from Amazon
Saturday, March 26, 2011
I agree with Babe Ruth
I’m sure I don’t agree with everything he says, but I want him on my side on the IBB issue.
Filter posts by...
In-game_Strategy
Tuesday, March 15, 2011
Bunting in a high-run, and poor-fielding, environment
In MLB, bunting makes more sense when the run environment is low, say close to 4 runs per game. But in college, when run scoring is very high, why bother bunting, beyond the “keeping them honest” aspect? Well, Jeff gives us good data (ESPN):
Last season, in over 3,000 sacrifice situations (runner on first and/or second with fewer than two outs), the bunter and all baserunners were safe over 17 percent of the time. An error was committed more than five percent of the time. Double plays? Less than one percent.
Let’s say that means 22% of the time, it’s like an infield hit, 1% of the time, it’s a DP. Jeff doesn’t report the other numbers, but let’s say 50% of the time, it’s a ground out with runners advancing, and 27% it’s a fielder’s choice (runner out, batter gets on).
Breaking out our trusty RE24 chart (WE would be better, but let’s take it simple):
- 5% of the time, you add +.9 runs (I put runners on the corners)
- 17% you add +.6 runs
- 50% you get -.2 runs
- 27% you get -.4 runs
- 1% you get -.8 runs
Adding it up, and we have -.07 runs. That’s with an average hitter in a 5 runs per game environment. That seems like the kind of numbers you’d get with MLB outcomes too. I don’t buy that you should bunt MORE in college because of the poor fielders (they are also poor fielders when they get the ball on the ground on swinging away, no?).
But, at least it seems defensible in certain situations, and is not the crazy idea I initially thought.
Thursday, December 16, 2010
No equilibrium in 2 point conversion? Run more!
More fascinating stuff from Brian. While the current overall success rate is 48%, when you break it up by rushes or not:
Still, even after removing all QB runs, conventional RB runs are successful 57% of the time. This suggests that if teams ran more often, the overall success rate would increase. Defenses would likely respond, and eventually the success rates for both running and passing would equalize at a success rate somewhat over 50%.
He had the QB plays at 75% for 47 runs and 43% for 525 non-runs (passes, sacks) for an average of 46%. So, if you handoff to the RB, you will get the 2 pointer 57% of the time. If the QB controls the ball, it’s 46% of the time. This means you should run more. As you run more, the defense will expect you to run more, and so, your success rate on runs will go down (while it will go up on QB plays). Now, it won’t necessarily settle at the mid point (51.5%). It’s possible that it will settle even less than 50%.
Here’s the question: is it even possible that it will settle at only 47%? That is, if you run “optimally”, then the success rate for QB and RB plays are both identical (say 47%). But, if you run sub-optimally, it’s 46% for QB plays and 57% for RB plays. Is it possible that increasing your running plays will have just negligible impact on the success rate of your QB plays (if the defense plays for the run, it would hamper the QB scrambles). So, can it be optimal to have a 46% / 57% split (overall average of 48%), rather than a 47% / 47% split? Or, does it necessarily follow that the split MUST be no less than 48% / 48%?
Tuesday, November 23, 2010
Situational wins of Barry Bonds’ walks
I’ve been meaning to write this article since forever, and now I don’t have to, because Studes did a bangup job on it.
When to walk Josh Hamilton
Ah, I love fishermen, as I get to eat the fish without effort:
The Book goes into great detail regarding the iBB; the idea is that given a particular game/out/base stat, the hitter at the plate’s expected wOBA versus the pitcher needs to be a certain higher proportion than the following hitter(s) in order for an intentional walk to be the right move.
...
Of course, if you think Hamilton is better than a .400 wOBA hitter versus righties, it changes things. If you Hamilton is a .430 wOBA true talent hitter versus RHP, the ratio of that to Vlad’s .358 is about 1.20. But even though the situations where an intentional walk is called for increases, most of them are still in the latter innings with the pitching team behind, one out, and runners on second and third. Again, in many of those situations the Giants probably be able to bring in Javier Lopez to pitch to Hamilton.
