[Anthropoid]

I think I've figured out an easier system than what I described above. Well, at least it seems to have served me well on auto-play (with strategy set up for both the team and the players, and player set to adjust to team strategy).

Once you have a screen with your prospective batters open a window with the Batting Stats (I have a template called "Lineup Stats" that has most of the batting ratings, and speed, plus Defensive value at position, and then the Batting stats that I think are most useful). I've come to realize that, the splitting function is actually quite useful.

Say you select Major Leagues for the top level, then "Overall" or "Career" for the split or double-split, you will of course get numbers that are reflective of each players entire major league career.

If you select anything else at the lowest level, it seems to simply make use of the most recent season's stats; actually, I'm not entirely sure of that but I think so. So for example if you select "Bases empty" you'll get all the stats recalculcted for all the plate appearances where the player faced a field without any teammates on base. Likewise, you can do the same with all the other combinations (player on 1st, player on 2nd, or on third, on 1st and 3rd, 2nd and 3rd, bases loaded).

Some of these other stats I was using before (SO%, GDP, wOBA) are to some extent derived from or related to batting average. If you got a hit, then you obviously got on base and didn't strikeout, though it is possible you grounded into a double play but still managed to get onto base, even while two of your teammates (who were on the field when you came to the plate) were retired. This of course could only happen in a situation with two players already on bases, and no outs. Probably would be unlikely to happen in any situation except player on 2nd and 1st, possible 3rd and 1st. with 2nd and 3rd, it seems quite likely that it would never be a double play, though I suppose anything is possible. Anyway, just an aside to say: special stats are neat, but when you can use the aggregate least divisible variable instead, that is generally preferable.

Long story, short, I think that by looking at batting average in different "splits" you can effectively come up with an even better estimate of a players proficiency at any given order in the lineup.

#1 needs a high batting average when bases are empty (if the lineup has gone all the way round, it could of course be any other permutation too, but that is a low occurrence event compared to tthe leadoff man facing all bases empty). Basically, the guy with the best BA with bases empty, as long as he has good speed, and nothing else that looks deficient (SO% and BB% in particular) = best leadoff man.

Using this method I actually determined that Honus Wagner was my best leadoff man, and so far it seems to have proven true.

Number two guy can also face bases empty, or (and restricting it to the most common permutations): runner on 1st, 2nd, or 3rd. Add those four batting averages together, and divide by four and you've got his "#2 batter" score. Usually you can figure out the two players who you need to do this with pretty easily: the ones with the 2nd and third best BA against empty bases. A bit tedious but only takes a couple minutes.

#3 faces a larger set of (typical) permutations: bases empty, or runner on 1st, or 2nd, or 3rd, or runners on 1st and 2nd, or 1st and 3rd, or 2nd and 3rd. So that is seven different batting averages to add up and calculate a mean value for. It sounds laborious but it does actually seem to confirm one's gut intuition for who should be in what slot.

#4 through 8 it is virtually impossible to predict with any certainty what they will face, but #4 in particular faces the highest inherent probability of "all possible" permutations of offensive distributions. So plain old Batting Average is bound to be the best indicator for #4! (whether overall career or overall this season might be a matter of context).

#'s 5 through 8 slots are similar to #4 in that it is virtually impossible to predict with much certainty what offensive permutation will be faced, so again the best predictor for best batter at each of these slots is again aggregate batting average. Out of the remaining players, the guy with the next highest BA (next below the guy who is in slot 4) goes in 5, next highest in 6, next highest in 7 and lowest of the fielders gets 8. Pitcher or DH go in 9!