Am I the only one that chokes on the "1 false positive out of 25000" ??!! Thats like..hundreds an hour in my environments...
Ill stick to my schema..
Not to give any credit to their claims, but that's much lower than the claimed FP rate of SA 2.64's STATISTICS*.txt
With a threshold of 5.0 and SA 2.64's corpus has between 0.07% (set0) and 0.04%(set3) FP rate. Theirs works out to 0.004% which is a factor-of-10 ahead of SA's claims.
Admittedly SA is tweakable to reduce FP's considerably, but being a SA user the 1/25,000 FP rate doesn't choke me up at all. If it's true, it's quite impressive. Very few spamfilters can claim a FP rate anywhere near that low.
I also find it impressive you're processing in excess of 2.5 million messages per hour with spamassassin. That's a very busy domain you have Dave.
