Confused spam filter with bayes?

[chris42]

Hi there,
I continuously have spam mails in my inbox and completely weird - non matching - headers:

X-Spam-Bayes: No, 0.24
[...]
X-Spam-Result: DMARC_POLICY_ALLOW (-0.50),
        DKIM_ALLOW (-0.20),
        RWL_MAILSPIKE_GOOD (-0.10),
        ARC_NA (0.00),
        DKIM_SIGNED (0.00),
        DMARC_POLICY_ALLOW_WITH_FAILURES (0.00),
        DNSWL_BLOCKED (0.00),
        DWL_DNSWL_BLOCKED (0.00),
        FREEMAIL_REPLY_TO (0.00),
        FROM_EQ_ENV_FROM (0.00),
        FROM_HAS_DN (0.00),
        HAS_REPLYTO (0.00),
        HAS_X_AS (0.00),
        HAS_X_GMSV (0.00),
        RBL_SENDERSCORE_REPUT_BLOCKED (0.00),
        RBL_SPAMHAUS_BLOCKED_OPENRESOLVER (0.00),
        RCPT_COUNT_ONE (0.00),
        RCVD_COUNT_ONE (0.00),
        RCVD_TLS_ALL (0.00),
        REPLYTO_DOM_NEQ_FROM_DOM (0.00),
        SPF_SOFTFAIL (0.00),
        TO_DN_ALL (0.00),
        TO_EQ_FROM (0.00),
        URIBL_BLOCKED (0.00),
        FAKE_REPLY (1.00),
        FORGED_RECIPIENTS (2.00),
        TO_DN_RECIPIENTS (2.00),
        MISSING_MID (2.50),
        FREEMAIL_REPLY_TO_NEQ_FROM_DOM (3.00),
        BAYES_SPAM (5.10)
X-Spam-Status: Yes, score=14.80

First summary of bayes gives: not spam
Bayes scoring gives: spam
Overall score gives: spam
It is treated as not spam and ends up in my inbox

Not sure what is going on here?

[mdecimus]

The BAYES_SPAM tag comes from the global Bayes filter while the X-Spam-Bayes: No, 0.24 headers is added by the account specific Bayes filter. If you have account Bayes filtering enabled I suggest that you disable the global one.

[chris42]

Understood on the filter.
I think I found the error, why this is not marked as spam. Got a sieve filter to mark them read and accidental put a stop in the end. I guess that stopped processing it further.