Sabermetrics is about more than just statistics. At least I think it is. To me, it’s a philosophy. Sabermetrics is a way of thinking about baseball that challenges the way things “have always been done.” A central tenet of that philosophy, and one that has caught the attention of Mets fans due to our new GM’s espousal of said philosophy, is that the manager is “middle management.” Perhaps unsurprisingly, this has proved controversial. But it really isn’t.
To illustrate, imagine you own a struggling computer company, let’s call it Mets Inc. After years of suffering, you somehow convince Steve Jobs to leave Apple and join your company. It’s well known that Jobs knows how to build products that consumers want.
Jobs’ first task at the company is to hire a new head of product engineering. This person will be responsible for overseeing your company’s 25 engineers on a daily basis. The engineers themselves are ultimately the people who design the products people buy, and largely determine the success of the company.
Even though Jobs will not be the head of engineering, it’s his vision that you as an owner of Mets Inc. were seeking when you hired Jobs. And Jobs is also the person who will hire the actual engineers. However, he can’t supervise them daily–that’s the head of engineering’s responsibility.
In such a scenario, no one would argue that Jobs should hire someone who is opposed to his philosophy of engineering, right? Of course they wouldn’t. The head of engineering is important, but he or she has to fit, and more importantly implement, the overall philosophy laid out by Jobs.
It’s no different in baseball. Sandy Alderson believes that many of the things in the so-called baseball “book” are simply not borne out by probability. For example, it makes little sense if you get a runner on first with nobody out to bunt that runner over to second (unless the pitcher’s at bat). The probability of scoring with a runner on first and no one out is about 41% in a given inning. A runner on second with one out? The probability drops to only 38%. (h/t Teaching Statistics Using Baseball by Jim Albert). Sound insignificant? Look at the below chart, which shows the number of runs you’re likely to score in an inning based on particular situations:
| No one on | Runner on 1st | Runner on 2nd | Runner on 3rd | Runners on 1st and 2nd | Runners on 1st and 3rd | Runners on 2nd and 3rd | Bases loaded | |
| No out | 0.48 | 0.85 | 1.08 | 1.35 | 1.38 | 1.63 | 1.90 | 2.14 |
| 1 out | 0.26 | 0.5 | 0.65 | 0.94 | 0.85 | 1.13 | 1.32 | 1.50 |
| 2 out | 0.1 | 0.22 | 0.31 | 0.38 | 0.41 | 0.49 | 0.54 | 0.66 |
The bunt in the situation described decreases the probable number of runs scored in an inning. But it’s worse than that. The best case scenario if you bunt–success–leaves you likely scoring 0.65 runs. If you fail, you’re at 0.5 runs (or you may bunt into a double play, dropping it all the way down 0.26 runs). If you swing away, the worst-case scenario is the same. But, the best case scenario is much, much greater. Even if there’s just a single and the runner doesn’t make it to third, you’re looking at an average of 1.38 runs in the inning (or more than double with a successful bunt). If you can get a double and put runners on 2nd and 3rd, your probability goes up to nearly two runs scored. Note that we can also see that bunting with runners on 1st and 2nd and no one out also isn’t a great play (if you want or need to score more than a run).
All of this is to say that there are things we know about baseball. Sandy Alderson has been successful in part because he’s put these concepts to use. It is hardly unreasonable to want a manager who will run the team accordingly. Does that mean the manager needs to be a dull robot? It does not. In fact, I would argue that recognizing that the appropriate role of the manager with respect to strategic direction places a premium on other important characteristics–such as the ability to relate to players, dealing with the media, and yes, having an identifiable personality. What we don’t want is a manager who when given the design for the ipod, goes ahead and builds a Zune.
144 comments
Mr North Jersey
11/8/2010-10:08am at 10:08 am (UTC -4)
ok here is the problem with your chart.
It states that with a runner on 1st and no outs your likely to score 0.85 percent of the time but it doesnt say how much of that 0.85 percent of the time was due to a bunt advancing the runner to 2nd.
The stat you needed to show was how many times did a runner score after being bunted over to 2nd. That would of been a better indicator than using what you did to try and explain the reason why bunting a runner at 1st with no outs is less likely to score you a run.
TRS86
11/8/2010-10:18am at 10:18 am (UTC -4)
I think what he is trying to show is that you decrease your chances of scoring by bunting the runner to 2nd because a runner at 2B with 1 out has only a 65% chance of scoring. Thus by bunting a runner to 2nd you decrease your chances of scoring by 20%.
Mr North Jersey
11/8/2010-10:20am at 10:20 am (UTC -4)
still the fact remains that the runner on 1st with no outs does not accurately display what percentage of those were due to being bunted from 1st to 2nd.
TRS86
11/8/2010-10:22am at 10:22 am (UTC -4)
Not sure why that would be entirely relevant however. His stats show you have 85% chance of scoring with 0 outs on 1B and only a 65% chance if the runner is on 2B with 1 out. Thus bunting a runner from 1st to 2nd with 0 outs lessens your chance of scoring.
Mr North Jersey
11/8/2010-10:24am at 10:24 am (UTC -4)
Because the 85% does not show how much of that is due to bunting and that is the entire point of the chart.
