably absent from Watson’s responses to the first two inquiries is any mention of his opposition to all wars. Rather, the answers express general beliefs of good will likely held by large numbers of people, including many presently serving in this country’s armed forces. Even more noteworthy is the lack of any mention in Watson’s third response of his purported rigorous study of thinkers and philosophies that (according to his written submissions) was the critical factor in expanding and fixing his beliefs from opposition to the Afghanistan and Iraq wars to opposition to all wars. Watson does not testify to any such expansion of his views. Quite the contrary, he says that in the summer of 2005 his discomfort with “the current situation” prompted him to break out his service agreement with the Army to see how he could get out of it.
The investigation officer, the reviewing officers, and DACORB itself are, of course, obliged to review the record in its entirety. That record, as Watson notes, contains numerous written submissions asserting that Watson’s opposition to the Afghanistan and Iraq wars was simply the first step on a road that brought him to his present opposition to all war. See Watson v. Geren, 569 F.3d at 132.19 But the investigation officer never made an explicit credibility finding on this point, and neither the reviewing officers nor DACORB were required to credit those submissions, particularly in light of the objective fact of Watson’s hearing testimony indicating that it was the “current situation,” ie., the ongoing wars in Afghanistan and Iraq, that in the summer of 2005 prompted him to start reading not philosophy but his service agreement and to contact conscientious objector groups to see what would get him out of his commitment. Whether or not this testimony might somehow be reconciled with Watson’s claim of conscientious objection is not the point. It is objective evidence from which DACORB could rationally draw an inference that Watson’s real opposition is to the specific wars in which he would be required to serve, not to all war. See Witmer v. United States, 348 U.S. at 383, 75 S.Ct. 392 (holding that, where Selective Service picks between “conflicting evidence” or competing inferences, its decision has a basis in fact); United States v. Corliss, 280 F.2d at 815-16 (observing that, when confronted with competing statements of applicant’s basis for his conscientious objection to military service, Appeal Board was free to accept or reject last statement most *175favorable to applicant).20
Still other objective facts in the record could provide rational support for a DACORB determination that Watson failed clearly and convincingly to prove his opposition to all wars. For example, Watson professes to view all war as “an entirely shameful human endeavor.” Watson App. at 6. The particular war actions that he references are ones subject to easy condemnation or, at least, to present debate. See id. at 10 (citing “the Crusades to the atrocities of the Third Reich, the Turks in Armenia, the Iraq-Iran conflict, Stalin’s murderous regime, the Japanese campaign in the Pacific, to include the rape of Nanking, Rwanda, Ethiopia, Sudan, the American conquest of the Native peoples”). But this begs the difficult question critical to Watson’s demonstration of conscientious objection to all wars, ie, whether he also views as “shameful” those war actions generally recognized for their liberating effects. Lexington and Bunker Hill, Omaha Beach and Iwo Jima are notably missing from Watson’s list of shame, as are the Civil War campaign that led to Lincoln’s pronouncement of the Emancipation Proclamation and the Allied offensive in World War II that led to the liberation of Auschwitz and Buchenwald.21 Yet this must be Watson’s view if his refusal to save a life, if the life belongs to an American soldier, is to be understood as a sincere opposition to all war.
This discussion is not meant to exhaust the objective facts that might be relied on by DACORB on remand to explain its denial decision, but only to illustrate why I cannot agree with Watson’s conclusion that no explanation with a basis in fact could be given on remand. For that reason, but more specifically because I do not think the futility doctrine is properly applied in this case as an exception to the remand rule, I respectfully dissent from the denial of en banc review.
. Courts review military classification decisions only for procedural fairness and a basis in fact. United States v. Corliss, 280 F.2d at 810-11. The "basis in fact” standard is deliberately narrow in order to "limit the freedom of a court to substitute its judgment for that of an official or board.” United States ex rel. Checkman v. Laird, 469 F.2d 773, 781 (2d Cir.1972). It "is satisfied if there is objective evidence, even though not preponderant or substantial, to support the finding in question.” United States ex rel. Foster v. Schlesinger, 520 F.2d 751, 755 (2d Cir.1975).
. The APA is relevant to this case only by analogy because the Act "applies to the Army in peacetime,” Ornato v. Hoffman, 546 F.2d 10, 14 (2d Cir.1976); accord Lunney v. United States, 319 F.3d 550, 554 (2d Cir.2003), and a statutory exception applies in any event to military personnel decisions "committed to agency discretion by law,” see 5 U.S.C. § 701(a)(2) (excepting actions committed to agency discretion from judicial review); id. § 554(a)(4) (excepting "conduct of military or foreign affairs functions” from adjudication procedures); see also Ornato v. Hoffman, 546 F.2d at 14 (noting statutory exceptions and observing that conscientious objectors would have no judicial remedy but for availability of habeas corpus).
