This is a multi-part message in MIME format.
--------------010509010807010403010807
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit



Per Jessen wrote:
> dirk@bonengel.de wrote:
>
>
>> The difference is that the .de domain is fed by input that's either
>> visually checked or stems from dedicated spamtraps, so I'm quite
>> confident the hashes contained really mark spam.
>>
>> The .ag domain contains hashes either from feedback loops (ie. end
>> users) or from mails marked as spam by other systems. Thus there's a
>> higher risk of getting FPs from that list - hence the lower score.
>>

>
>


One thing to note about how the iXhash works is that if it were fed a
false positive that wasn't some mass mailing then that false hash would
probably never be hit again. That's why I think this meathod is highly
false positive resistent.

--------------010509010807010403010807
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: 7bit











Per Jessen wrote:

dirk@bonengel.de wrote:



The difference is that the .de domain is fed by input that's either
visually checked or stems from dedicated spamtraps, so I'm quite
confident the hashes contained really mark spam.

The .ag domain contains hashes either from feedback loops (ie. end
users) or from mails marked as spam by other systems. Thus there's a
higher risk of getting FPs from that list - hence the lower score.







One thing to note about how the iXhash works is that if it were fed a
false positive that wasn't some mass mailing then that false hash would
probably never be hit again. That's why I think this meathod is highly
false positive resistent.




--------------010509010807010403010807--