30-11-2012, 03:33 AM
Quote:James H. FetzerSo you think these 50 points of identification, the similarity in their right ears, in their left eyes, in the shape of their skulls and their hairlines IS A MATTER OF COINCIDENCE? What would the probability be? If we assume similarity would occur BY CHANCE one time in 10 for each of these features, then we are talking about an improbability of 1/10 to the 54th power or 1/1 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 more or less. I infer you are no student of probability, but the probability of 54 points of similarity if they were the same shirt would be approximately 1!
First off Jim, I do probability for a living... your "inferences" as I've read over the years leaves much to be desired.
The 10% by chance has only to do with the ABILITY TO CORRECTLY CONCLUDE THEY MATCH... not whether they do or not...
If you have two shirts and one is a buttoned shirt and the other a Tshirt and both are white and size M... I can say with 100% accuracy that of the 2 items that were looked at to see if they match 100% of them match - Color and Size... so they MUST be the same shirt.... right? of course not... there are other variables but you CAN CLAIM that on these two items your MATCHING ABILITY is 100% accurate.
YOU or RALPH stating something is a MATCH has it itself odds of you even being right to begin with...
What are the odds that you've incorrectly "matched" any two items on the different PHOTOGRAPHS of the two shirts (one of them from HOW SMALL AN AREA on that photo?)
Shall we say 50%?, 25%? or are you of the opinion that y'all got it right 50 out of 50 times? and that it is part of the tautological argument that you are assumed right in each case?
Your 10% BY CHANCE statement has no bearing on the probability of YOU AND RALPH being right about the analysis...
THAT's the % that matters Jim... how RIGHT are you and how do you substantialte any claim of 50%, 75% 99% correctness?
If the chances are 50% on any ONE ITEM then the CHANCES of YOU AND RALPH being right about 2 of the items is 25%, 3 items? .5 to the 3rd... 50 items?
[TABLE="width: 481"]
[TR]
[TD="class: xl63, width: 641, bgcolor: transparent"]1,125,899,906,842,620 : 1 are the odds that YOU AND RALPH are correct on ALL 50 items if you are correct 50% of the time in your matches... 50% to the 50th power
When you can give a % to the probability of you being RIGHT about the match, for each and every one of the 50 items... you then multiply ALL THOSE %'s to get the probability of you being right... for ALL of them. Or you can pick two items like I did above, and with 100% certainty claim a MATCH... but just cause they are both white and size M, it does not follow they are the same shirt.
Does not prove the shirt is the same... the COLOR may even be just a little off... would that make you wrong about the shirts being THE SAME - yes, it does.
So list the 50 items, and give each a % of correctness in your matching ability....
Since we do not know ahead of time whether the shirts are the same OR whether YOU AND RALPH are correct even about a SINGLE ITEM....
Let's say you are 90% sure on ALL 50 items... the chances are still 194:1 that you are indeed right about all 50.... Obvioulsy if you are 100% right... they'd have to be the same shirt...
So you see Jim... the question is not whether the SHIRTS ARE THE SAME OR NOT...
The QUESTION is how reliable is your MATCHING on each and every item looked at?
and HOW many items can we find that DO NOT MATCH... which is evidence on the other side - like my example above (buttons versus none) that helps dtermine if the two "shirts" are the same...
Next time you want to make probability statements call me... the CA Lottery and LAPD, among a number of others, already do....
Cheers
DJ
[/TD]
[/TR]
[/TABLE]