Jump to content
Sign in to follow this  
DavidT

"Bayesian experiment" has ended?

Recommended Posts

It would appear that the experimentation with the "Bayes" feature of SpamAssassin on one of the SpamCop servers ("blade4") has ended, but I've not seen any announcement to that effect. From the headers of spam that I've received in the last 24 hours, it appears that the experiment ended very early on Wednesday morning.

However, as I've mentioned in another thread, there seem to be residual discrepancies between the way that identical spam messages are handled by the various "blades," with "blade4" not behaving quite like the others, even after the apparent end of the experiment. If there's been a clear explanation of what's going on posted elsewhere in the last 24 hours, I must have missed it.

David T.

Share this post


Link to post
Share on other sites
It may have been better to have added your remarks/query to the existing Topic/Thread started by JT about the "test" ... this would have brought it back to the top ot the listings, highlighted it as un-read to all others, etc ...

I think that "may" is the key word here. I've tried doing that with other topics that are a few days old, including ones that I've started myself, and even though it does bring the topic to the top, it doesn't necessarily result in any new responses. One possibility is that people click back into the topic, but don't scroll all the way down to see the most recent response.

Furthermore, by simply adding another response to an existing topic, you don't get to alter the topic Title or Description. By starting a new topic, you do, such as is the case with this one.

Regarding the "Bayes experiment," I downloaded almost 300 messages from my Held Mail and found that before the experiment was (apparently?) ended the Bayesian filters were getting MUCH better, in terms of scoring. However, each one of the messages with signs of Bayesian filtering had also been blocked by "bl.spamcop.net," so not one of them was blocked only due to the filtering.

Share this post


Link to post
Share on other sites

Well, your ending remarks are pretty much exactly what JT said over in the other Topic/Thread ...

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×