SCHWELB, Senior Judge:

On May 5, 2008, following an evidentiary hearing, an Administrative Law Judge (ALJ) of the District of Columbia Department of Employment Services (DOES) issued a Compensation Order in which she held that claimant David H. Combs, formerly a bus driver for WMATA, had sustained only a 1% disability to his right arm as a result of an on-duty accident that occurred on August 24, 1991. Combs appealed from the ALJ’s decision to the Compensation Review Board, which unanimously affirmed the Compensation Order on November 3, 2008.

Combs has asked this court to review the Board’s decision. He contends that the ALJ enred with respect to the facts and the law, and he maintains that he has sustained a 32% disability. Although we are troubled by some aspects of the Compensation Order, we conclude that the ALJ’s dispositive findings are supported by substantial evidence. Accordingly, we affirm.

Combs, who is now 73 years old, began his employment with WMATA in 1980. He remained with WMATA until his “Normal Retirement” (as distinguished from “Disability Retirement”) twenty years later.

Combs testified at the hearing before the ALJ that on August 24, 1991, while he was driving a Metrobus on P Street near Delaware Avenue, S.W., a car ran into the side of his bus, pushing it towards a tree. Combs suffered an injury to his right shoulder and arm. He applied for workers’ compensation, and he received temporary total disability (TTD) payments from August 25, 1991 to September 17, 1991, and temporary partial disability payments (TPD) from April 30,1992 to May 12,1992. Combs made no further application for benefits for more than ten years.

Combs further testified that he was involved in a second accident in 1996. He asserted that on this occasion, he was “rear-ended” by a Cadillac. Combs stated that this accident occurred on Georgia Avenue. Combs asserted, however, that the August 24, 1991 collision on P Street, S.W. was the only accident during his employment with WMATA in which he suffered an injury to his right arm or shoulder.

On February 24, 2003, Combs filed a new “Employee’s Claim Application” for workers’ compensation. In his application, he asserted that he suffered injury to his right shoulder on “09/03/96 at Delaware Avenue, S.W.” He further stated that “[t]he bus I was driving was hit by another vehicle.” Delaware Avenue is in the immediate vicinity of P Street, S.W. It is not near Georgia Avenue.

Combs testified that from 1991 he suffered pain and other difficulties with his right arm and shoulder. In 2003 and 2004, *1007 and on February 2, 2005, Combs was examined or treated by, inter, alia, Sankara R. Kothakota, M.D., 1 V.P. Chandar, M.D., and Michael A. Franchetti, M.D. The reports by all of these physicians specify the year of Combs’ injury as 1996, not 1991. 2

On May 18, 2005, a DOES Claims Examiner held an informal conference with Combs and with counsel for the parties with respect to Combs’ claim that he injured his right arm and shoulder on September 3, 1996. The Claims Examiner concluded, in pertinent part:

Based upon the testimony of the claimant, and my review of the record submitted at the conference, I believe that the claimant sustained job-related right shoulder injury in 1991 (which was not discussed), and in 1996, which is the main claim; however, the claimant did not submit any evidence to support the claim. Moreover, claimant failed to file for workers’ compensation until 2/2/2003 (seven years) after the injury. Consequently, his claim is time-barred by the statute of limitation.

In her “Recommendation” the Claims Examiner stated that “[t]he claim is time-barred.”

After the Claims Examiner concluded that Combs’ claim relating to a 1996 accident was untimely, Combs and his counsel apparently changed their theory as to when Combs sustained his injury. On March 21, 2007, at the request of Combs’ counsel, Dr. Franchetti conducted a second “Independent Medical Examination” of Combs. Dr. Franchetti, an orthopaedic specialist, now concluded that Combs suffered from a 32% right shoulder impairment “as a result of his injuries sustained on August 2k, 1991.” (Emphasis added.) In his report, Dr. Franchetti wrote that

I previously evaluated the patient on February 2, 2005 for right shoulder injury that occurred at work and it was mistakenly reported that this right shoulder injury occurred on September 3, 1996. There was no injury that he sustained to his right shoulder on September 3, 1996 and his work-related injury to his right shoulder occurred on August 24,1991.

Dr. Franchetti provided no explanation of the reason for the change (from 1996 to 1991) in the reported date of the accident. In all other relevant respects, his 2007 report was consistent with his report of February 2, 2005.

On June 12, 2007, at the request of WMATA, Combs was examined by Louis E. Levitt, M.D. Dr. Levitt concluded that

although the records are somewhat scarce in this case and do not show a consistent history of medical treatment, there is enough information from his history and medical records to causally relate complaints of shoulder pain to the 1991 motor vehicle accident. Perhaps he had a chronic tendonitis to the rota-tor cuff that resulted from the original injury.

