OPINION

Carl C. Thorn appeals an order of the Benefits Review Board (BRB) affirming the denial of his claim for black lung benefits. We remand for reconsideration.

I.

Carl Thorn worked in the coal mines for over twenty years. His last jobs were very heavy labor — laying track and loading coal— and his last employer was the respondent, Itmann Coal Company.

Thorn’s final day in the mines was July 27, 1978. He filed this claim for black lung benefits on November 14, 1979. Itmann Coal, the responsible operator, contested an initial finding of eligibility. Eight years later the claim was heard by an administrative law judge (ALJ).

The record before the ALJ was primarily a contest of physicians’ opinions. There were no valid pulmonary function studies, and no blood gas tests conducted after exercise. ^[1] Every physician agrees that Thorn suffers from at least simple pneumoconiosis. Thorn had a heart attack in 1978 and suffers from coronary artery disease. He is also mildly mentally retarded and has only a fourth-grade education. Though he has never smoked, Thorn suffers from a chronic cough, dyspnea, and shortness of breath on exertion. No one disputes that Thorn is totally disabled from performing his past work.

Dr. Gary Craft examined Thorn on March 13,1979. He found that Thorn suffered from simple pneumoconiosis, but that “he had good breathing capacity and does not have any major disabilities secondary to his chronic Lung disease.” These findings were based on both a physical examination and an invalid pulmonary function study: “... the patient made no effort at all. However, what results we could obtain indicate that he had good breathing capacity at rest.”

On June 13, 1980, Dr. A.R. Piracha examined Thorn. He found that Thorn “does not[*715] show evidence of any significant pulmonary insufficiency.”

Dr. Shawn Chillag examined Thorn on January 21, 1981. He attempted to perform pulmonary function tests, but Thorn could not be made to understand how to perform the test. Chillag concluded that Thorn does have simple pneumoconiosis, but could not venture an opinion whether Thorn had any pulmonary impairment.

On May 6, 1981, Dr. George 0. Kress issued a consultative report based on. the examinations and testing performed to date. He lamented that all pulmonary function tests had been invalid. Though a blood gas study performed by Dr. Chillag showed a “mild degree of resting hypoxemia,” Dr. Kress postulated that “this might well be explained on the basis of hypoventilation.” Dr. Kress concluded:

It is my opinion that there is no objective evidence available to indicate that this man has any pulmonary or respiratory impairment. It is unfortunate that this gentleman has not been able to cooperate on the attempted ventilatory studies to gain additional objective evidence for evaluating his pulmonary function. There certainly is no evidence to indicate that he is permanently and totally disabled as the result of pulmonary or respiratory problems, either pneu-moconiosis or any other respiratory condition. He is probably disabled, however, as the result of his arteriosclerotic cardiovascular disease but this, of course, has no relationship to pneumoconiosis or to any primary respiratory condition. There simply is no objective evidence available to indicate that any disability he has is caused either in whole or in part by pneumoconio-sis. As has been pointed out simple coal workers’ pneumoconiosis does not as a rule produce total disability.

On September 28, 1988, surgery was performed to repair an eventration of Thorn’s left hemidiaphragm. In July, 1985, Thorn spent three days in the hospital because of chest pain. His attending physician, Dr. Barit, diagnosed and treated Thorn for chronic obstructive pulmonary disease, pleu-ritis, and acute coronary insufficiency.

In the fall of 1986, there was a flurry of examinations and reports relating to Thorn’s long-dormant black lung claim. On September 24, 1986, Thorn was examined by Dr. Jose Floresca. In conjunction with this examination, an x-ray was taken, and blood gas and pulmonary function tests were performed. Dr. Maurice Bassali read the x-ray as showing arteriosclerotic heart disease and “diffuse interstitial lung disease compatible with complicated pneumoconiosis category.A superimposed upon pneumoconiosis type q/t with profusion of ]é affecting all lung zones.” Dr. Bassali also noted surgical clips in the left hilum and a small density in the upper left lung that could represent progressive massive fibrosis or malignancy. Without earlier x-ray films for comparison and a complete history, Dr. Bassali would not give a definite opinion concerning the nature of the left lung density.

The resting blood gas test conducted by Dr. Floresca showed acidosis and moderate to severe hypoxemia (pH of 7.34, p02 of 63.2, and oxygen saturation of 90%). A pulmonary function test (considered invalid by all reviewers) performed that same day yielded qualifying values. Based on his examination of Thorn and the laboratory results, Dr. Floresca diagnosed severe chronic obstructive pulmonary disease, “far advanced” coal worker’s pneumoconiosis, and hypertension. He opined that, as a result of coal dust exposure, Thorn was totally and permanently disabled.