. The law and regulations governing objector classification assign no review role to the judiciary. See 50 U.S.C. app. § 456(j); Department of Defense Instruction 1300.06; Army Regulation 600-43. In no small part this is because conscientious objector classification is a privilege, not a right, granted by the executive consistent with its constitutional command authority over the armed forces. See U.S. Const, art. II, § 2; Nurnberg v. Froehlke, 489 F.2d 843, 849 (2d Cir.1973). Nevertheless, classification challenges can come before the courts in two contexts: (1) as defenses to prosecutions for failure to submit to military induction, see, e.g., United States v. Corliss, 280 F.2d at 810; and (2) on petitions for writs of habeas corpus, see, e.g., Hammond v. Lenfest, 398 F.2d 705, 715-16 (2d Cir.1968); see also Witmer v. United States, 348 U.S. 375, 377, 75 S.Ct. 392, 99 L.Ed. 428 (1955) (recognizing two contexts).
. Indeed, the Army went further, allowing Watson to defer his military service for five years after his 2001 graduation from medical school so that he could complete his internship and a residency program in radiology.
. "Conscientious objection” is defined as "a firm, fixed, and sincere objection to participation in war in any form or the bearing of arms, by reason of religious training and/or belief.” DOD Instr. 1300.06 § 3.1. Army regulations construe "religious training and beliefs” broadly to encompass “a sincere and meaningful belief which occupies in the life of its possessor a place parallel to that filled by the God of another, or, in the case of deeply held moral or ethical beliefs, a belief held with the strength and devotion of traditional religious conviction.” AR 600-43, Glossary. The applicant bears the burden of demonstrating "by clear and convincing evidence” his qualification for conscientious objector classification. DOD Instr. 1300.06 § 5.3; AR 600-43, ¶ l-5c.
Two conscientious objector classifications are possible: "1-0,” which applies to persons who object to serving in the military in any capacity and results in full discharge, DOD Instr. 1300.06 § 3.1.1; and "1-A-0,” which applies to persons who refuse to serve in combat but whose objections permit them to perform non-combatant military service, see id. § 3.1.2. Watson sought the former classification.
. Although Watson initially stated a willingness to serve in a veterans' hospital, he withdrew that offer upon learning that active duty soldiers and reservists are sometimes treated at such hospitals.
. The four Army reviewing officers were Brigadier General John E. Sterling, Jr., Commander of the Maneuver Support Center and *163Fort Leonard Wood, Missouri; Colonel Robert T. Marsh, Commanding Officer of the Human Resources Command, St. Louis, Missouri; and their respective legal advisors, Colonel Jerry J. Linn and Major Matthew D. Ramsey, both of the Judge Advocate General's Corps. In their individual reports, these officers stated various reasons for their denial recommendations, e.g., the timing of Watson’s application, the vagueness of his professed moral code, inconsistency between Watson's professed commitment to the sanctity of all human life and his refusal to treat wounded soldiers, and minimal evidence that Watson's beliefs carried over into other aspects of his life. One reason for denial, however, was cited by all four officers: Watson's failure convincingly to demonstrate that his opposition extended beyond the wars in Afghanistan and Iraq to reach "all wars.” DOD Instr. 1300.06 § 3.5.1 (stating that applicant's "objection must be to all wars rather than a . specific war”); AR 600-43, Glossary, Section II (defining “war in any form”); see also Gillette v. United States, 401 U.S. 437, 443, 91 S.Ct. 828, 28 L.Ed.2d 168 (1971) (“[Cjonscientious scruples relating to war and military service must amount to conscientious opposition to participating personally in any war and all war.”).
. AR 600-43, ¶ 2-8d(3) requires that denials of conscientious objector applications be accompanied by a statement of reasons. We have construed this requirement to demand more than "a bare recitation ... of the ultimate [regulatory] criteria.” United States ex rel. Checkman v. Laird, 469 F.2d at 787. DACORB must state "[t]he facts or factors relied upon” in reaching its conclusion. See United States v. Stewart, 478 F.2d 106, 113 (2d Cir.1973). DACORB’s explanation for its denial decision in this case, quoted in its entirety by the district court, see Watson v. Geren, 483 F.Supp.2d at 238, failed to satisfy this standard. DACORB’s President stated simply that Watson’s application was "not convincing,” and observed, "I am not drawn to believe in the applicant's sincerity and find him to be disingenuous and the application expedient.” Id. DACORB’s Staff Judge Advocate concluded that Watson “failed to demonstrate by clear and convincing evidence that he has a firm, fixed and sincere objection to war in any form.” Id. Even DACORB's Staff Chaplain, the only Board member to recommend conscientious objector classification, stated only that he found Watson's "beliefs to be sincere and fixed.” Id.
