@figuerres: The fact is, I'm not sure what works either.
Rely on Microsoft account creation may or may not work, because of the proxy problem I mentioned. Nothing stops you from repetitively creating accounts on multiple VMs pointing to different anonymous proxy servers.
Making any restriction based on how many time / how fast you can do something will always push the spammers to use multiple accounts to publish spams, making the situation more difficult to manage. When you designing mechanisms against spammers you'll want to keep in mind regarding this.
The methods with less drawbacks are posting pattern recognition, but unless you can code some AI method, it'll need someone to maintain or it'll lose effectiveness over time. Btw, AI can be trained against normal users so it can backfire too.
IMO, the most effective way is to have site admins watching the forum and amend spam blocking strategy as needed, but seems that's not the case here. It would be nice if they can grant users with 10+ years registration time some limited level right to temporarily remove spams but they don't do that. They've already granted some MSFT members here right to remove spam posts but they're not always here.
The current situation can be roughly stated as above.