TRS86
11/8/2010-10:29am at 10:29 am (UTC -4)
No matter how you look at it, IF you bunt a runner to 2B it decreases you chances based on those stats. What does it matter what percentage of the guys at 1B score due to bunting. That’s not what the chart is showing. It is showing that you have a better chance of scoring with 0 outs on 1B than with 1 out on 2B. Unless you can bunt with 0 outs and a runner on 1B and not make an out then you just decreased your chances of scoring by 20% according to the chart.
What am I missing here?
Mr North Jersey
11/8/2010-10:38am at 10:38 am (UTC -4)
I have explained it as best I could I don’t know what is so difficult to understand but in any event maybe someone else will and they can try explaining to you what i mean.
TRS86
11/8/2010-10:40am at 10:40 am (UTC -4)
I guess I am just missing it. Seems pretty clear to me that if you bunt the runner to 2B then you decrease your chances by 20%.
stickguy
11/8/2010-10:56am at 10:56 am (UTC -4)
NJ, if you look at the base numbers, you have a better chance of scoring from 1st, no outs. That includes giving up the out to sac. If you take that out, the % should be even higher (since the remaining occurances are more positive).
Not sure that makes sense the way I wrote it, but the logic s, even giving up an out sometimes, you still score more the rest of the time.
Prismo
11/8/2010-11:21am at 11:21 am (UTC -4)
It’s not 85%.
It’s .85 runs on average – the chart is runs, not percentage of scoring.
Mr North Jersey
11/8/2010-10:19am at 10:19 am (UTC -4)
Keep in mind if you hired Jobs to work at Mets inc. it is without a doubt he would hire people he feels comfortable with and trusts that they can do the job. This is standard business 101 in my opinion. This is something that all mlb teams do it has nothing to do with some new philosophy.
It’s about finding the man that can best do the job where Alderson will have trouble. He did it with Tony LaRussa and Tony Larussa has been a good manager even after Oakland. He failed with his 1st 2 managers Steve Boros and Jackie Moore.
Ceetar
11/8/2010-10:20am at 10:20 am (UTC -4)
I know the bunting thing is one example, and one i agree with for the most part. However, it’s still flawed, as all stats are. I want a manager that’s most of the time going to do the smart thing, and not bunt and give away outs, but is connected to the game and his players enough to know when maybe he should do things outside of statistics.
For one, that chart represents all of MLB. the Mets don’t have all of baseball due up after the bunter. (For arguments sake, Reyes on first, Castillo bunting) Perhaps it’s Beltran, Wright, Bay. The runs scored for not-bunting pr obably is higher than that. But perhaps it’s an off day, or 2009 and you’ve got J. Feliciano, Daniel Murphy, and Josh Thole coming up. Perhaps Murphy is slumping too. Perhaps in this situation it might make sense to bunt. Maybe it’s the 7th inning and you’re down by one at home. maybe it’s a tie game, and the bullpen has been lights out lately and you really are focused on one run (this may never be a sound philosophy, but just saying..)
So while I want am manager that understands statistics, I also want a manager that sees a lefty on the mound, but instead of pulling the lefty batter for a pinch hitter, notices that the pitcher is a bit off (or maybe tipping his pitches?) and trusts in his guy to get the job done.
A manager is more than a computer, and has to be able to read the small sample size. Players don’t have a hot week because of statistical quirks involved with small samples. It’s a constant game of adjustments and reading pitchers and pitches and sometimes you get hot. streaks exist.
Mr North Jersey
11/8/2010-10:22am at 10:22 am (UTC -4)
Well said.
Mr North Jersey
11/8/2010-10:33am at 10:33 am (UTC -4)
I don’t know if you can do this “Metsfan20XX” but do you think you can find out of that 0.85 percent of runners that score from 1st with no outs is broken down?
How much is due to bunting and how much is due to sacrificing and how much is due to a wild pitch and how much is due to a error and how much is due to a hit or a walk following the runner getting on 1st with no outs to begin with?
That would be a better indicator on understanding how one gets that 0.85.
TRS86
11/8/2010-10:35am at 10:35 am (UTC -4)
I agree that info would be interesting, however no matter what the stats still show that if you bunt you decrease your chances by 20%.
Mr North Jersey
11/8/2010-10:39am at 10:39 am (UTC -4)
You still dont see it
TRS86
11/8/2010-10:40am at 10:40 am (UTC -4)
Nope.
TRS86
11/8/2010-10:41am at 10:41 am (UTC -4)
I just don’t see how it’s relevant that some of those runners on 1B with 0 outs scored with a bunt to 2nd. It’s a smaller percentage because once you arrive at 2B with one out you have a lesser chance of scoring.
Mr North Jersey
11/8/2010-10:55am at 10:55 am (UTC -4)
you keep looking at it form 2nd base when it is the 0.85 from 1st that you need to be looking at.
The chart says that you have a greater chance of scoring from 1st with no outs as opposed to scoring from 1st with 1 or 2 outs. That is a rather obvious stat by itself.
What it doesnt say is how did it come to be that a runner on 1st scored with no outs 0.85 of the time.
How much of that is due to a bunt?
Why is that important?
Because the point being made is bunting is a bad strategy with no outs and unless you know how much of that 0.85 is due to bunting “specifically” How can you know how much bunting had to play in achieving that number?