. It is noteworthy that judicial authority to reverse decisions of the Commissioner of Social Security without remanding the case for rehearing is expressly provided by statute. See 42 U.S.C. § 405(g). Even in this context, however, we have signaled a strong preference for an agency statement of reasons preliminary to undertaking substantive judicial review. See generally Halloran v. Barnhart, 362 F.3d 28, 33 (2d Cir.2004) (observing that our circuit "do[es] not hesitate to remand” and “will continue remanding when we encounter opinions from ALJ’s that do not comprehensively set forth reasons for the weight assigned” treating physician opinions).
. This case is distinguishable from George Hyman Construction Co. v. Brooks, 963 F.2d 1532 (D.C.Cir.1992), also cited by Watson to support its refusal to remand, see 569 F.3d at 129, because the error identified in that case was one of law. Specifically, a Benefit Re*166view Board misconstrued a fee statute despite established Supreme Court precedent. On such a point of law, courts owe agencies no deference. See INS v. Cardoza-Fonseca, 480 U.S. 421, 446, 107 S.Ct. 1207, 94 L.Ed.2d 434 (1987) (observing that “pure question[s] of statutory construction” are "for the courts to decide”); George Hyman Constr. Co., 963 F.2d at 1537 n. 3. Moreover, when the court declined to remand the case to see if the Board could explain its fee award consistent with applicable law, see George Hyman Constr. Co., 963 F.2d at 1539 (observing that “no rational factfinder conscientiously applying [the applicable law] could find” the interrelatedness of claims necessary to support challenged award), it was operating in an area, fee awards, routinely addressed in the first instance by federal courts and, unlike conscientious objector applications, requiring neither agency expertise nor difficult assessments of the scope and sincerity of a person’s beliefs.
. Indeed, despite the identification of commission errors, Checkman did not go so far as Watson in ordering the objector’s immediate discharge. Rather, Checkman approved a conditional grant of the writ of habeas corpus, affording DACORB four weeks to conduct new proceedings “shorn” of the identified errors. 469 F.2d at 788.
. In Coates, the court's remand ruling referenced the existence of valid as well as invalid grounds for the challenged decision. See 494 F.2d at 712 ("The proper procedure in such a case where the record evidences alternative grounds, one possibly valid and the other invalid, is to remand the proceedings to the service for reprocessing and for compliance with the requirement of a statement of reasons.”). I understand the reference simply to acknowledge the state of the record in Coates, not to hold that remand depends on a court's own identification in the record of a valid reason for decision. If that had been so, the Fourth Circuit would hardly have concluded that judicial review of the record "to glean ... possible reasons for the denial ... was unnecessary." Id. at 711 (emphasis added).
Nor can Checkman be understood to impose such a condition on remand. It references record evidence of valid and invalid reasons for decision to explain why judicial review must be limited to the stated reasons, for: "where no reasons are given for a denial of CO status, and the record before the selective service officials contains both valid and invalid grounds for the denial, the denial cannot be sustained because it is impossible to know whether the invalid rather than the valid grounds were relied upon.” 469 F.2d at 779 n. 7. Checkman's observation that "a court, if it sustains a decision by recourse to reasons outside those specified, opens the door to an improper substituting of the court’s judgment and evaluation of evidence in place of that of the agency ... with responsibility,” id. at 781 (emphasis added), applies with equal force if a court reverses a decision by recourse to reasons outside those specified. If it is presumptuous for a court to assume that an agency decision was based on reasons identified by the court, see id. at 779 n. 7, 781, so it is presumptuous for a court to assume that it can identify every possible reason for agency decision, particularly in an area informed by agency expertise, see McGee v. United States, 402 U.S. at 486, 91 S.Ct. 1565.
. The concurring opinion identifies this characterization as one of two "factual inaccuracies” because "there is no information in the record about whether Dr. Watson would be required to make final treatment decisions for American soldiers." Ante at [160 n. 5], It proceeds to state that "the dissent's analysis would make all radiologists, if not all doctors, ineligible for discharge as conscientious objectors.” Id. This misstates my position, which identifies Watson’s specialty as a factor properly considered by DACORB in assessing the sincerity of his particular conscientious objector claim, not a determinative factor in this or any other conscientious objector case. Moreover, DACORB would consider the fact with a considerable advantage over the Watson panel, namely, familiarity with how the Army presently uses doctors with specialized medical training, particularly radiologists. In any event, Watson clearly refuses to serve even to the limited extent discussed in text.
The second charge of factual inaccuracy is similarly unwarranted, as explained infra note 18.
. To the extent some reviewing officers noted that Watson’s opposition to caring for wounded soldiers seems at odds with the Hippocratic Oath, Watson is dismissive in light of Watson's non-opposition to other doctors providing necessary treatment. See Watson v. Geren, 569 F.3d at 134. The record provides us with no information as to whether this responsibly construes the Hippocratic Oath. No matter. For tire reasons stated in text, Watson’s argument that it would be immoral for him to treat a wounded soldier, but moral for other doctors to do so, does not bear close scrutiny.
