In the last note, some particular pieces of evidence were examined. They seemed to say one thing, but other evidence showed that the first reading, or the more obvious interpretation, was not true. In some cases the fault was ours for assuming something. In at least one case, the original record was simply in error. This points up that any piece of evidence might be in error for a variety of reasons.
Given the less than perfect reliability of evidence, what is the impact on the final outcome? When compounded over several generations, the chance that a particular sequence of events is true starts falling very rapidly.
For example, a birth recorded in the Hebron Church Register is assumed by most people to be true, i.e., in a thousand cases, it will be correct one thousand times. But yesterday we discussed the case of Johannes Becker whose father was not the man that we would normally assume from the record. If, in one thousand births, the recorded event is true nine hundred and ninety-nine times, we could assign a probability of 0.999 to the event being true. Seldom is evidence this good. This is actually a high probability.
Suppose that we have a chain of events, say male ancestors back for eight generations. Say the probability that each male is correctly identified is 0.999. What is the probability that great-great-great-great-great-great-grandfather has been correctly named? The probability is quite high, still better than 99 out of 100 (or 0.99 plus).
But a probability of 0.999 is extremely high. In many of the generations or steps backward to an eighth ancestor, the odds are quite low. Many times, the evidence that can be accumulated in any one generation would warrant odds no better than 0.8, i.e., in ten similar cases the facts would be true only eight times. If these odds were applied to each of the eight generations, the chance that the sixth great-grandfather has been identified correctly, falls to slightly less than 0.17. Or stately differently, it is rather unlikely that he has been correctly identified. Out of eight similar cases, only one of the eight would be correct.
The point is that the odds compound. At each step, the chance that the sequence is correct weakens. Considering that some of the best evidence is not as good as a first glance might indicate, it becomes important to find new or improved sources or data to buttress the case.
There are a lot of guesses floating around in the world of genealogy. A lot of these guesses do not even rate an assignment of 0.05 for being correct.
The old adage that "a chain is no stronger than its weakest link" is true. But in genealogy the problem is even more difficult. We not asking if there is one rotten apple in the barrel. We are asking if there are any rotten apples. There are many different ways that the chain can be broken. Some of the events are not what they seem to be. The probability that a fact is true should never be assigned as a rating of certainty, i.e., a probability of 1.0.
We gratefully acknowledge the work of John Blankenbaker who published over 2,500 Germanna History Notes via the Germanna-L@rootsweb.com email list from 1997 to 2008. We are equally thankful to George Durman (Sgt. George) for hosting the list and republishing the notes via rootsweb.com.