Wednesday, October 20, 2010
A Good Intentional Walk Day
Poz is quickly becoming my go-to guy for sabermetric scenarios. IBB. I like the chart he made that ended with:
Murphy—runner on second, two outs.
That’s a total of 139 runs. The inning continues 35% of the time.Molina—runner on first and second, two outs.
That is a total of 153 runs. The inning continues 29% of the time.
That was per 500 PA. So, he’s saying 0.278 runs for Murphy hitting and .306 runs for Molina.
After that, you have to add in the “ending” base/out state following their plate appearance. I’m feeling lazy, so I’ll wing this part. Let’s see, when they reach base, there’s another close to .30 run potential in the inning. And if they make out, well, it’s zero. So, you add .30 runs x 35% = .105 runs to Murphy, and add .30 x 29% = .087 runs to Molina. We have: .278 + .105 = .383 runs for Murphy, and .306 + .087 = .393 runs for Molina. That’s pretty darn close for such a quick look at it, with a slight edge to Molina.
What does The Book say? Let’s see, top of sixth, runner on 2B, 2 outs, pitching team ahead by 1, go to Table 126, the breakeven point is “1.35”. That is, if the batter at bat has a wOBA that is at least 1.35 times the on deck batter’s wOBA, you have a justifiable IBB scenario. Benjie Molina has a career .309 wOBA, but is 36 years old. Let’s say his true talent level is .290. David Murphy has a career .349, and should be near his peak. Let’s make him .350. On top of which, we gain the platoon advantage. Let’s say that means against LHP Murphy is .360 and Molina is .280. And .360/.280 = 1.29. That’s just below the breakeven point, and therefore, a slight edge against issuing the IBB, as Poz’s method above also determined.
I mean, it’s close enough that if you think Molina is really a .270 wOBA hitter in that case and Murphy is really a .370 hitter, then it makes it a justifiable IBB situation.
But, I agree with Poz’s main thrust that I despise the IBB.
Do managers have much influence over their team’s pythag o/u?
An interesting study from last year suggests that they don’t (have much influence on whether their team’s record is less than or greater than its pythag record), although there is probably a large uncertainty surrounding the results (basically the “r” of the regression) because of small sample sizes (20 managers with at least 10 years of managing).
Then again, these are long-tenured and thus, likely successful/winning, managers (in fact, as a group they outperform their teams’ pythag record, which is to be expected), such that they may be a sample of managers who DO have a talent for outperforming their pythag records, and those that are not included in the study, especially those that go from team to team or don’t manage very long in the majors, are the ones who have a talent (or lack thereof) for underperforming.
IOW, if you have pairs of observations with a high correlation (or at least SOME statistically significant correlation) and you create a subgroup of those observations that has high or low values, and then you run a regression on that subgroup, you might find little or no correlation, either because the spread of talent in that subgroup is so small, your sample size is too small, or your original distribution was bi-modal (lots of high talent and lots of low talent, but not much in between and not much spread within the low and high ends).
It would have been much better if he used all managers and all years…
Tuesday, October 19, 2010
Upset at not bringing in Rivera in a low leverage situation?
Neyer is absolutely correct:
This time, there was a tomorrow. Two tomorrows. You want to guess how many times in 2010 Rivera has pitched on three straight days?
Zero. He hasn’t done it, not even once. And while the Yankees scored so many runs that save opportunities were scarce in 2010, it’s a fact that Rivera didn’t pitch on three straight days even once. And Girardi might be loath to ask him to start now. Especially with two big games coming up, in both of which Girardi might want the greatest postseason relief pitcher in history to toss a couple of innings.
I’m usually a big proponent of being aggressive with your best relievers in big games. But with two terribly important games right around the corner, this might have been the time to keep it in the holster and live to fight another day. Or two.
