Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: AR confidence and one person's opinion of 'best practices' (Read 14570 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

AR confidence and one person's opinion of 'best practices'

Reply #50 –
If you take into account the birthday paradox, which doesn't exactly apply since it would assume equal probability for each unique hash, then the third individual getting a match should seal the deal for you.  How you wish to guess as to what the confidence of the third person's hit is up to you, but don't delude yourself, you're only guessing.  The confidence could be as low as two in this case, since the third person's hit does not automatically bump the confidence level and it may never bump the confidence level if the result is never submitted or is rejected for some reason.  EDIT: Of course if you haven't previously submitted results, you are this third person.

Taking this silly assumption that two people will submit twice but not once or three or more times and that these two people will have unique physical discs with scratches that result in a hash collision (I'd love to see your math showing the likelihood of this happening, btw), then the confidence need only be 4, not 5.  Once we dismiss the possibility that two people will have two discs that are damaged resulting in a hash collision, and I believe it should be dismissed until you come up with some numbers, we're back where I started, a confidence of one is sufficient provided it is not your submission.

Again, I am ignoring the other mechanisms that can cause multiple bad entries.  As was mentioned earlier, none of these mechanisms can be categorized as neatly as "X is bad" and "X+1 is good"; despite your insistence, re-submission can't be categorized this way, either.