For example, > there is just about no situation in which a sacrifice bunt is a good idea. > Those in which one is are definable - basically the bottom of the ninth > inning in a game where you're tied or down by one run. _In any other > situation_ making a sacrifice bunt actually decreases your probability of > winning the game. That's something that the statisticians figured out a > long time ago.
Bob So let me get straight; independent of variables such as the players involved there is never a reason to bunt. It seems to me the notion that you can manufacture a single run is key to understanding when this strategy can be useful. Late in a game that you are within one run of the other team (1 down 1 even or 1 up) and you have a runner on first. A weak hitter is up against a dominant pitcher. (Todd Green against Pedro). The rationale as I understand it is that it easier to get a bunt out and a single than two singles. You are sacrificing the chance for more than 1 run so you better think that one run will help enormously (if you have no closer the strategy makes no sense). Man on second the rationale is stronger. Now you can get two outs and score the run. You are telling me that this strategy has been deemed to be unsuccessful in these specific circumstances? G Who are some managers who are famous for their aversion to sacrifice bunts? Earl Weaver, Davey Johnson, and so on. Not by coincidence, they won a lot. B Well is that why they won or were there other reasons. I would point out that by the Zimmy criteria (WS victories) Davey hasn't been all that successful. G Statistical predictions even have empirical > data to verify them, Bob. Players who draw more walks will hit for more > power - see, Sammy Sosa or Alfonso Soriano for examples. B But is there a causal relationship or are these measures correlated with another variable that is more significant. Are they correlated in a trivial way (e.g height and age in humans) And that's not even a prediction supported by a lot of data, it's just something that a lot > of us believe because of what we've seen in the data. It just works out very well. Teams that don't use sacrifice bunts will score more runs. But will they win more games. Or alternatively do teams with lots of offense not have the need or personal to sacrifice. (A bunch of big boppers who can't run worth stink may not be the ideal running team). See every team ever managed by Weaver and Johnson. The ratio of runs scored to > runs allowed is an exceptionally accurate predictor of winning percentage. > If Brian Cashman says, in essence, I think the sabermetricians are right and he >beats the pants off everyone else, don't you think he might be right? Is that what he says? It seems to me that the Yankees play some "small" ball when they need to. They stress the "little things" advancing runners stealing when the time is right sacrficing when the situation requires. > When Billy Beane beats the snot out of the rest of the league on a $30M > payroll and he says the reason is that he applies statistical techniques to > how he manages his team - don't you think it's just possible that there's > something to that? I don't believe anything someone named billy beane would say. But seriously I am not arguing about the value of statistics in evaluation of players with relationship to other active players or for their use in teaching players how to play or in building a team. I have argued that they cannot be used to compare across different eras. I have also argued that the definition of the "best team" cannot be determined by JUST looking at the the performance of individual players. Winning as a team is an emergent phenomena that correlates with but is not completely described by its components. So there is our arguement or discussion (it is fun and it should be neat to see how this plays out over the next few weeks). I think that stats you quote can be misleading in two specific circumstances; The comparison across eras (especially in attempting to compare the best players in each era) and because individual stats can fail to capture team performance (emergence).