A week later, on October 2, 1986, Thorn was seen by Dr. Naeem Qazi. Dr. Qazi noted decreased breath sounds and rhonchi, and his clinical impression was chronic obstructive pulmonary disease possibly secondary to pneumoconiosis. Dr. Qazi then examined the laboratory results obtained by Dr. Floresca and concluded that Thorn has moderate chronic obstructive pulmonary disease. In a supplemental report, Dr. Qazi noted that Thorn’s pulmonary function studies and x-rays were “compatible with complicated pneumoconiosis,” and he opined that Thorn was “markedly impaired” and should not be employed in coal mining or any other comparable job, because any such exertion “would further increase his bronchospasm[s] and[*716] cause more shortness of breath and dyspnea and [could] cause further injury to his lungs.”

A few weeks later, on October 20, 1986, Dr. Mario Cardona examined Thorn, obtained another x-ray, and attempted to perform a pulmonary function study. Thorn became confused and was unable to perform the FVC part of the test, though he completed the MW satisfactorily. The x-ray was read by Dr. DeRamos as pneumoconiosis, type q/q, with profusion 2/2. DeRamos also noted surgical clips and a “rounded density” in the left lower lung. Like Dr. Bassali before him, DeRamos thought attention should be paid to this density, but he felt that it could represent post-surgical changes.

In reliance on his examination and DeRa-mos’ report, Dr. Cardona diagnosed chronic obstructive pulmonary disease secondary to pneumoconiosis, and he opined that Thorn was totally and permanently disabled because of the effects of occupational exposure to coal dust.

On December 17, 1986, Dr. Kress supplemented his 1981 consultative report in light of the new medical evidence. He reiterated that simple pneumoconiosis “does not, as a rule, produce total disability,” and he summarized his 1981 opinion: “the information available did not indicate that this man would be considered impaired to a degree from a strictly respiratory standpoint that he would be unable to do his former coal mine work.” Dr. Kress again noted that there were no valid pulmonary function tests, though he assembled a composite result from the “best results of his multiple function studies,” from which he thought “it is possible to conclude that these function studies do not demonstrate evidence of any significant ventilatory impairment.” Though conceding that Thorn ‘.‘has the symptoms and signs of obstructive lung disease,” Kress concluded, “I still do not believe the available objective evidence indicates that this man has a significant pulmonary impairment. Certainly, these objective test results, where valid, would indicate that from a respiratory standpoint, he really retains the capacity for his former coal mine work.” Dr. Joseph Renn reviewed the medical evidence developed from 1978 through Dr. Car-dona’s examination, and offered a consultative report dated December 19, 1986. From a review of the tracings accompanying the seven pulmonary function tests in the record, Renn concluded that “there were many deficiencies of quality control rendering [all of] them invalid for accurate interpretation or for the derivation of significant data from which to assess his true ventilatory function.” Notwithstanding these defects, Renn culled the MW results from the March 13, 1979, and October 20, 1986, tests and opined that they “imply excellent ventilatory reserve.” Renn noted that two of three resting blood gas studies showed “metabolic acidosis” ^[2] and “moderate hypoxemia,” and suggested that the “incongruity” among the results should be “better defined with arterial cannulation and exercise studies.”

From the examining physician’s reports, invalid pulmonary function tests, and incongruous resting blood gases, Dr. Renn concluded that Thorn has no ventilatory impairment. “When considering only his respiratory system, it is with a reasonable degree of medical certainty he is not totally and permanently impaired to the extent he would be unable to perform heavy manual labor such as that found in and around the underground coal mines.”

Dr. George Zaldivar examined Thorn on December 17, 1986. He took an x-ray and reviewed five previous x-rays in making his report. He attempted to perform a pulmonary function test, but, like all others before him, was foiled by Thorn’s poor effort and inability to understand instructions. A resting blood gas was tolerably normal for Thorn’s age (pH of 7.44, pO2 of 75, and pCO2 of 33).

From comparison of recent x-rays with ones taken before Thorn’s 1983 surgery, Zal-[*717] divar concluded that the suspicious area on Thorn’s left lung was a pleural shadow resulting from the surgery rather than evidence of massive fibrosis. He based this conclusion on the sudden appearance of the area on post-surgery films and because the shadow’s position did not vary with the degree of rotation or inspiration in the films.

Dr. Zaldivar concluded that Thorn did have simple pneumoconiosis, and angina that appeared to be disabling, but no “significant pulmonary impairment at all.”

Dr. Barit, Thorn’s treating physician, reported on December 31,1986, that Thorn had suffered from pneumoconiosis for years, and had “increasing shortness of breath on slight exertion, and at times, even at rest.” Barit concluded, “this man could no longer engage in any gainful job, and is disabled throughout life, due to his occupational Pneumoconiosis and its residuals.”