Now if you look at runner scoring from 2nd with 1 out it says you stand a 0.65 chance in scoring. Again another obvious stat in that the less chances you have to score regardless of where you are on the basis the harder it will be to score.
So of course a runner at 3rd with no outs will tend to score more times than a runner at 3rd with 2 outs. This is something we all agree on I would hope.
But the point is that the charts focus is on the bunt specifically. so if a runner scores from 1st with no outs 0.85 of the time then what is relevant is what percentage of that 0.85 was due to a bunt?
TRS86
11/8/2010-11:00am at 11:00 am (UTC -4)
And I am saying that regardless of how many of those scored due to a bunt, once bunted over to 2B they have a 20% less chance of scoring.
Mr North Jersey
11/8/2010-11:01am at 11:01 am (UTC -4)
You still don’t see it.
TRS86
11/8/2010-11:05am at 11:05 am (UTC -4)
NJ I think we would continue to go around in circles on this all day. IF there is a runner on 1B with 0 outs they have an 85% chance of scoring. Some of those will be by HR, steals, 2B, bunts, errors whatever. HOWEVER, once they are bunted to 2B they have a 20% less likely chance of scoring.
Mr North Jersey
11/8/2010-11:08am at 11:08 am (UTC -4)
It’s ok I tried to make another attempt into trying to explain myself but I am failing at it.
Maybe with time someone else may see what i was trying to explain and can better get the point across that I currently am unable to do so.
GravediggerHebner
11/8/2010-11:07am at 11:07 am (UTC -4)
Gentlemen, despite my inclination to just stay out of this due to my own lack of knowledge of and ability with mathematics, I don’t think the chart is showing “percentages” at all but rather “number of runs you are likely to score” which could probably be broken down into percentages but isn’t being displayed here as such.
For example, I don’t think the “2.14″ in the “no outs bases loaded” category is to suggest that in that scenario you will score one run 214% of the time, but rather that on average when a team has the bases loaded with no outs it winds up scoring 2.14 runs.
Mr North Jersey
11/8/2010-11:12am at 11:12 am (UTC -4)
I believe that is correct I kept referring to it as percentage of the time but Metfan20xx clearly says the number of runs you will score.
metsfan4decades
11/8/2010-11:15am at 11:15 am (UTC -4)
That was my take on this as well.
As far as scoring at a higher % from first with no outs, as opposed to a lesser % on second with one out….I would think this would be rather obvious. I would have thought that the odds of scoring from any base (be it 1st, 2nd or 3rd) with a lesser amount of outs would be a higher % and go down with each out. Which…I believe is what this chart says. (warning: definitely mathematically challenged fan here).
Ceetar makes some good points as well. Overall, if I never see a bunt again in the first inning by the 2nd batsman after the leadoff guy gets on, it will be too soon. IMO, Jerry relied on bunting way too much.
Ceetar
11/8/2010-11:19am at 11:19 am (UTC -4)
I don’t even agree with the pitcher bunting in the the first inning.
metsfan4decades
11/8/2010-11:30am at 11:30 am (UTC -4)
LOL….if the pitcher is bunting in the first inning and there are less than two outs, that opposing starter is really having a bad 1st inning – so yeah, I’d let my pitcher bat away also. Heck, with a pitcher that much out of control, have a better chance of our pitcher walking, getting a hit, etc.
TRS86
11/8/2010-11:17am at 11:17 am (UTC -4)
Just noticed that as well. Don’t guess it changes our argument any.
stickguy
11/8/2010-11:12am at 11:12 am (UTC -4)
I think NJ is simply asking this question. Of the guys that scored from 2nd, how many were only there (to be able to score) because they were bunted over from 1st.
or from another angle, if there was no bunt, what would have happened to that guy on 1st base.
TRS86
11/8/2010-11:18am at 11:18 am (UTC -4)
Right but the chart is trying to show that you lessen your chances of scoring by bunting the runner from 1B to 2B which I believe it shows quite well.
kingman 26
11/8/2010-11:23am at 11:23 am (UTC -4)
No it does not.
It is NOT taking into account the total number of the runners who score from 1st with no outs who were moved over by a bunt.
The data is flawed.
Plus, remember, guys can be on 2nd with one out from a double, a single/BB, errors, etc.
I think Mr N is right on the money here guys.
TRS86
11/8/2010-11:35am at 11:35 am (UTC -4)
It does not matter how he got to 2B.
Point still remains that once you bunt him from 1B to 2B his chances of scoring decrease according to the chart.
kistics
11/8/2010-12:15pm at 12:15 pm (UTC -4)
It’s the return on the chance that decreases. That doesn’t necessarily mean that the % of scoring decreases.
Mr North Jersey
11/8/2010-11:22am at 11:22 am (UTC -4)
It’s ok stick i appreciate your attempt at helping better understand.
My focus has been on the singling out of the bunt overall but my attempts have been poor in doing so.
Ceetar
11/8/2010-11:18am at 11:18 am (UTC -4)
NJ: Doesn’t it stand to reason that of the .85 runs you score from first with no outs, .65 of that is accomplished by bunting/sacrificing? (Actually slightly less, thats assumping 100% sac rate)
But that’s basically the best case scenario (a _decrease_ in the rate you score runs)
If you don’t bunt, even if you fail you still have a .5 runs scored number. that’s .15 difference versus nearly double, 1.35 runs, if you simply walk and advance the situation to first and second.
