An term applied to the denizens of the USENET newsgroup rec.sport.baseball who would try to boil every aspect of the game of baseball down to probabilities and statistics. It was originally meant to be perjorative, but many posters wore the label proudly. The patron saint of Statdrunk computer nerds (SDCN's) is Bill James, author of Total Baseball and other texts full of dense statistics.
The SDCN's seem to have two main missions in life

Find (or invent) the statistic that determines the quality of a baseball player.
 For every baseball situation (e.g. Score tied, bottom of the ninth, 1 out, righthanded pitcher, righthanded batter, runner on first, 21 count) find the optimum managerial choice based on expected values, which in turn is based on statistical analysis of past situations and the baseball abilities of all the players involved.
The main whipping boy of the SDCN's is the Runs Batted In (RBI) stat and how much importance people have given to it as a measure of a player's value (Many winners of the Most Valuable Player (MVP) award led the league in RBI's). SDCN's say that RBI's partially depend on how many people are onbase when a player is atbat, and so is also a function of the quality of your teammates. SDCN's created the statistic OPS (Onbase percentage Plus Slugging percentage) as a much more effective stat in judging the quality of a hitter.
Traditionalists decried the SDCN's reliance on stats, especially since the SDCN's would routinely attack conventional wisdom.
Eventually, a subgroup was created specifically for the SDCN's, rec.sport.baseball.analysis.