Dr. Levitt added that “[i]n my opinion, Mr. Combs is entitled to a 20% impairment to the right upper extremity as it relates to the original work injury in 1991.” There is no indication in Dr. Levitt’s report, however, that he was aware of Combs’ claim that he had suffered injury to his right arm and shoulder in 1996.

*1008 II.

In a nine-page Compensation Order, the ALJ resolved substantially all of the contested issues adversely to Combs, and she concluded that Combs suffered a 1% disability to his light arm as a result of the 1991 accident. Specifically, the ALJ found that Combs was involved in a motor vehicle accident on P Street, S.W. on August 24, 1991 and a second motor vehicle accident on Delaware Avenue, S.W. on September 3, 1996. The ALJ wrote that “Claimant’s testimony that he was not involved in a motor vehicle accident in 1996 is not credible.” The ALJ repeated in several places in the Compensation Order that Combs had denied being in an accident in 1996, and on each occasion she reiterated her finding that Combs’ testimony was incredible. Indeed, in footnote 4 to her order, the ALJ acknowledged Combs’ testimony on cross-examination that he had been involved in an accident in 1996, but she continued to assert, later in her order, that Combs was involved in a motor vehicle accident in 1996, and that she “reject[ed] Claimant’s testimony to the contrary.” In any event, the ALJ found that the injury suffered by Combs which led to his May 2004 surgery was incurred in 1996.

The ALJ was likewise unimpressed with Dr. Franchetti’s 2007 report, which had been submitted to her on Combs’ behalf. After noting that Dr. Franchetti had written in 2005 that Combs suffered his injury in 1996 but wrote in 2007 that the accident occurred in 1991, the ALJ did not mince her words:

Just as I do not find Claimant’s testimony that there was no 1996 accident credible, I cannot find Dr. Franchetti’s opinions credible. I reject them in their entirety.

(Emphasis added.)

Turning to the opinion of Dr. Levitt, the ALJ wrote as follows:

Dr. Levitt opines Claimant has sustained a 20% permanent impairment to his right arm as a result of Claimant’s limited range of motion, weakness to the supraspinatus tendon, acromioplasty, and the five (5) subjective factors, but Dr. Levitt relied upon “[scarce]” and “incomplete” records (which may or may not have included the medical records detailed above in regards to Claimant’s 1996 injury) and a determination that “[pjerhaps he had a chronic tendonitis to the rotator cuff that resulted from the original injury.” Because of Dr. Levitt’s lack of information and lack of definitiveness, I reject his opinion regarding Claimant’s permanent impairment as well.

(Emphasis added in Compensation Order.) Noting on the basis of the medical reports in 1991 and 1992 that Combs’ injury incurred as a result of the 1991 accident was not then deemed especially serious, that Combs had received “minimal medical treatment approximately fifteen (15) years ago,” that he continued working until 2000, that he applied for “Normal Retirement” rather than “Disability Retirement,” and that workers’ compensation benefits are available only for injuries that result in disability, the ALJ concluded that Combs qualified only for a 1% PPD award for the 1991 injury to his right arm and shoulder. The Compensation Review Board affirmed, holding that the ALJ’s “thorough, well-reasoned decision” was supported by substantial evidence.

III.

In this case, the ALJ’s findings of fact have been sustained by the Compensation Review Board, and our review of these findings is limited to determining whether there is substantial evidence to *1009 support them. Zhang v. District of Columbia Dep’t of Consumer and Regulatory Affairs, 834 A.2d 97, 101 (D.C.2003). “Substantial evidence is more than a mere scintilla. It means such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” King v. District of Columbia Dep’t of Employment Servs., 560 A.2d 1067, 1072 (D.C.1989) (quoting Consolidated Edison Co. v. NLRB, 305 U.S. 197, 229, 59 S.Ct. 206, 83 L.Ed. 126 (1938)); see also Marriott Int’l v. District of Columbia Dep’t of Employment Servs., 834 A.2d 882, 885 (D.C.2003). Significantly for present purposes, we must uphold the agency action if it is supported by substantial evidence even if there is substantial evidence to support a contrary conclusion. Marriott Int’l, 834 A.2d at 885. Although the Compensation Review Board expressed no reservations regarding the ALJ’s findings, we perceive some of them to be problematical.

First, the ALJ repeatedly characterized Combs’ testimony as incredible because, according to the ALJ, Combs falsely denied being involved in an accident in 1996. In fact, Combs acknowledged that he was in an accident during that year, but he placed it on Georgia Avenue. Although the record discloses beyond peradventure that Combs contradicted himself as to the date of the accident that caused his injury, the ALJ’s repeated intimation that he falsely denied that he was in an accident in 1996 is not supported by the record. The ALJ found, in effect, that Combs lied by testifying to something that in fact Combs did not say.