So over the course of everything, unless the guy only has to get a hit once in every five tries. (1 hit increases the runs scored number by .7 (more if it’s more than just a base to base thing) 4 strikeouts decreases it .6 for a net gain of .1 run scored for having a .200 batting average (It doesn’t even make sense for Ruben Tejada to bunt) with a runner at first with no outs.
Ceetar
11/8/2010-11:21am at 11:21 am (UTC -4)
You have to realize that to get a guy at second base with one out basically means you either bunted, or the second guy up hit a double. (single and error, SB, etc)
TRS86
11/8/2010-11:21am at 11:21 am (UTC -4)
Well not really. A the chart allows for a runner to be on 2B with one out regardless of if he ever got to 1B. Thus one out and a runner hits a double then he would still count in the runner at 2nd one out column. However, I thought the chart was trying to show that if you bunt the runner from 1B to 2B you lessen your chances of scoring and it looks pretty clear to me.
kingman 26
11/8/2010-11:25am at 11:25 am (UTC -4)
“NJ: Doesn’t it stand to reason that of the .85 runs you score from first with no outs, .65 of that is accomplished by bunting/sacrificing?”
Uh, no.
kingman 26
11/8/2010-11:20am at 11:20 am (UTC -4)
I have always agreed with the info that this chart specifies, but Mr N’s point is so outstanding that it is making my head hurt thinking about it.
Perhaps enough of those times you scored with a man on first and no outs were due to bunting the man over; then the chart above would indeed be irrelevant.
Great point Mr N! I think TRS is missing the point here. Yes, once they are on second with an out, the chances to score decrease BUT the overall numbers which this decrease are based on DO NOT take into account how many of the times the runners on first with no outs scored after being moved over by a bunt!
I think!
We need more research!
Mr North Jersey
11/8/2010-11:27am at 11:27 am (UTC -4)
Kingy your getting it.
stickguy
11/8/2010-11:28am at 11:28 am (UTC -4)
I think we really need beer.
Ceetar
11/8/2010-11:29am at 11:29 am (UTC -4)
Okay, let’s try it this way. TRS and NJ are talking about two separate things.
TRS is talking after, NJ before. but the chart does not care how you got to second base. On average you score .85 runs when you have a leadoff runner. on average you score .65 runs when you have a runner at second with one out.
If you bunt successfully, every single time you have a runner on first with no outs, you will score .65 runs.
TRS86
11/8/2010-11:34am at 11:34 am (UTC -4)
Right, thus no matter what if you bunt a runner from 1B to 2B according to these stats you lessen your chance of scoring. I thought that was the author’s point all along.
TRS86
11/8/2010-11:31am at 11:31 am (UTC -4)
Not missing it at all. Yes there are runners who will score from 1B with 0 outs due to being bunted to 2B. However, it still applies that no matter what once they are bunted to 2B they have a less likely chance of scoring than they previously did.
stickguy
11/8/2010-11:35am at 11:35 am (UTC -4)
that is true. But, it is also possible that if you eliminated all sac bunts, the .85 number might change. the .65 number would not, since the sac is irrelevant to it (that is, how you got to 2nd base).
TRS86
11/8/2010-11:37am at 11:37 am (UTC -4)
It is possible that if you eliminated all sac bunts the .85 might change but that still does not make the stats irrelevant considering that once you arrive at 2B you have a decreased chance of scoring.
Mr North Jersey
11/8/2010-11:36am at 11:36 am (UTC -4)
I understand how you are coming to the conclusion you are and there was a time I viewed it the same way but allow yourself to step outside the box if you will and focus on the following.
How much of the 0.85 runs scored from 1st with no outs is due to a bunt following them?
kingman 26
11/8/2010-11:38am at 11:38 am (UTC -4)
You are right.
There is zero doubt.
This piece is a CLASSIC example of the limits of sabermetrics and the limits of stats.
Along with stats which say that Bay is a bad outfielder or Pagan is a top 5 anything.
TRS86
11/8/2010-11:54am at 11:54 am (UTC -4)
Wow. No. However, the discussion does illustrate how there are many different ways you can take and use stats and they can mean different things to different people. Now we are going to have the saber and non-saber stats debate. LOL. Both are relevant. No use in using one damning stat to damn them all.
Ceetar
11/8/2010-11:38am at 11:38 am (UTC -4)
Think about waht you’re asking? You score .85 runs on first with no outs. How many runs score if you bunt?
What happens if you bunt?
You get a runner on second with one out.
This is a known condition, and we have a chart for how many runs score. it’s .65.
TRS86
11/8/2010-11:39am at 11:39 am (UTC -4)
Why does it matter? They still get to 2B with one out where the runs drop. I get what you are saying as well. Might as well move on.
kingman 26
11/8/2010-11:36am at 11:36 am (UTC -4)
Wrong.
We need to know how many of the .85 on first with no out who scored got to second by being bunted over. THIS would tell the efficacy of bunting.
If a huge number of them were bunted over, the data is meaningless.
It is possible that MORE of the .65 on 2nd with one out who scored scored after being bunted over than after a 1-out double or a single and an SB or error.
