Saturday, May 05, 2007

What's It Going To Be?

Early in the season, Kevin Youkilis suggested that Sox fans were making judgments other words, based on too small a sample size. How much of a sample size is 'enough'?

Statistics can be a tricky subject. To reach statistical significance for a given event not happening (at the 0.05 level), a ratio of 3/n is sometimes employed. In other words, if something doesn't happen in 100 tries, that would be 3/100 (.03) and would be 'statistically significant at the .05 level' that the event wasn't going to happen. That obviously doesn't mean it couldn't happen, and in baseball it seems that anything can happen.

When the Tigers started the 1984 season (hope I'm right here) 35-5, it seemed inescapable that they were going to win their division. On the other hand, because Josh Beckett starts 6-0, I doubt that any sane Red Sox fan would predict 30 wins, as some New York fans projected for Randy Johnson in his first campaign with the Bombers.

What can we say about the Red Sox at this point? Offensively, it seems fair to say that they are likely to be in the top thirty percent of the AL. They are second in both runs and OPS, and not that many players are having 'career' years. Defensively, I expect them to be in the bottom third of the league, possibly abetted by having a pitching staff that has a K/9 inning ratio of 7 (4th in the league). On the hill, the Sox are second in ERA, and lead the league in opponents batting average (.228). You need skill, health, and luck to maintain these figures, and only time whether the cosmic forces will cooperate.

I've argued the importance of quality starts in the past, and here's the proof. Quality starts tend to result in a team winning about 70 percent of those games. The Fantastics Baseball Blog lists some previous results for both Quality Starts and Disaster Starts. With Johan Santana going tonight, hoping for a Disaster Start doesn't tickle reality. The Quality Starts blog seems to have run aground, but as Youkilis would say, it lacked statistical validity to the point it dried up.

Is it better to have your number five starter going against Johan Santana or your top dog? It's hard to expect the Sox to score a lot of runs against the Twins' ace. On the one hand, "it takes a diamond to cut a diamond", but on the other, is your diamond hard enough?

No comments: