Score counting error

["Andrew Hearn (AAISP)" <an...@aaisp.net.uk>]

Hi,

In my headers I see:

X-Spam-Status: No, score=4.3 required=4.4 tests=BAYES_99,NO_RELAYS
	autolearn=disabled version=3.1.7
X-Spam-Report:
	* -0.0 NO_RELAYS Informational: message was not relayed via SMTP
	*  4.4 BAYES_99 BODY: Bayesian spam probability is 99 to 100%
	*      [score: 1.0000]

Seems odd that score doesn't add up? (4.4 + 0.0 = 4.3!!)


-- 
Andrew Hearn

Re: Score counting error

[Kelson <ke...@speed.net>]

Andrew Hearn (AAISP) wrote:
> X-Spam-Status: No, score=4.3 required=4.4 tests=BAYES_99,NO_RELAYS
> 	autolearn=disabled version=3.1.7
> X-Spam-Report:
> 	* -0.0 NO_RELAYS Informational: message was not relayed via SMTP
> 	*  4.4 BAYES_99 BODY: Bayesian spam probability is 99 to 100%
> 	*      [score: 1.0000]

It's just differences in rounding.  The scores actually have three 
digits past the decimal point, but the report only shows one.

NO_RELAYS is actually -0.001, so the final score, assuming BAYES_99 is 
4.400 in your setup, is 4.399.  IIRC the final score is always rounded 
down in the report to avoid confusion when people see things like this:

	X-Spam-Status: No, score=4.4 required=4.4

-- 
Kelson Vibber
SpeedGate Communications <www.speed.net>