The Leverage Index of being down by two entering the top of the 9th is 0.3. You would like to bring in your ace reliever when the LI is at least 1.5, and preferably at least 2.0. Anything under 0.7 is a definite non-ace condition. The chance of winning the game is only 6.6%.
Now, if there were no games the next day, then, sure, you can use that bullet, because you get to replenish that bullet. But, with two more games to go, it’s not clear that that bullet will get replenished in time. Or, as Neyer notes in his article, bringing in Rivera down by 1 and with runners on base would have been defensible. The LI there would at least have been over 1.0.
Putting in Rivera would have been a desperation move at a point in time when you don’t need to be desperate.
Sunday, October 17, 2010
Bunting against great pitchers
Poz has a post about the breakeven points about bunting. As we know, The Book has 51 pages on various parameters that need to be considered. I don’t remember if MGL discussed the run environment. Anyway, Poz has a post about it, with regards to bunting against Mariano Rivera.
The basic rule is that the lower the run environment, the more small-ball makes sense. So, for a guy like Rivera, where scoring from 1B with 0 outs is pretty low, but scoring from 2B with 1 out is only low, then it will often be the case that bunting in that situation makes sense (base more valuable than the out). If you need to think about it in a general sense, the value of the base (between 1B and 2B, or 2B and 3B) is pretty static, at 0.17 runs per base. But, the cost of the out goes way down, the lower the run environment. Whereas for the average pitcher, the cost of the out is .28 runs, the cost of the out for a great pitcher is under .20 runs. And, if we are to believe that Rivera is superhuman in the playoffs, the cost of the out goes down substantially.
There’s a time and a place for a sac bunt, and when a great pitcher is on the mound, that time happens alot more than when you have a bad pitcher.
***
Oh, and just as an aside, I saw that Rivera has given up two (!!!) homeruns in his post-season career (500 batters faced). His batted ball distribution in the playoffs since 2002 mirrors his regular season career totals, so we can extrapolate that for his post-season career. With nearly 400 contacted balls, that would mean about 110 flyballs allowed, or 11 HR expected if he had no skill there. His actual performance is 3 SD from the mean. Of course, his regular season career is also obscenely low: 62 HR on close to 1000 FB, which is also 4 SD from the mean.
As a general rule, I don’t like any stat that excludes the number of HR a pitcher allowed. I understand why the stats do that, because at some level of sample size, the correlation is under r=.50, and therefore, given the choice between the pitcher’s rate for that sample, and the league average, you should choose the league average. However, we don’t need to restrict our sample to seasonal data. Rivera has given up 1000 flyballs in his career, and that means his HR rate per FB tells us more than not knowing it at all.
Monday, October 11, 2010
Did Cox make the right decisions in game 3?
If you are a Braves fan, you are justifiably emotional about last night’s game.
Here are the 3 major criticisms I have read:
1) Not pulling Conrad for Diory Hernandez in the 9th, for defensive purposes.
2) Kimbrel throwing a slider to Sanchez on the 1-2 count.
3) Choosing to bring in Dunn, the lefty reliever, to pitch to Huff, who has a higher BA against lefties this year.
OK, number one is a legitimate beef I think. I do no know how good a defender Hernandez is. I have a feeling that Cox never even thought of replacing Conrad in the 9th. I don’t think any of the commentators did either. No one remarked about it on TV that I remember.
Despite Conrad’s recent woes in the field, is he really that bad of a fielder? According to Chone’s TZ for the minor leagues, he was a slightly below average defender at 2B. Now maybe, with all his miscues recently, he was nervous and thus a terrible fielder at the time. I don’t know. Only someone like Cox might know that. And of course if Conrad is much the worse fielder, how much does that actually cost in terms of WE?