TRS86
11/8/2010-11:41am at 11:41 am (UTC -4)
However, looking at the chart if you are standing on 1B with 0 outs you have a better chance of scoring than you do at 2B with 1 out. That part is clear. Does not matter to me how you get to 2B. 1B with 0 outs is has a better chance of scoring than 2B with one out.
Prismo
11/8/2010-11:43am at 11:43 am (UTC -4)
Yes, and it’s as simple as that. No reason to read into it any more.
Ceetar
11/8/2010-11:34am at 11:34 am (UTC -4)
Obviously, these stats are based on the league, and the league bunts. Therefore that .85 runs scored from first base does factor in bunting. But we know what happens when you bunt. You get a runner at second base with one out. The chances of scoring from there are .65. _that_ is the runs scored of the .85 that happens from bunting (it’s not a percentage, it’s number, so all possible scenarios won’t add up to .85) You actually have a greater number than .85 without bunting.
kingman 26
11/8/2010-11:26am at 11:26 am (UTC -4)
Unless I am seriously missing something, you are totally right Mr N.
And it is a truly outstanding point.
Prismo
11/8/2010-11:27am at 11:27 am (UTC -4)
This is fantastic MF’11. I’m so glad you’re on board and writing some stat-heavy pieces for RDM!
The chart is excellent and truly shows how some myths about small-ball must be overturned. If I may recommend a book: “Hard Facts, Dangerous Half-Truths and Total Nonsense: Profiting from Evidence-Based Management” which is more about business than baseball. But the book actually references Moneyball and how sabermetrics (which is basically evidence-based management) is best for decision-making.
And as TRS said, it doesn’t matter how the runner got to 2B with 1 out, whether by bunt or something else. Once there’s a runner on 2B with 1 out, the situation is the same regardless.
It’s like if you flip a coin 2 times. It doesn’t matter what the first flip is, there’s still a 50/50 shot of heads/tails on the second flip.
kingman 26
11/8/2010-11:34am at 11:34 am (UTC -4)
If a significant number of the runners on first eventually scored after being bunted over, then the above numbers are flawed and meaningless.
Sorry buddy; think about it.
It is indeed possible that if ALL bunting was taken out of the equation, then maybe 61% of runners on first with no outs scored eventually, and 58% of runners on second with one out scored.
Without ANY bunting being figured in.
We don’t know; that’s why these numbers mean nothing, without the info on HOW MANY of the runners on first with no outs eventually scored AFTER BEING BUNTED OVER.
At least I am virtually certain this is correct.
“And as TRS said, it doesn’t matter how the runner got to 2B with 1 out, whether by bunt or something else. Once there’s a runner on 2B with 1 out, the situation is the same regardless.”
This is nonsense without knowing HOW MANY of the runners got to second by being bunted over.
stickguy
11/8/2010-11:37am at 11:37 am (UTC -4)
knog, the only thing that elimnateing (or hell, increasing) sac bunts could possibly change is the 1B, not outs run #. The 2B, 1 out # would not change, since it is not dependant with how the runner got to 2B.
Prismo
11/8/2010-11:39am at 11:39 am (UTC -4)
Sorry buddy, you and MrNJ don’t have the right analytical frame of mind for deciphering statistics apparently. However, unfortunately I don’t know how to explain why you guys are wrong.
I mean no offense, but you guys sound pretty confident/cocky in your corrections, while going up against myself, TRS, our sabermetrician, and a friggin’ statistics book!
Mr North Jersey
11/8/2010-11:42am at 11:42 am (UTC -4)
OK well b4 this gets ugly let me say i apologize if anything i said was viewed as cocky that was never my intent.
Maybe Prismo is right I don’t have the right analytical frame of mind for deciphering statistics to understand such things.
Math was never my strength.
Prismo
11/8/2010-11:44am at 11:44 am (UTC -4)
No need to apologize. That was more aimed at King with his “Sorry buddy; think about it” as if I wasn’t thinking about it in the first place! Agreed, don’t want to see this get ugly.
I also don’t mean to offend anyone’s intelligence or math skills.
TRS86
11/8/2010-12:00pm at 12:00 pm (UTC -4)
I understand what you and NJ are saying. However, it still applies that standing at 1B with 0 outs you have a better chance of scoring than you do once you arrive at 2B with one out.
Mr North Jersey
11/8/2010-11:32am at 11:32 am (UTC -4)
The 0.85 runs you score because you have a runner on 1st with no outs is good and all but if you have to take into acct what percentage of that number is due to a bunt following it.
The fact that you score 0.65 runs with a runner at 2nd with 1 out is not a direct refection that a bunt was responsible for that 1 out.
The continued focus on the bunt is the point that is being missed here. You can not say a “bunt” is a bad strategy unless you can accurately explain how much of thos 0.85 runs were the result of it or even the 0.65 runs at 2nd.
GravediggerHebner
11/8/2010-11:43am at 11:43 am (UTC -4)
If you bunt with a runner on 1B and no outs you are roughly but not exclusively assuring that your result will be a runner on 2B and one out. The opposition can make an error, or you can just beat out the bunt, but a lot of the time you’re just out.
If you don’t bunt you allow for myriad other possible results, some good, some bad. You could walk, you could single, double, triple, homer, get hit by a pitch, reach on an error. Or you could strikeout, flyout or lineout not advancing the runner, or you could hit into a double play eliminating the runner.
