On Wed, 01 Oct 2003, Ian Clarke wrote: > On Wed, Oct 01, 2003 at 12:45:33PM -0700, Ian Clarke wrote: > > On Wed, Oct 01, 2003 at 12:35:08PM -0700, Ian Clarke wrote: > > > Line 160: > > > double pNotConnectFailedOrSearchFailed = > > > (1 - pConnectFailed) * (1- pSearchFailed); > > > My high-school math is a bit rusty, but shouldn't that be: > > > ... = 1-(pConnectFailed+pSearchFailed); > > Well, thinking about it what is there is: > > not(connectFailed) and not(searchFailed) > > My suggested correction was: > > not (connectFailed or searchFailed) > > Which, if I am not mistaken, is the same thing. > > Thinking about it some more - perhaps not. > > Lets say pConnectfailed was 0.1 and pSearchFailed was 0.3. the original > way of calculating it would tell us that pNotConnectFailedOrSearchFailed > is 0.63 - but my method would tell us that it is 0.6. I think my method > recognises the fact that you cannot have a ConnectFailure and a > SearchFailure at the same time.
To illustrate, compare the throwing of two coins at once verses in series. The probability of getting a head out of two coins thrown at once is three in four. The probability of getting a head out of two coins thrown back to back is two in three. (You wouldn't throw one if you already had a head, right?) Problem is, the second scenario is wrong from a probability perspective. The odds of getting heads are still three in four. The double negatives are very confusing. NOT connection failure and NOT search failure In other words, exactly what we have already. The only other way to express it would be 1-pSF - pCF + pCF * pSF. Ugly. P.S. Disregard my previous post, I found a nice stache of crack rock and was smoking furiously. A or B does not translate to pA / pB. I really don't remember my probability math.
pgp00000.pgp
Description: PGP signature
_______________________________________________ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