Number two is somewhat of a legitimate beef I think, although it is very much a 20-20 hindsight thing. If Sanchez had gotten out on that slider, no one would have said a thing. Plus, it was a little bit of a hanger. Had Kimbrel thrown that down and away, like he wanted to, again, no one likely would have said a thing. Yes, he has a good fastball, and yes, Sanchez had just swung at some fastballs out of the zone (and missed them badly), but is he supposed to throw all fastballs? He has a good slider. If he had thrown a fastball right down the middle and Sanchez had gotten a hit, fans could easily have criticized him for throwing too many fastballs. I did not think that the slider was a terrible pitch to throw in that situation. He obviously missed his location which can happen with any pitch. I don’t think there is much of a legitimate criticism here. A little maybe, as I said, but not a lot.
#3 is way off the mark, I think. There are several reasons why. First of all, despite the fact that Huff is hitting better versus lefties, BA-wise this year, all lefties batters have positive platoon splits. Virtually no lefty batter has a reverse platoon split. In fact, if you look at Huff’s last 3 years, he has a 1.15 platoon ratio for OPS, and even a 1.08 ratio for BA. If we go back 5 years, it is even higher than that. So, the idea that Huff’s higher BA against lefties in 2010 means anything (predictive) is nonsense.
Now, when you bring in a same-side pitcher, it is not just the batter’s true platoon ratio that you care about, but the pitcher’s as well. Let’s say that Huff had a 1.00 ratio or even a reverse one. If you brought in a pitcher with a large platoon ration, it still might be correct. The platoon ratio for a particular batter/pitcher matchup, like anything else, is roughly a log-5 (odds ratio) combination of the pitcher and batter’s ratio.
And obviously the overall true talent of the pitcher has to be considered as well. If you bring in an overall (much) worse pitcher just to get a platoon advantage, it still might not be correct, although when you bring in lefty against lefty, in most cases you gain so much from the platoon advantage that even a mediocre overall lefty becomes a very good pitcher when facing a lefty batter.
So what about Dunn? What is his true platoon ratio? In a small sample, his MLB career platoon (OPS against) splits are large - 1.29.
For all pitchers, I compute estimated platoon splits from their actual numbers in the majors and minors, regressed the appropriate amount, given the types of pitches (and frequencies) they throw. For Dunn, I have that number at 1.27. For Kimbrel, based on his minor league stats, I have that at 1.31 (mostly because he throws the 2 pitches with the highest platoon ratios - fastball and slider). So it looks like Dunn is WAY the better choice for Huff than Kimbrel. What about Moylan or Farnsworth? Moylan is a siderarmer. He has an estimated true split of 1.41. And Farnsworth’s is 1.21, also a large platoon split, probably because he throws mostly fastballs I think.
Finally, I also, in my computations, estimate the average ERA that a reliever would have against RHB and LHB with average platoon splits:
Dunn
vs. RHB: 3.80
vs. LHB: 3.03
Kimbrel
vs. RHB: 2.91
vs. LHB: 3.75
Farnsworth
vs. RHB: 3.23
vs. LHB: 3.93
Moylan
vs. RHB: 3.51
vs. LHB: 4.86
So, it looks like Dunn was clearly the right choice, even if Huff had a slightly smaller platoon split than the average LHB (and there is not much evidence that that is the case - in fact, my estimated OPS true platoon split for him is quite high - 1.21 ).
(16) Comments • 2010/10/12
• Sabermetrics
• Batter_v_Pitcher
• Fielding
• In-game_Strategy
• Platoon
Sunday, September 26, 2010
I wonder how third base coaches think about these things…
In the Yankee game last night, down 4-0, with one out in the sixth, the Yankees’ third base coach Robby Thomson waved Kearns in on a Jeter single to left. He was thrown out at the plate for the second-out, on a fairly close play.
“I thought he was going to score and he didn’t,” Thomson said. “It was a mistake.”
Now, we don’t know what Thompson was really thinking when he said that (hence the title of this thread), but it suggests (if he was being completely forthright) that he is being results-oriented, which is exactly the wrong way to be as a coach or manager, especially for a third base coach.
You also have to wonder what Girardi said to Thompson after or during the game. Surely he said something.