What I take from the chart is you are likely to score .20 more runs if you take the chance to allow any of these possibilities to unfold as opposed to eliminating many of them by bunting.
TRS86
11/8/2010-11:58am at 11:58 am (UTC -4)
Bingo.
Mr North Jersey
11/8/2010-12:08pm at 12:08 pm (UTC -4)
Hey Metfan20xx I don’t want you to take away from this that I am anti Sabe-stats that can’t be further from the truth.
While I don’t value stats that deal in predicting the way that Babip does for example i do believe in absolutes like OBP and OPS for example and yes I am of the idea that there is no one way to do things. New ideas are always welcome and have their worth so long as everything is kept in perspective.
To wrap things up my position today was more about that the stats provided don’t correlate well to the idea that a bunt is a bad strategy in my opinion.
Also that the idea of a mgr being middle management is nothing new that was brought about due to sabermetrics but rather that is something that has been part of mlb for some time now.
I write this so you don’t come away with any wrong impressions of me being somehow anti-sabermetrics.
stickguy
11/8/2010-11:32am at 11:32 am (UTC -4)
anyway, to the premise of the article, yes, the manager should be on board with the philosophy set by his superiors int he organization. so if that means he needs to be a middle management company man, fine.
it makes no sense to implement an organizational development stratagy and then have the manager of the big club toss it in the trash and go a different direction.
Now, that does not mean he should be getting daily line ups from above, or phone calls during the game telling him what to do. But he should have an overall managing plan that fits what the team believes in, and which (in theory) should make best use of the players supplied to him.
and if it is corporate policy to not sac bunt in the 1st inning, then by God, he had best not do it!
Mr North Jersey
11/8/2010-11:39am at 11:39 am (UTC -4)
Agreed stick all Alderson is doing is what any GM would do when he comes into a new org. Bringing in the people that he feels will best work with him.
Minaya did it also and so has every other GM in mlb. From the fo to the field mgr. This is nothing new as far as I am concerned.
oleosmirf
11/8/2010-11:45am at 11:45 am (UTC -4)
these are where stats are flawed. it assumes that every batter has an equal chance of driving in runs which we all know is not the case.
it all depends on the situation like using a weak hitter like Tejada, Castillo or Cora or a pitcher to sac bunt a runner into scoring position is still a good baseball play if the on deck batter is a good hitter…
Prismo
11/8/2010-11:47am at 11:47 am (UTC -4)
He did clarify that this doesn’t count for pitchers as they’re such poor hitters. And in giving us that assumption, I think we can also say it doesn’t count for batters who hit like pitchers (aka Tejada and Cora).
Ceetar
11/8/2010-11:58am at 11:58 am (UTC -4)
Nah. If you think about it if Cora or TEjada or a pitcher strike out/make out without advancing the runner, it only drops to .5 versus the .65 of bunting. So you only lose .15 runs by failing. You get .53 runs by walking or singling and making it first and second. (And this is ignoring the extra from a double, a first-third advance, a home run..) That’s works out against the bunt for anyone with higher than a .283 OBP, and that number would get even lower if you factor in the possibility of a bigger hit.
Prismo
11/8/2010-12:00pm at 12:00 pm (UTC -4)
Tejada’s OBP last season was only .305, he’s not far off. lol
Ceetar
11/8/2010-12:07pm at 12:07 pm (UTC -4)
Then again, these numbers are factored out for everyone. So if you’re guy is under league average, lefty on lefty, etc, his runs scored might actually be a bit lower. And this is why in certain situations I might accept a bunt, late in a game, with a poor hitter up.
Hoffman in his prime for example, agaisnt Tejada, Reyes on first down one run in the bottom of the 9th. And even then, probably not.
Ceetar
11/8/2010-11:47am at 11:47 am (UTC -4)
The numbers are so heavily in favor of non-bunts that it doesn’t even favor a good hitting pitcher bunting.
Prismo
11/8/2010-11:46am at 11:46 am (UTC -4)
Another way to read this chart:
For every 5 times you bunt a runner over to 2B and are out, you’re sacrificing 1 run.
That’s huge.
metsfan4decades
11/8/2010-11:54am at 11:54 am (UTC -4)
Well, I’m not sure how you can say that. Wouldn’t that depend on what the percentages were on if you hit away? In other words, what happens with a man on first with no outs, if the next batter does not bunt?
Does the percentages go up, go down, or stay the same that the runner winds up on second and the batter is out? (which would mimic the scenario of a batter getting down a sacrifice bunt).
And by the way….my head now hurts.
Prismo
11/8/2010-11:55am at 11:55 am (UTC -4)
I find when thinking about stats, it’s best to think about it for a few minutes…then take an hour break before trying again.
On a related note, I have no idea what you’re saying.
haha
metsfan4decades
11/8/2010-12:00pm at 12:00 pm (UTC -4)
Concerning your last statement, I can’t say as I blame you.
I think I’m probably one of those that would do better staying out of most of these discussions and just reading and trying to learn something. I have this tendency to start at the end and work my way backwards with analysis, instead of starting at the beginning and going foward. Probably a big part of the reason I’m mathematically challenged.
