Hi, > I can understand adjusting the values slightly for each user's
> > particular situation, but doesn't it generally throw off the balance of > > how the email as a whole is weighted when you adjust values in that way? > > If a rule is causing undesired behaviour (for this particular rule, > false negatives due to this hitting eg Hotmail or Yahoo mail), then the > stock score is not suitable for that mail flow. > > I add a score override on a stock rule a couple of times a month > (usually due to a FP report); I review them every couple of months to > see if a) the stock score is now closer to the local override or b) the > rule has been removed completely. > > Thanks for everyone's input. We did get an FP with it set to zero, so we'll continue to monitor the stock values. I didn't realize the corpus could lack the volume to get a more accurate calculation. Wish we could help there. Thanks, Alex