I was watching the game and at first it looked like Kearns had little chance of scoring (maybe 25%), but it was a good throw and the play, as I said, was fairly close, so maybe he had a 75% of scoring. I am not really sure. It could have been 50%.
I don’t, off the top of my head, know the BE point for 1 out and down 4 runs in the sixth, with those teams and in that exact situation. It is probably pretty high, maybe 80%, given the large run deficit.
In any case, that is the way Thompson (and all third base coaches) is supposed to think. Maybe they don’t need to memorize the exact (not that ANYONE knows them exactly of course) BE percentages, but they should all have proper rules of thumb for just about any situation. Again, whether Thompson or Girardi does or doesn’t, I don’t know. His comment (mildly) suggests that he doesn’t and that his thinking is not that which you would want your third base coach to have ("he got thrown out, therefore I made a mistake").
My guess also is that there are at least a few organizations that go over this stuff properly (sabermetrically) with their third base coach (among other things like that), but not too many. (And of course even if you “know” the proper numbers for every situation, you will still occasionally make an error in judgment in the heat of the battle.)
If a team does do that, I wonder how many runs/wins it would add per season. And then of course, think of the other similar things they could discuss and plan with their manager and coaches in order to optimize strategy.
The other day, Carl Crawford, I think, was thrown out at third base with 2 outs to end a game (maybe I got some of that wrong). He and the Rays were criticized for that - considering the adage that, “You don’t want to make the first or third out at third base.”
Joe Maddon was talking about that play on the radio and he said that he was perfectly fine with Crawford’s play (and I am pretty sure that he told him so). He said that it took an absolutely perfect throw to nail him, which it did, implying that there was a very high chance that Crawford would be safe if you redid that play X amount of times (again, correctly so).
He added that he does not go by the old adage mentioned above, again implying that it is all about the chance of being thrown out (going strictly by the adage, the BE point for being safe would have to be 100% with 0 our 2 outs, which it is not of course) given the outs, runners, score, inning, etc.
I don’t hate everything about Maddon, and I have always said that he is a very smart guy. I should probably also add that I think he is very knowledgeable about sabermetrics. I may even be wrong about him overall. When I say things about managers and GM’s and teams, there is a very high uncertainty level surrounding what I say, given my lack of information, the emotion of the various situations, the fact that when you see one bad play, it gets magnified, etc.
Thursday, September 16, 2010
Derek being Alex
Poz’s take.
My take is that these two are equivalent, and no line can separate them:
Sunday, September 12, 2010
Sac bunting on 3-0?
The league has a wOBA of about .570 or so on 3-0 counts. Basically, better than Barry Bonds at his best. Would YOU ask Barry Bonds to sac bunt?
Tuesday, August 31, 2010
Typical bad managing by a bad manager…
Giants are winning 1-0 in the bottom of the 8th with their starting pitcher, Sanchez, leading off the inning at bat. He has already thrown over 100 pitches. Sanchez is a good but not great pitcher. Somewhere around league average.
They let him bat. Nothing happens in the inning for the Giants. In the top of the 9th, he walks the leadoff batter and they immediately take him out. What was the point of letting him bat? Either you think he is still good enough to pitch the entire 9th, whether he gets the first batter out or not, or you don’t. Not to mention the fact that the next batter (after the walk) is a lefty (Gonzalez).
IOW, if you plan on taking him out if the first batter gets on, then obviously you think that he has little or nothing left in the tank. If that be the case, pinch hit for him and then bring someone else in in the 9th.
Bochy’s decisions were clearly of the “I’ll do whatever it takes to avoid criticism,” rather than actually think about what are the best moves to help his team get into the post-season.
First, “I won’t take my starter out while pitching a shutout, lest my relievers blow the game and I get lambasted for that.”
Second, “If the first batter gets on, I’ll take him out, lest I be accused of leaving him in too long.”