Ceetar
11/8/2010-11:59am at 11:59 am (UTC -4)
What happens with a man on first with no outs if the next batter does not bunt? .85 runs score.
metsfan4decades
11/8/2010-12:03pm at 12:03 pm (UTC -4)
Um….just asking but is it that cut and dried?
Wouldn’t you have to look at what the next batter actually did, assuming he didn’t bunt?
Not arguing the point as I hate the sac bunt most instances almost as much as I hate the IBB in most instances. I was just asking….
GravediggerHebner
11/8/2010-12:09pm at 12:09 pm (UTC -4)
I’m no statistics expert but I would say, while it’s not quite that cut and dried it’s pretty close and thus the chart and it’s number of runs based on different situations.
If you bunt you eliminate many other possible outcomes, some good some bad. What the chart is telling me is that deciding to eliminate those outcomes is a bad choice.
Prismo
11/8/2010-12:12pm at 12:12 pm (UTC -4)
Grave, you’re 100% correct!
Sure, there is a small bit of risk to these stats. It does depend on who’s at the plate, who’s on deck, the speed of the buy on the base, the chance that a bunt doesn’t necessarily mean an out, who’s pitching, etc. But these are all relatively small variabilities, and can lean in either direction. So it’s not cut and dry, but this is a very large sample size and very clear results.
Ceetar
11/8/2010-12:12pm at 12:12 pm (UTC -4)
yeah. nothing is cut and dried in the future. But what is cut and dried is in that whatever sample this is cut from (not sure if it’s all of the 2010 season, all of baseball ever, etc) when a runner is on first with no outs. .85 runs will score. If Reyes gets on first 100 times in 2011, he’ll score 85 of those times. If you bunt every time, he’ll score 65 of those times.
those 85 runs are actually based on league average bunting. It stands to reason that if you actually separated out the bunting you’d get 65 runs with bunting, and more than 85 without.
Ceetar
11/8/2010-12:18pm at 12:18 pm (UTC -4)
This is an oversimplification of course. It’s not Reyes that scores, but the Mets. that .85 number also factors in the case where you GIDP and then hit a home run.
Mr North Jersey
11/8/2010-12:19pm at 12:19 pm (UTC -4)
exactly
Ceetar
11/8/2010-12:22pm at 12:22 pm (UTC -4)
And I believe what you’re trying to get at, which isn’t what the chart is showing, is that does the percentage of scoring 1 run increase with a bunt? (It doesn’t, but that’s not what this chart is showing) If Reyes leads off, the Mets will score .85 runs. But, sometimes they’ll score 7 in the first. If you bunt, you’re less likely to score 7 runs, which is part of the reason the number is lower. But what’s the % chance of scoring 1 run?
Mr North Jersey
11/8/2010-12:28pm at 12:28 pm (UTC -4)
“And I believe what you’re trying to get at, which isn’t what the chart is showing, is that does the percentage of scoring 1 run increase with a bunt?”
Your close Ceetar, I don’t care to know if the percentage of scoring 1 run increase with a bunt I just want the percentage regardless if its greater or smaller.
kistics
11/8/2010-12:26pm at 12:26 pm (UTC -4)
I don’t think you are correct Ceetar. I think what the table is showing is that if there’s a runner at 1st with no outs, average return is .85 runs. This has nothing to do with probability of scoring.
I think the table is showing the investment and returns. With a runner at 1st with no out, if you invest an out, your average return is reduced by 0.2 runs. But the probability or scoring cannot be determined just by looking at the above table.
It’s a bit like gambling. I think what the author is saying is that given the returns on investment/bet (an out in the case of bunting), it might be smarter NOT to invest/bet (bunt) because the odds are not as attractive. But if you are a conservative gambler (e.g. Jerry), the risk/return may make more sense.
Prismo
11/8/2010-12:27pm at 12:27 pm (UTC -4)
what’s your occupation, if I may?
kistics
11/8/2010-12:35pm at 12:35 pm (UTC -4)
I’m not involved in investments or work for wall st companies. But I play poker and that’s how I can relate to this. LOL
But my job is in corporate finance.
Prismo
11/8/2010-12:36pm at 12:36 pm (UTC -4)
Don’t act as if your finance background doesn’t add to this! You sounded like one of my professors, which is what tipped me off. LOL
oleosmirf
11/8/2010-12:03pm at 12:03 pm (UTC -4)
the real solution to the problem is to not have guys like Castillo or Cora batting 2nd!!!!
Prismo
11/8/2010-12:05pm at 12:05 pm (UTC -4)
Did you see the ESPNNY article from 1-2 days ago where the author said the Mets should sign Eckstein to play 2B because he can hit-and-run so well.
It almost made my head explode with rage.
kistics
11/8/2010-12:07pm at 12:07 pm (UTC -4)
lol
Prismo
11/8/2010-12:09pm at 12:09 pm (UTC -4)
http://espn.go.com/blog/new-york/mets/post/_/id/11795/if-not-lee-or-crawford-or-werth-than
read it and weep
Mr North Jersey
11/8/2010-12:17pm at 12:17 pm (UTC -4)
why the weep exactly? is it because of the likeliehood of no Lee or Crawford or Werth?
I am not sure that is a bad thing myself if that is what you was refering to.
