While the guide is, itself, very specific to Movable Type, it, also, gives a very good high-level description of comment spam, spammers' techniques, and the pros and cons of different countermeasures. So, if you have a blog (even a non-MT one), it is worthwhile read.
At the end, they give recommendations or best practices for battling spam with MT-based blogs. Blog@Case follows nearly all of them. For throttling, force-moderation-on-older-entries, and content analysis, we use the excellent MT-Blacklist plugin. The comment CGI script has been moved from the default location. And, I have been experimenting with combining in the MT-DSBL and MT-Bayesian plugins.
This, also, illustrates an important concept. On a personal site with a couple friends running 5 weblogs or what-have-you, it's very easy to just slop all of these plugins together and see what happens. On a site like Blog@Case, I don't have that luxury. I need to know how these will interact; whether or not their interaction(s) is going to help or just confuse users; whether or not I can make it a per-blog configurable thing so each person can decide for him or herself how complex he or she wants their anti-spam countermeasures to be; will it scale to 5000 blogs (!); etc.
Right now, I am taking a wait-and-see approach. In my pockets, I have additional countermeasures ready to whip out at a moment's notice, but at the cost of additional complexity. If they are not necessary (or something better comes along in the meantime), I won't use them. We'll see how Blog@Case performs after it moves into Open Beta. I will adapt the system accordingly after that.