Dr. Zaldivar’s deposition was taken on January 12,1987. He stated that the pulmonary function test he performed was invalid, but that, because Thorn had given submaxi-mal effort, “it would imply that his pulmonary function study should be normal if he had cooperated.”

Thus, four examining physicians (Floresca, Qazi, Cardona, and Barit), including Thorn’s treating physician, ^[3] opined that Thorn is totally disabled by pneumoconiosis. Three examining (Craft, Piracha, and Zaldivar) and two consulting (Kress and Renn) physicians thought otherwise.

The ALJ found that Thorn was entitled to the interim presumption of total disability due to pneumoconiosis under 20 C.F.R. § 727.203(a). However, the ALJ found that the presumption was rebutted under § 727.-203(b)(3) — the evidence showed that Thorn’s disability did not arise, in whole or in part, from coal mine employment. In reaching this conclusion, the ALJ gave “decisive weight” to the opinions of Drs. Zaldivar, Renn, and Kress, because of their superior qualifications and because “their reports, which are among the most recent, are also the most comprehensive, take into account all of the clinical testing obtained in the course of the Claimant’s many examinations, and are consistent with the reported results.”

Thorn sought review before the BRB. His petition for review challenged both the AL J’s finding of no complicated pneumoconiosis and his finding of rebuttal under § 727.203(b)(3). ^[4] Thorn briefed the complicated pneumoconio-sis issue more thoroughly, and the employer asserted that he had waived the (b)(3) issue. The BRB took no note of the employer’s argument, and instead addressed and rejected both of Thorn’s arguments on their merits.

II.

The employer renews its assertion that Thorn waived his challenge to the finding of (b)(3) rebuttal because he merely noted it as an issue presented in his petition for review to the BRB and did not brief it extensively. In the face of the same argument from the employer, the BRB gave the (b)(3) issue plenary consideration.

Our rule on administrative waiver of claims later asserted in court is “flexible.” Rana v. United States, 812 F.2d 887, 889 n. 9 (4th Cir.1987). Had the BRB ignored Thorn’s (b)(3) argument notwithstanding its having been asserted in his petition for review, or had explicitly found a waiver, we would perhaps face a difficult question. Here, though, with the BRB’s full consideration of the issue, the policy reasons behind administrative waiver — preserving the requirement of exhaustion of remedies and respect for the agency’s expertise — are simply not present.

[*718] III.

A.

Rebuttal of the interim presumption under § 727.203(b)(3) is difficult — substantial evidence must be adduced that “rules out” any causal connection between the miner’s coal mine employment and his disability. Bethlehem Mines Corp. v. Massey, 736 F.2d 120 (4th Cir.1984). ^[5] We will have more to say about Massey later.

We review the ALJ’s application of the law de novo, and we must affirm factual findings if they are supported by substantial evidence. Wilson v. Benefits Review Board, 748 F.2d 198 (4th Cir.1984).

B.

In Adkins v. Director, OWCP, 958 F.2d 49 (4th Cir.1992), we struck down, as arbitrary and irrational, the practice of blindly ascribing more weight to the most recent evidence. Adkins bears directly on this case, because the ALJ cited the recency of Zaldivar, Kress, and Renn’s opinions in giving them decisive weight. In fairness to the ALJ, he did mention proper criteria for weighing evidence — expertise, completeness of explanation, etc. — but, from the opinion, there is no way to know how much of a role “later evidence is better” had in the final outcome.

The employer argues that a recent physician’s opinion can be more reliable than an old one. Of course it can be. There may be new or additional evidence developed that discredits an earlier opinion; a comparison of medical reports and tests over a long period may conceivably provide a physician with a better perspective than the pioneer examiner. The reasons for crediting such an opinion could be perfectly rational.

But “recency,” in and of itself, is not one of those reasons. A bare appeal to “recency” is an abdication of rational decisionmaking. In this case, the ALJ’s invocation of recency is particularly barren. Four, perhaps five, ^[6] physicians examined Thorn in a space of eighty-four days between September 24 and December 17, 1986. All of these examining physicians, save Dr. Zaldivar, found Thorn totally disabled by pneumoconiosis. The consulting reports of Drs. Renn and Kress were, of course, generated in response to the opinions of the examining physicians, and they naturally postdated the actual examinations. To describe a relevant medical opinion as “among the most recent” is simply to describe this record. Dr. Qazi’s report was “among the most recent.” So were Dr. Car-dona’s and Dr. Floresca’s. Dr. Barit’s opinion would have been last of all, except that the employer duplicated Dr. Zaldivar’s opinion by taking his deposition twelve days later.

