The opinion
MURNAGHAN, Circuit Judge,
concurring:
Here we had a sympathetic plaintiffs case where the evidentiary material submitted to the trial judge was indicative of either of two possible results. There were facts which would adequately support a finding for the plaintiff. There also, however, were facts adequately to support a finding for the defendant. Both the majority and dissenters on en banc appeal appear to accept the rule of Anderson v. City of Bessemer City, 470 U.S. 564 , 105 S.Ct. 1504 , 84 L.Ed.2d 518 (1985). That case of binding authority made crystal clear that, in a court trial, a district court’s determination which the appellate court finds uncongenial must, nevertheless, be affirmed as to an issue of fact unless clearly erroneous.
The en banc dissenters (two of twelve appellate judges, that is two of thirteen judges who have heard the case if the district judge is included) allowed themselves to be swayed by the apparent uncon-geniality. By doing so, they gave lip service to the rule but then, by proceeding to find the district judge to have been clearly erroneous where there nonetheless was adequate evidence to support his finding, actually emasculated the ruling. If the holding of the majority at the panel level (the dissenters on en banc appeal) had prevailed, we would be met with a departure from the spirit of the ruling in the Anderson v. City of Bessemer City case, which we would have found difficult to correct in view of the precedential value of such a holding.
Chief Judge Winter seeks to dismiss the case as merely a judgment call on disputed facts. He regrets the time consumed in rehearing the case en banc and the waste of valuable court resources in a sitting requiring the presence of all the judges when otherwise other cases could be heard. However, this was not merely one involving a disagreement as to which of two disputed factual findings should prevail. Even if every appellate judge found more acceptable the version of the facts as they appear to the dissenters, they were, nevertheless, bound to accept the contrary findings of the district judge inasmuch as they had sufficient evidentiary support.
To me it is somewhat disturbing to imagine handing down an incorrect and, therefore, unjust decision, but I recognize that there can be cases of that kind which do not merit en banc reconsideration. However, one reason for en banc hearing is the security and maintenance of uniformity in *75 our decisions. See FRAP 35(a). That requirement, as use of the word “or” indicates, is disjunctive of the grounds for an en banc hearing or rehearing “when the proceeding involves a question of exceptional importance.” The holding in a Supreme Court case is one as to which we must strive to secure or maintain uniformity. The case presented here had to be heard en banc to secure or maintain such uniformity, and I regard it as important that we do not encourage the ill-conceived notion of “achieving justice” for the sympathetic party by ignoring the factual evidence and the district court judge’s decision. 1
I have already indicated my concurrence with the fine opinion produced by Judge Haynsworth. I concur specially, not to suggest that my enthusiastic acceptance of what Judge Haynsworth has written is in any way diminished, but to present a view that it was necessary and important that we take the trouble to hear the case en banc and that Judge Haynsworth trouble himself to write and to write correctly.
Judge Widener has authorized me to say that he concurs in my opinion. In doing so, he does not intend to reduce in any way the strength and enthusiasm of his concurrence in Judge Haynsworth’s opinion for the en banc majority.
. One of the court's techniques for promoting justice is randomly to select panel members to hear cases, customarily in groups of three. Here the technique has led us into an unfortunate bog — a veritable Slough of Despond. When the judges were originally assigned for panel consideration of the case, there was, of course, no way of knowing prior to argument how the panel members, or the non-panel members, of the court would vote. However, in retrospect, it now appears that there was only an extraordinary concatenation of circumstances which led to the problem Judge Haynsworth’s opinion en banc addresses. If any judge, other than either of the two members making up the panel majority, had been selected, the decision below would have been correct and, in all probability, would never have been successfully petitioned to rehearing en banc. If either one of the panel majority had been replaced by any of the other nine judges (t.e., the full complement of active circuit judges not counting Judge Hall or Judge Sprouse) in regular active service, he would not have voted at the panel level and his replacement would have sided with Judge Haynsworth, converting his dissent at the panel level to the majority opinion.
It appears that with twelve judges, namely, the eleven active circuit judges and one senior circuit judge, available for random selection to form the panel, there were 220 separate possible panels. Of those 220, ten only could be composed of both Judge Hall and Judge Sprouse.
The permutations of twelve judges organized into panels of three judges yields 1320 distinct possible panels. However, for our purposes a panel of Judge Hall, Judge Sprouse and Judge Haynsworth would be the same as Judge Hayns-worth, Judge Sprouse and Judge Hall. Since there are six such irrelevant changes of order, to arrive at the correct number of combinations of randomly selected different panels we must divide by six. (The six different panels which are identical for these purposes are: Judge Hall, Judge Sprouse, Judge Haynsworth; Judge Hall, Judge Haynsworth, Judge Sprouse; Judge Haynsworth, Judge Hall, Judge Sprouse; Judge Haynsworth, Judge Sprouse, Judge Hall; Judge Sprouse, Judge Hall, Judge Haynsworth; Judge Sprouse, Judge Haynsworth, Judge Hall). That yields 220 (1320/6) as the number of possible panels. Using the same technique we find that ten different panels of the 220 possible could include in their makeup both Judge Hall and Judge Sprouse. The likelihood, therefore, of the wrong decision here being reached at the panel level was 4.54% (10/220). I submit that to leave the panel decision undisturbed would embody an error of 4.54% in a system which we have introduced to bring about success in the quest to achieve full and fair justice. It is, in short, necessary to say that the oft repeated rule that judges on the Fourth Circuit are fungible, or at least fungible enough to insure the correct result, did not work in the panel selection here.
The defendant's counsel here would have much to complain of since 95,46 times out of 100, he would and should have won. (He should have won 100% of the time.) It is distasteful to me to see the work of the court take on the guise of a roulette wheel operated by chance. I stick emphatically by my suggestion that there has not been security and maintenance of uniformity in our decisions. It may even be argued that we have here a proceeding which involves a question of exceptional importance.