. Only when asked about the possibility of his having "unknowingly” treated military personnel in his civilian practice did Watson indicate that he might have to limit his practice "in the broadest terms to non-members of the militaiy.” Investigation Hearing Memorandum, July 12, 2006, at 2.
. What is remote is the likelihood of any United States hospital acceding to a doctor’s refusal to treat patients in the nation’s armed forces.
. The interviewing chaplain expressed reservations about the sincerity of Watson's claim precisely because he failed to report any actions taken in his civilian life manifesting the beliefs asserted. Watson dismisses these reservations by noting that, although the chaplain cited abortion as an example of a procedure that a doctor committed to the sanctity of life might seek to avoid, the chaplain apparently did not specifically ask Watson about abortion. See Watson v. Geren, 569 F.3d at 134. That seems beside the point, given that the burden of proof was on Watson, and the cited DOD instructions alert objector applicants to the relevancy of civilian conduct consistent with professed beliefs. Cf. Hanna v. Sec'y of the Army, 513 F.3d at 7 (noting that doctor seeking conscientious objector classification sought and obtained leave from hospital to abstain from participation in abortion procedures). In any event, whatever Watson’s position on abortion, the objective fact remains that although he professes an opposition to "weaponizing human beings” so complete that he would refuse even to read the X rays of a wounded American soldier, he took no steps in his civilian practice to ensure that none of his patients were members of the United States armed forces.
. Because the district court and panel opinions quote these submissions at length, there is no need to do so here. See Watson v. Geren, 569 F.3d at 119-25; Watson v. Geren, 483 F.Supp.2d at 229-34. To be sure, Watson cannot be faulted for seeking the assistance of counsel in presenting his application. See Watson v. Geren, 569 F.3d at 134. Nevertheless, in assessing whether he demonstrated a sincere opposition to all war, rather than to specific wars, the Army was reasonably entitled to note differences in how Watson himself explained his beliefs and how they were described in his counseled submissions. Cf. Gillette v. United States, 401 U.S. at 457 (counseling caution that conscientious objection claims not produce different results based on whether applicant is "more articulate, better educated, or better counseled”); United States v. Corliss, 280 F.2d at 815 (observing that where answers on application form "seem more learned than personal” and are prepared with assistance from another person, factor may be considered in assessing scope and sincerity of professed belief).
The concurring opinion identifies purported factual inaccuracy in the characterization of Watson’s submission as counseled, stating that "there is no indication in the record that Watson’s written application was prepared by anyone but Watson himself." Ante at [160 n. 5], The fact that Watson’s 16-page, single-spaced application was submitted by an attorney, however, is certainly some evidence that his written submission was “counseled.” Even without that inference, the inconsistency between Watson’s written application and his hearing testimony, discussed in the ensuing text, is objective evidence on which DACORB could rely in explaining its decision to deny conscientious objector classification.
. For example, in his application for conscientious objector classification, Watson stated that it was as a result of "a culmination of readings, meetings, discussions and other experiences” that his “beliefs” opposing all war "crystallized by early Summer 2005.” Watson App. at 9.
. This precedent prompts me to question Watsons conclusion that the obvious emphasis throughout Watson's written submissions on his opposition to the wars in Afghanistan and Iraq could not rationally support a determination that Watson clearly and convincingly demonstrated only his opposition to those wars, not war generally. See Watson v. Geren, 569 F.3d at 132. Watson complained that the four hearing officers who reached that conclusion took his statements opposing the Afghanistan and Iraq wars out of context. This is an argument properly considered by DACORB, but not by a reviewing court precluded from weighing the evidence. The law plainly permits a decisionmaker to credit evidence in part without accepting it in whole. See, e.g., United States v. Jackson, 778 F.2d 933, 942 (2d Cir.1985). The rule of completeness ensures only that a party does not introduce misleading evidence. See United States v. Kopp, 562 F.3d 141, 144 (2d Cir.2009). It does not limit a factfinder’s discretion in deciding what evidence to credit. Thus, I respectfully reject the idea that a rational fact-finder could not credit Watson’s opposition to the Afghanistan and Iraq wars without also crediting his professed opposition to all wars.
. In the final Allied campaign in Europe, large numbers of wounded American soldiers owed their lives to the extraordinary efforts of conscientious objectors serving as Army medics, a history recounted by Stephen Ambrose in his book Citizen Soldiers in the chapter entitled “Medics, Nurses, and Doctors.” In detailing the extensive readings that informed his professed moral opposition to treating American soldiers, Watson does not indicate familiarity with this history.