I updated to SpamSieve 2.8.4 and suddenly a lot more spam is getting through, including seemingly easy catches, like Viagra. A lot is still getting caught, but I used to only see 1 or 2 spam emails a day and this morning, I already had 15-20.
Michael, I tried to email you with this, but got this rejection:
The following message could not be delivered because the address spamsieve@c-command.com was rejected by host c-command.com (208.113.200.128).
554 5.7.1 <spamsieve@c-command.com>: Recipient address rejected: Access denied
Reporting-MTA: dns; intellexcorp.com
Arrival-Date: Thu, 14 Oct 2010 06:52:29 -0500
" updated to SpamSieve 2.8.4 and suddenly a lot more spam is getting through, including seemingly easy catches, like Viagra. A lot is still getting caught, but I used to only see 1 or 2 spam emails a day and this morning, I already had 15-20."
I am in the same boat but I am not as popular so my numbers are smaller.
Do we have to go back to the previous version and if so how?
Are there “Predicted:” entries for those messages in SpamSieve’s log?
By the way, I do seem to be having problems receiving mail right now, so please use the private message feature in the forum if you need a quick response.
Is there a problem?
I am confused as to whether or not my problem is even being addressed. RandySK at the beginning talks about v.2.8.4 appears to letting a lot of spam through. I also have that problem.
I don’t know what all the other threads are talking about.
Is there something wrong with v. 2.8.4? Should I go back a version?
I examined MartyA’s log file and found that the 6 spam messages that got into the inbox do not seem to be related to the SpamSieve 2.8.4 upgrade. Most of the messages got through because the senders were on the whitelist.
-regardless of what is occurring now the new rev, has increased the number of spam getting through. If this is my fault or just a change in how the program is behaving, can be explained but I will not understand it.
-I am assuming (bad??) that if I just keep training for spam things will end up even better than before. If not…
I doubt that’s the case, but if you want I could send you the previous version and you could try that and see if things suddenly improve.
I think the problem is that you’re giving SpamSieve contradictory information. I described several examples of this via e-mail.
For example, there was a newsletter that you’ve been receiving since August, and so SpamSieve had whitelisted it. Then earlier this month you trained one of them as spam. Same with the alerts/notifications from Fidelity. If these were spam all along (I doubt it, but it’s hard to know without seeing the messages) then you should have been telling SpamSieve that when the messages were in your inbox. If these are not spam messages, then the recent ones shouldn’t count as SpamSieve mistakes, and telling it that they’re spam will confuse it. Does that make sense?
To clear out this contradictory information you would need to choose Reset Corpus and then re-train SpamSieve.