Thus, the ALJ committed an analytical error in the course of making the crucial factual determination in the ease — which group of physicians to believe. Keeping in mind, as we must in any black lung case, that “[t]he Act embodies the principle that doubt is to be resolved in favor of the claimant,” Mullins Coal Co. v. Director, OWCP, 484 U.S. 135, 156 n. 29, 108 S.Ct. 427, 439 n. 29, 98 L.Ed.2d 450 (1987), we conclude below that the ALJ’s error is not harmless, because there are several gross deficiencies in the employer’s proof.

IV.

A.

As we have stated several times above, there are no valid pulmonary function studies[*719] in the record, and those physicians who relied in part on these invalid results to find Thorn disabled can be fairly criticized for their error. All three of the physicians credited by the ALJ trumpeted the invalidity of the results. However, in almost the same breath, all three culled selective (i.e. the highest) numbers from various tests to support an opinion that Thorn had no ventilatory impairment. Dr. Zaldivar even went so far as to predict that a valid pulmonary function study, if one could be performed, would yield “normal” results.

We have criticized the practice of routinely ascribing greatest weight to the highest results among valid studies. Greer v. Director, OWCP, 940 F.2d 88, 90-91 (4th Cir.1991). Imputing selective reliability to the highest results of invalid tests strikes us as highly speculative. Predicting the result of a hypothetical test is rankly so.

B.

We quoted Dr. Kress’ reports at length in Part I above so that the viewpoint from which he rendered his opinions would be apparent. Dr. Kress’ stated credo is that simple pneumoconiosis does not “as a rule” cause total disability. Therefore, because, in Dr. Kress’ opinion, Thorn did not present “objective” evidence of total disability, Dr. Kress would apply his presumption that simple pneumoconiosis has not resulted in total disability. But the law — the interim presumption and rebuttal provisions of 20 C.F.R. § 727.203 — is precisely the opposite. Simple pneumoconiosis, combined with ten years of coal mine employment, is presumed to be totally disabling (§ 727.203(a)), and the burden is on the respondent to rebut the presumption by producing persuasive evidence that rules out any causal connection between pneumoconiosis and total disability. § 727.203(b)(3); Massey, 736 F.2d at 123. Dr. Kress may disagree with the clinical accuracy of the interim presumption, but his mere disagreement is not rebuttal.

It is perfectly reasonable to discredit an expert’s conclusion with regard to whether a condition defined by statute and regulation does or does not exist when that expert bases his conclusion on a premise fundamentally at odds with the statutory and regulatory scheme.

Penn Allegheny Coal Co. v. Mercatell, 878 F.2d 106, 109-110 (3rd Cir.1989). Several other circuits have recognized that a physician’s opinion based on a premise antithetical to the Black Lung Benefits Act is not probative. Robbins v. Jim Walter Resources, Inc., 898 F.2d 1478, 1482 (11th Cir.1990); Wetherill v. Director, OWCP, 812 F.2d 376, 382 (7th Cir.1987); Kaiser Steel Corp. v. Director, OWCP, 757 F.2d 1078, 1083 (10th Cir.1985) (doctor’s opinion that simple pneu-moconiosis is not disabling “is in direct conflict with a basic premise of the Act”). Having inverted the presumption, Dr. Kress then bases his opinion on the claimant’s failure to persuade him. True rebuttal is more than merely questioning or criticizing the quality of the claimant’s proof. Black Diamond Coal Mining Co. v. Benefits Review Board, 758 F.2d 1532 (11th Cir.1985).

C.

A further infirmity in both Dr. Renn and Dr. Kress’ consulting reports is their failure to address the issue at hand. Neither Renn nor Kress ask the relevant question, and their ultimate opinions are stated in terms we have rejected before. Dr. Renn’s opinion on disability expressly considers “only [Thorn’s] respiratory system.” Dr. Kress’ 1986 disability opinion is likewise stated from a “respiratory standpoint,” and he characterized his own 1981 opinion as being from a “strictly respiratory standpoint.” These opinions are not helpful because a claimant need not prove that pneumoconiosis is a self-sufficient cause of disability. Indeed, Massey’s central holding rejects opinions like Dr. Renn’s and Dr. Kress’ as competent for (b)(3) rebuttal.

The “in whole or in part” language of section 727.203(b)(3) simply reflects the current uncertain state of medical knowledge concerning the exact consequences of prolonged exposure to coal dust. The regulation does not attempt to articulate a precise medical standard of proof.

Rather, its very flexibility represents a considered legal judgment that the broad remedial purposes of the Black Lung Act[*720] cannot be achieved if claimants are held to a standard of proof approaching medical certitude.... In cases in which the combined effects of several diseases disable the miner, the employer obviously cannot meet its burden of proof by focusing solely on the disabling potential of the miner’s pneumoconiosis.

736 F.2d at 124. ^[7] In sum, the consulting-reports of Doctors Renn and Kress are not probative evidence on the issue of (b)(3) rebuttal.

We remand the claim for reconsideration.

REMANDED.