Second, the ALJ “rejected in their entirety” Dr. Franchetti’s opinions (apparently including his medical diagnosis) solely because of the discrepancy between his 2005 and 2007 reports with respect to the date of Combs’ injury. Since Dr. Fran-chetti did not examine Combs until fourteen years after the first accident, and since he could have had no personal knowledge as to when the accident occurred, he was obliged to rely on what he was told by his patient. Therefore, the only reasonable inference from the record before us is that the change of the date of Combs’ accident from 1996 in Dr. Franchetti’s first report to 1991 in his second was based on information from Combs or his counsel to the effect that the 1996 date was incorrect. Dr. Franchetti did not testify, and the ALJ had no opportunity to assess his demeanor as a witness. 3 In our view, the ALJ’s *1010 effective characterization of the doctor as altogether unworthy of belief was, at least, overstated.

Third, the ALJ found that Combs was involved in two separate accidents in southwest Washington, D.C., one in 1991, the second in 1996, and that in each of them, Combs incurred an injury to his right arm and shoulder. She found that the first accident occurred on P Street, S.W. and the second on Delaware Avenue, S.W. Combs testified, and a glance at a street map confirms, that P Street and Delaware Avenue are very near each other, and close to Combs’ stated destination of Half and 0 Streets, S.W. on the day of the 1991 accident. Under these circumstances, it would be more than a trifle surprising if Combs had been involved in two separate accidents five years apart in almost the same location and had suffered injuries to the same part of his body. See, e.g., Poulnot v. District of Columbia, 608 A.2d 134, 139 (D.C.1992).

Finally, if Combs in fact incurred an accident in southwest Washington, D.C. in 1996 which ultimately led to his surgery and caused his alleged disability, it would appear somewhat odd that WMATA apparently had no record of such an event. At the hearing, WMATA introduced medical records relating to the 1991 injury, but although Combs remained in WMATA’s employ until 2000, WMATA produced no documentation of the 1996 accident which allegedly caused Combs’ partial disability.

The foregoing problems with the conclusion reached by the ALJ and sustained by the Compensation Review Board suggest, at least, the possibility that a dozen years after the 1991 accident, Combs was confused as to when it occurred, that he wrote the wrong date on his 2003 Claim Application, and that he likewise provided the wrong date to various physicians in the years that followed. That possibility, however, does not and cannot permit us to hold that the ALJ’s dispositive finding — namely, that the accident which resulted in surgery by Dr. Kothakota and in Combs’ other difficulties occurred in 1996, not in 1991, and that there is no evidence of an appreciable permanent disability arising out of the 1991 accident — lacks support by substantial evidence. After all, it was Combs who applied for workers’ compensation in 2003, claiming that the injury to his arm and shoulder occurred in 1996. It was Combs who told at least three examining and treating physicians that he was injured in 1996. Combs has a twelfth grade education, and an impartial finder of fact could reasonably find that, on such an important matter, Combs would not repeatedly make such a major and devastating error. If, indeed, time after time, Combs misrecol-lected the date of his accident by five years, 4 then even the most sympathetic *1011 appellate tribunal cannot rescue him from the unfortunate consequences of his mistake. Combs not only admitted, but asserted to an agency of the government, that he had incurred an injury to his right arm and shoulder in 1996 for which he was entitled to compensation. We cannot hold that a finding based on his own oft-reiterated assertion is not supported by substantial evidence. 5

Moreover, the sequence of events is significant, and it provides support for the ALJ’s decision. From 2003, when he filed his second claim application, through at least February 2005, Combs represented to the agency and to his physicians that the accident which caused his injury occurred in 1996. However, on May 18, 2005, Combs and his attorney presumably learned that his claim of a 1996 accident would probably be dismissed as untimely, in conformity with the Claims Examiner’s recommendation. 6 It was after that recommendation that he or his counsel evidently represented to Dr. Franchetti and Dr. Levitt that the injury was suffered in a 1991 accident, as to which he had made a timely claim.

Affirmed.' 7

.Combs claims that the ALJ's findings that he suffered only a 1% disability necessarily lacked support by substantial evidence because Combs produced medical evidence and WMATA provided none. But see Golding-Alleyne v. District of Columbia Dep't of Employment Servs., 980 A.2d 1209, 1214 (D.C.2009) (rejecting the claimant's contention that where the employer presents no medical expert of its own, the ALJ is bound to accept the opinion of the treating physician). Combs also asserts that it was error for the ALJ to rely, in part, on the fact that Combs elected “Normal Retirement” rather than “Disability Retirement.”