Third, “I’ll bring in my closer, Wilson, against the lefty batter, Gonzalez, even though my closer threw over 30 pitches the night before, lest I lose the game without bringing in my stopper.”
The other reason for bringing in the lefty to pitch to Gonzalez, or even leaving Sanchez in, is to keep Fowler from stealing (a generally underused strategy - bringing in a lefty to keep a runner from stealing second).
Needless to say, it all blew up in Bochy ‘s face, and they lost a game that they could ill-afford to lose…
Interesting play
In the Boston/Tampa game tonight, in the bottom of the 7th with Boston leading 1-0 and a runner on third and 1 out, Matt Joyce hits a clearly foul fly ball to deep RF. Drew makes the catch and the runner on third scores easily.
Drew could have not caught the fly ball. Should he have? The count was 3-2 on the batter and Dan Johnson was on deck. Buchholz was still on the mound for Boston.
Of course all teams should know the answer to that question and should be relaying it to the outfielders. My guess is that virtually no teams do that.
Thursday, August 19, 2010
Game theory on the squeeze
Drayer reporting:
The bunt that beat the Mariners Monday night surprised many but was not a complete surprise to those on the field at the time. Adam Moore said Adam Jones did tip them off on the previous pitch. When he did, there was a decision to be made.
“On the first pitch he kind of scooted up in the box and that way you can either go up and in with another fastball but then you are probably looking at 0-2 with Adam Jones so I thought a slider down and away would be a tough one for him to get it out there and get. And he did. A great bunt and tough loss.”
Josh Wilson saw the scoot also.
“It looked like he was doing something on the first pitch. I didn’t know if it was a bluff or whatever. In that situation you aren"t expecting many guys to have the guts to lay down a bunt for a hit to get him in.”
I asked if they saw it why the corner infielders weren’t pulled in?
“A lot of times guys will fake it to try and get you to come in there. It’s the first bunt for a hit I have seen to win a game. I’ve never seen it before. You pick your poison going in. If he comes in he may swing and as a defense that puts you at a disadvantage because you are playing in with a guy with a heavy sinker and then a lot of times you end up getting stuff chopped through a hole that wouldn’t otherwise be there. It’s a tough call.”
Glove-slap Jeff.
Wednesday, August 18, 2010
Should a pitcher take three strikes, rather than (possibly) swing into a DP?
Asked of Bill James:
I have a feeling you covered this at some point, but… Bases loaded, one out, .120 hitting pitcher at the plate, .370 obp leadoff hitter on deck. Would it make sense statistically to have pitcher take 3 strikes rather than risk hitting into a DP?
Asked by: Tony K
Answered: August 14, 2010Well. . .the question assumes that attempting to bunt is not an option. As long as you have the option to attempt to bunt--which you always have--that seems to obviate this situation.
If, for some reason, you COULDN’T bunt. . .my guess is that you would be better off to accept the strikeout, rather than risk the double play.
Starting with an average baseline, we know that the run expectancy of BL (bases loaded), 1 out is 1.65 runs. A strikeout brings it down to .815 runs. Because the top of the order is batting, let’s say it’s .90 runs with 2 outs, BL. And because the pitcher was batting, let’s say it’s 1.35 runs.
So, a DP loses 1.35 runs, a strikeout loses .45 runs, and a weak single adds 1 run.
If you have a pitcher that gets 10% singles, 40% DP, and 50% outs, that’s a run value of -.665 runs. If he strikes out 100% of the time, it’s a loss of .45 runs.
So, yeah, you can easily construct a case to take the strikeout rather than swinging away. That you can construct a reasonable scenario for the “intentional strikeout” tells me how horrible it is to have the pitcher batting to begin with.
***
You may enjoy this link as well (run expectancy by batting order):
http://www.tangotiger.net/retrosheet/reports/re_bo.htm
Recent comments
Older comments
Page 3 of 342 pages « First < 1 2 3 4 5 > Last »Complete Archive – By Category
Complete Archive – By Date