from Amazon
Tuesday, May 13, 2008
If I worked for a team and was allowed to do whatever I wanted (and they actually listened)….
Here is a partial list of what I would do:
1) Tell the manager to throw away his batter/pitcher index cards, or book, or whatever it is that they use, and never look at another batter/pitcher historical result again. I’ll get to the alternative in a minute.
2) Do the same for any other small sample size of historical performance. For example, don’t ever look at how a batter is doing lately, either yours or the opponents’. That should have no bearing on any of your decisions.
3) Never spend another minute worrying about the best lineup. Use a set lineup against RHP and LHP and leave it alone unless you change players or someone is injured.
4) Along those lines (#3), I will give the manager the 2 or 3 best lineups to use against RH and LH pitchers. I can even tweak those for GB and FB pitchers (remember that there is a significant GB/FB platoon advantage - it is just that it doesn’t come up very often).
5) Along the lines of #2, I will give the manager a book or index cards of each batter/pitcher matchup. It is comprised of the batter’s current projection, adjusted for the park (maybe) and the pitcher’s current projection (again, maybe adjusted for park), combined using each player’s platoon ratio and a log5 method. I might give him several versions: one for in general, another for when he needs a K, another for when he wants to avoid a BB, another for when he wants to avoid a HR or extra base hit, etc. I would also have two numbers: one for when the batter is already in the game and has seen the pitcher several times, and another one for when the batter is a pinch hitter (includes the pinch hitter penalty). That way, the manager, if he wants to compare the batter in the lineup with a potential pinch hitter, all he has to do is to compare the two players applicable “matchup” projection. I would also have a column of each player’s projection (displayed in some manner that the manager can easily understand, like EQA or wOBA) versus a RH and LH opponent. That way, for example, he can decide between two or more players he is considering bringing in as a pinch hitter or reliever, given the likely opponent or opponents’ handedness.
6) Give him a guideline for estimating leverage in any situation. Give him a list of his pitcher’s projections and explain to him that he should put in his best pitchers (versus the upcoming batters) in high leverage and his worst pitchers in low leverage, as a general guideline. Along those lines, we would throw out the window the concept of the closer, set-up man, etc. Of course, we would explain that to the players before the season begins. We would explain that pitchers would still get the same number of saves overall, but that the “closer” might get a few less saves and the other pitchers might get a few more. We can still call the “best pitcher” the closer, and the next-best pitcher the “set-up man” if that will make everyone happy.
7) Tell the manager to tell the batters to not worry about getting called out on strikes on a 3-2 count, especially when a baserunner is important. IOW, we want the batter to take more on a close pitch with a 3-2 count and we want him to know that he won’t get chewed out for being called out on strikes some percentage of the time. In fact, anyone who doesn’t get called out on strikes with a 3-2 count is probably swinging at too many pitches.
8) I would study the pitch f/x data for a lot of things of course. One of them would be to tell each batter whether he should or should not swing at 2-0, 3-0, and 3-1 counts in the various situations, against various types of pitchers. Of course game theory is implicated there (e.g., if the other team realizes that so-and-so never swings at a pitch in situation A, then he will throw a BP fastball right down the middle, which may require the batter to now swing at that kind of pitch, etc.).
9) Basestealing. I would make sure that we had a database of all pitchers’ time to home plate and all catchers’ time to second base, and combine these into one number. Then before the game, I would go over with all my players these numbers and tell them, you either have or don’t have the green light in these situations. The manager can also give the green light or red light to the 3rd base coach during the game, according to the chart. No one EVER attempts a steal versus a LHP, and consequently, if you ever get picked off by one, you are in a lot of trouble. No batter EVER takes a pitch (that he might otherwise swing at) in order to let the runner steal.
