The implication is that the large absolute number of witnesses supporting the conspiracist position is compelling.
The logic of their position is apparently something like this. Let's assume that the probability of any single randomly chosen witness being wrong is .2.
Then what is the probability of two randomly chosen witnesses both being wrong?
.2 x .2 = .04
What is the probability of three randomly chosen witnesses all being wrong?
.2 x .2 x .2 = .008
Thus, the odds of four or five randomly chosen witnesses all being wrong are exceedingly small.
For example, if there are 100 witnesses to a given situation, and the probability of each individual witness being right is .8, that means that on average 80 of the 100 witnesses will be right, and 20 will be wrong. So it will be quite easy to find five or ten witnesses who are wrong. Thus conspiracists have no trouble trotting out witness after witness from Dealey Plaza who believed he or she heard shots from the direction of the Grassy Knoll. This in spite of the fact that a clear majority of those who had any perception of the direction thought they came from the Depository.
But suppose there are several salient elements?
For Dealey Plaza witnesses, for example, they may have heard shots from the direction of the Grassy Knoll (implying a conspiracy), or from the Depository (consistent with the lone assassin theory). They may have seen the right top of Kennedy's head explode (consistent with a shot from behind), or the back of Kennedy's head explode (consistent with a shot from the Grassy Knoll). Bill Newman, for example, was a "conspiracy witness" with regard to where he located the source of the shots, and a "lone assassination" witness with regard to his description of the head wound.
In the case of medical witnesses, they may or may not have placed the large wound of the head in occipital bone (implying a shot from the front), or in parietal, temporal, or frontal bone (implying a shot from behind). They also may have denounced the autopsy photos and x-rays as faked (implying a conspiracy), or authenticated them as accurate (supporting a single shooter from behind).
With some witnesses, there are indeed many elements to their testimony. The JFK autopsists are an extreme example. On the issue of whether the "great defect" of Kennedy's head included occipital bone, the following data points exist:
What is the probability that two perceptions will both be correct?
.8 x .8 = .64
What is the probability that three different perceptions will all be correct?
.8 x .8 x .8 = .512
It follows from this that if there are 40 witnesses to a particular reality, and if each has three different perceptions that are relevant to the question of conspiracy, only 20 or 21 will be right about all three elements. Thus there will be 19 or 20 who have at least one perception that is mistaken. If a writer or TV documentary maker is careful to show the mistaken perception, and to avoid the accurate ones, he or she can produce 20 mistaken witnesses.
Of course, this leaves out of account a lot of real-world processes that might inflate or perhaps reduce the number of mistaken witnesses. The accuracy of perceptions may not be independent if, for example, some witnesses are having a good time telling conspiracy stories to researchers, or are simply being manipulated by those researchers. Conspiracists would be quick to point out that this is also true if witnesses are being careful to support the "official story" and refrain from saying things that support the notion of a conspiracy.
Ironically, the more a given witness knows about the case, the less useful his or her testimony becomes. "Knowledgeable" witnesses are likely to give testimony that is consistent with some particular theory of the assassination, and know how to avoid testimony inconsistent with that theory.
Sometimes, certain notions become part of the folklore of a particular society, and are known to pretty much anybody who might become a witness. One can quickly reach a point where there are dozens or even hundreds of witnesses telling stories that support these notions.
Thus nobody in Scotland has, to this researcher's knowledge, ever claimed to have seen Bigfoot. Likewise, few if any people in the western forests of the U.S. have claimed to have seen something like the Loch Ness Monster in some stream or lake. Nobody in the Bermuda Triangle has claimed to have seen either. The stories, in other words, are likely to be consistent with the folklore.
Partly this is because attention seekers telling tall tales know that tales consistent with the folklore will be more likely accepted. But partly it's because sincere but suggestible people peering through the mists in the twilight are likely to "see" something they are half expecting to see. This certainly happened in the case of the "Roswell incident" the supposed crash of a flying saucer in the New Mexico desert in 1947. For many years, nobody, not even avid UFO buffs, paid any attention to the incident. But then in the late 70s the story began to "take off."
Over the ensuing decades [after 1980] media coverage, book publications, and witnesses of various stripes emerged to expand the myth and integrate it with the larger mythology of UFO culture. The original story was subsumed into the framework of popular belief about a government conspiracy to withhold the truth about flying saucers, and the Roswell Incident ballooned to accommodate details provided by new witnesses. Along the way, an unexpected tourist boom dawned for an obscure New Mexico town, writers and TV producers made good money, and complete unknowns became famous. . . .Thus most of the accounts probably had some basis in some real-world incident. But as the folklore of the Roswell "saucer crash" spread, more and more witnesses interpreted what they had seen in terms of UFOs and extraterrestrials rather than mundane military activity. And a certain number simply lied.
As the Roswell Incident became an industry, TV networks and book authors aggressively pursued new witnesses during the 80s and 90s, often soliciting new accounts with advertising. Don Berliner and Stanton T. Friedman, for example, ran a newspaper ad in 1992 soliciting new witnesses, "In order to strengthen their case for government knowledge of what they call 'the truth behind almost 50 years of UFO sightings.'"
Regardless, the strategy of soliciting new witnesses was fruitful for media sources, and a number of fantastic stories emerged. Soon it appeared that the Roswell incident had involved flocks of flying saucers crashing into New Mexico during July of 1947 (or one saucer crashing in several places), with up to 11 completely different crash sites claimed. These locations neatly bracket the White Sands Missile Range, and define the operational area of Mogul and subsequent balloon programs, which suggests that military activity was the source for the accounts. (B.D. Gildenbert, "A Roswell Requiem," Skeptic, Volume 10, Issue 1 , p. 60)
Thus if a convincing number of witnesses claimed to see Lee Oswald and a woman (presumably Marina) driving around Alice, Texas and the surrounding area just weeks before the assassination, we might be impressed. If several apparently reliable witnesses saw Oswald at the Sports Drome Rifle Range in the weeks before the assassination, we might be inclined to think it really happened. If several witnesses claimed to have seen Lee Oswald with Guy Banister in the Summer of 1963 in New Orleans, we might ask "could all those witnesses be wrong?"
The answer is straightforward. Yes, they could be. Very easily.
Thus the bottom line is that, even if most witnesses are right most of the time, it's not that hard, in any complex case such as this, to mobilize the testimony of a superfically impressive number of witnesses to support one's case. Even if one is dead wrong.