These contentions provide no support for Combs’ position. In light of the ALJ’s finding sustained by this court, that Combs was operated on and treated for injuries sustained in a 1996 accident, and because the claim for the 1996 accident was time-barred and not pursued, there is ample support in the record for a finding that any permanent partial disability caused by the 1991 accident was minimal especially since Combs’ treatment following *1012 that accident ended in 1992. Combs made no further claim for more than ten years.

. Dr. Kothakota performed an arthroscopic debridement of Combs’ right shoulder in May 2004.

. As noted in the text, infra, Dr. Franchetti also wrote a second report in 2007, and on this occasion he stated that the accident occurred in 1991.

. We note in this regard the Compensation Review Board's statement, in affirming the Compensation Order, that

In this jurisdiction, it is well settled that where credibility questions are involved, "the factfinding of the hearing examiner is entitled to great weight,” since the hearing examiner is in the best position to observe the demeanor of witnesses. Washington Metro. Area Transit Auth. v. District of Columbia Dep't of Employment Servs., 683 A.2d 470, 477 (1996); Dell v. District of Columbia Dep’t of Employment Servs., 499 A.2d 102 (1985).

(Emphasis added.) We agree that the general nde was correctly articulated by the Board. In many and even most cases, deference by an appellate tribunal to the assessment of credibility by the trier of fact is, for obvious reasons, obligatory. As we stated in Morris v. United States, 728 A.2d 1210, 1215 (D.C.1999),

The judge had a front-row seat as the testimony unfolded. We, on the other hand, are limited to a paper transcript which, while capturing the words of a case, may often miss its heart and soul. Cf. In re S.G., 581 A.2d 771, 774-75 (D.C.1990). Indeed, as Judge Jerome Frank has recognized,

[a] stenographic transcript correct in every detail fails to reproduce tones of voice and hesitations of speech that often make a sentence mean the reverse of what the words signify. The best and most accurate record is like a dehydrated peach; it has neither the substance nor the flavor of the fruit before it was dried.

*1010 Broadcast Music Inc. v. Havana Madrid Rest. Corp., 175 F.2d 77, 80 (2d Cir.1949) (quoting Ulman, The Judge Takes The Stand 267 (1933)).

In this case, however, the contrast so effectively described by Judge Frank simply does not exist, at least as to Dr. Franchetti. The doctor did not testify, and this court has access to precisely the same documentary evidence as were available to the ALJ, namely, Dr. Franchetti's two reports. Although the "clearly erroneous” rule applies to the findings of the trier of fact with respect to documentary evidence as well as oral evidence, see Super. Ct. Civ. R. 52(a); American Sec. Bank v. American Motorists Ins. Co., 538 A.2d 736, 739-41 (D.C.1988); see also Anderson v. City of Bessemer City, N.C., 470 U.S. 564, 574-75, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985), this is not a case in which, as to Dr. Franchetti, the ALJ's peach was appreciably juicier than the appellate court’s dehydrated one.

. It should be noted that it was on May 19, 2005 that a DOES claims examiner recommended that a claim based on a 1996 accident be dismissed as untimely. It was after *1011 that date that Dr. Franchetti and Dr. Levitt attributed Combs' medical problems to the 1991 accident, and not to one that allegedly occurred in 1996.

. Combs asserts that the 1% disability finding is not supported by substantial evidence, and indeed, the Compensation Order does not illuminate how the ALJ arrived at that precise figure. The ALJ found, however, that the PPD allegedly suffered by Combs was attributable to a 1996 accident (as to which no claim is before us) and not to the 1991 accident. The comparatively minor treatment received by Combs in 1991 and 1992 is consistent with any permanent disability resulting from the 1991 accident being very slight. The employer does not challenge the finding of a 1% PPD as being too high. Accordingly, we discern no basis for reversal of the 1% disability finding.

. The statute of limitations for workers’ compensation claims is one year. D.C.Code § 32-1514 (2001). The claim for the 1996 accident was made in 2003, and was therefore time-barred.

A claim based on the 1991 accident, on the other hand, was timely. Because Combs had made a claim for TTD within a year of the 1991 accident, he was not precluded from requesting compensation for TPD in 2003, for the one-year limitation period does not apply in this situation. See, e.g., Capitol Hill Hosp. v. District of Columbia Dep’t of Employment Servs., 726 A.2d 682, 684-86 (D.C.1999) (construing earlier version of statute and holding that an initial claim for TTD or TPD preserves a subsequent claim for PPD). It was therefore potentially to Combs' substantial advantage, after May 18, 2005, to attribute his injury to the 1991 accident rather than to one that occurred in 1996.