10) I would figure out if the hit and run is actually a good or bad thing. If it is close, then I would explain to the manager when it is correct and when it is not (IOW, with whom on first, whom at bat, etc.). I would have the managers explain to the baserunner that a hit and run is NOT a license to get a bad jump and jog to second while looking back to home plate 2 or 3 times. When you get a hit and run sign, you get a jump and run to second as if you are stealing, other than the fact that we want you to look back to see where and if the ball is hit.
11) I would figure out if and when it is correct to “guard the lines” and when to play the outfield up a little and back a little. The depth of the outfield and how close the 1st and 3rd basemen play to the lines should be predicated on the difference in WE between the single and extra base hit. It should change on virtually every batter. Teams don’t do that. Also, I would determine whether it is correct to shift against certain batters or not. Some teams do it and other teams do not. Some of these teams are making a mistake (unless it is dead even, which maybe it is). I want to know which one is correct. I would also tell my players that are shifted against that it is OK to drop a bunt down until they stop shifting. No shame in that. I will even stay at the park late to give them extra bunting practice if they want!
12) IBB. Don’t ever IBB anyone unless it is late in the game and the IBB is obvious. That is a general rule of thumb. If I am feeling generous, I might give some more specific guidelines.
13) Sacrifice bunt. I would take out The Book, and have them memorize my general guidelines. Generally, it is: If the infield is expecting a bunt, NEVER bunt. If the infield is not expecting a bunt, you can bunt sometimes with almost anyone, assuming that they can bunt decently. As soon as your opponents catch on to your strategy (which should be pretty soon), you mix up your bunting and not bunting randomly. I can give them the mean percentages depending on certain rules of thumb, although the exact percentages are not nearly as important as just mixing it up random (or at least so that it appears random to the defense). This also applies to pitchers. They should NOT be automatically bunting. If a pitcher bunts with runners on 1 and 3 with 1 out, you (the manager) immediately get fired. As a very general rule, if your batter cannot bunt very well or is slow, rarely bunt him. If he is fast and a decent or better bunter, bunt him some percentage of the time. The worse the hitter overall he is, especially in the power department, the more you bunt him. The better the bunter and/or the faster he is, the more you bunt him. But, still, you bunt everyone some of the time and NOT some of the time. If the defense knows with 80 or 90% certainty what you are going to do, unless it is obvious (like Ortiz or a crappy hitting pitcher is at the plate), then you are doing something wrong (or they are stealing your signs). ALL pitchers must run hard on a decent bunt. Not hard enough to possibly hurt themselves, but hard enough to occasionally force an error or get a single.
14) In a high leverage situation, try and avoid having your pitcher bat. You can obviously adjust that according to the inning, how good of a batter he is, how good of a pitcher he is, and whether your pen is rested or not. But, in general, you need to have your pitchers bat A LOT less than they currently do. Good rule of thumb: In a low leverage situation, you can always let your pitcher bat if you want to. In a high leverage situation, pinch hit for your worst pitchers (no matter how well they are pitching in the game thus far) in the 5th or later and in the 7th or later for your best pitchers. And that is being generous! Explain that to your pitchers before the season starts. (Actually, I would implement the “4th and 5th starters never bat” policy, as explained in The Book, other than I would add, “unless it is low leverage in the early innings.”
15) Tell them to tell their outfielders when to drop a foul ball with a runner on 3rd and less than 2 outs (generally in a close game with 1 out and/or with 2 strikes on the batter, although I can get more specific than that).
16) Tell them to keep a running database on each pitcher’s pitch speeds, each batter’s bat speeds (bat speeds need to be tested throughout the season, which they are not), and each player’s running speed. That can help to detect injuries and aging patterns, and the running speed can help me with the basestealing strategies (and baserunning) above.
17) Tell the manager that his pitching coach gets fired if any pitcher walks a batter in the 9th inning or later when their team is ahead and the batter is not the tying or go ahead/winning run, and he doesn’t say anything to the pitcher afterward. (Just kidding about firing him.)
Again, this is a partial list off the top of my head, but I am guessing that full implementation would add from 3-5 wins to a team. You guys can add to the list at your pleasure and leisure.