Prismo
11/8/2010-12:22pm at 12:22 pm (UTC -4)
Noooo. Weep because this guy thinks the Mets should sign David Eckstein!
Mr North Jersey
11/8/2010-12:24pm at 12:24 pm (UTC -4)
lol ahhhh
GravediggerHebner
11/8/2010-12:29pm at 12:29 pm (UTC -4)
While in general I agree that this should be cause for weeping, specifically in the context of this post, run scoring and the relation of bunting to that, wouldn’t a player who might be utilized more to hit-and-run as opposed to sac bunting be a good thing?
Prismo
11/8/2010-12:32pm at 12:32 pm (UTC -4)
I can’t remember where, but I think I saw somewhere that Sandy doesn’t like the hit-and-run. I’m going to make the huge leap that that means h&r isn’t statistically favorable.
I’d prefer to see neither hit-and-runs nor sac-bunting (in most cases), and instead a #2 batter who can actually hit!
But you’re right, would be nice to see some more evidence for that.
GravediggerHebner
11/8/2010-12:38pm at 12:38 pm (UTC -4)
Excellent point on needing more data.
It’s a whole other topic in itself but I think given Jose Reyes’ overall (emphasis on recent) injury history, Luis Castillo’s seeming ineffectiveness, and Ruben Tejada’s youth/inexperience/uninspiring offense it’s my personal opinion that one type of player the Mets need to acquire this off season is one capable of playing adequately for potentially extended periods at both 2B & SS (not simultaneously of course).
While this doesn’t have to be one player it would be better if it were. I’m not sure I feel good about Eckstein playing at SS and thus I don’t think he’s the ideal guy for the roster spot. To be fair I haven’t given much thought or study to who I do think is a good target for such a role.
TRS86
11/8/2010-12:48pm at 12:48 pm (UTC -4)
Not sure we could get him but Uribe would certainly fit that role.
Prismo
11/8/2010-12:51pm at 12:51 pm (UTC -4)
Davidoff predicted Uribe gets a 2yr/$12MM contract, for what it’s worth.
I have no idea what it’s worth though.
TRS86
11/8/2010-12:52pm at 12:52 pm (UTC -4)
While sticking with the Giants, not sure if Renteria would be open to manning 2B some but he could be that type of guy as well.
Personally I would much rather stay in house for 2B and for the MI backup.
TRS86
11/8/2010-12:31pm at 12:31 pm (UTC -4)
Yeah I don’t get it. Why would we not just plug Castillo back out there another year if we wanted that type of player. I could almost bet that leaving Castillo at 2B all year gives you the same stats as Eckencrappy.
kistics
11/8/2010-12:06pm at 12:06 pm (UTC -4)
The numbers in the table are not % of scoring runs, but # of runs scored on average, correct?
Prismo
11/8/2010-12:08pm at 12:08 pm (UTC -4)
correct
Prismo
11/8/2010-12:22pm at 12:22 pm (UTC -4)
If I may say a bit about the overarching philosophy of sabermetrics/moneyball/evidence-based decision making, which MF’11 does bring up…
The philosophy is, at it’s core level, that everything should be questioned. Old assumptions and theories should be questions as well as new ones. One of the easiest and most factual-based ways of questioning such assumptions is the use of actual data, aka statistics.
However, and fangraphs nerds may disagree, this does *not* mean that all statistics/sabermetrics are the end-all-be-all to baseball analysis. Going back to the original philosophy, everything must be questioned. This *includes* sabermetrics.
Taking UZR for example, which I think is mostly BS. The first issue is that UZR is meant to quantify range, but most use it as a total defense statistic. The second issue is that the creator(s) of UZR have explicitly said that it only means something when using at least a 3-season average! Now personally, what use is a stat that can only be used as at least a 3-season average? Little use, IMO.
Ceetar
11/8/2010-12:23pm at 12:23 pm (UTC -4)
next to useless.
TRS86
11/8/2010-12:34pm at 12:34 pm (UTC -4)
Agreed, the only thing you can really use it for is to compare similar players over that span but even then it gets murky when you start to factor in how other players could take that out from them in their range OR manager defensive positioning changing their range. Saberstats are just another means to make assertions. They are not the end all be all for sure.
Mr North Jersey
11/8/2010-12:40pm at 12:40 pm (UTC -4)
I think we all can agree that whoever the Met manager will be we will see plenty of bunting a runner over to 2nd with 0 outs and many hit and runs despite what Alderson may have said regarding his dislikes toward those 2 strategies.
Alderson said it best when he said the manager is “not” an extension of himself. He is his own man and as such does things his own way.
Alderson just wants the manager to value certain qualities a player should have and be taught about when playing the game.
The startegy of the game is something that his mgr will decide on a case by case basis.
Prismo
11/8/2010-12:42pm at 12:42 pm (UTC -4)
I wish I didn’t have to, but I must disagree!
As Sandy is attempting to get full-scale autonomy for the Mets organization, I have to believe he’ll hire a manager with similar values.
Prismo
11/8/2010-12:44pm at 12:44 pm (UTC -4)
Or maybe that’s wishful thinking on my part…haha.
Mr North Jersey
11/8/2010-12:44pm at 12:44 pm (UTC -4)
i wonder then how will Alderson react when the manager he hires bunts a runner to 2nd with no outs the 1st time in 2011 or attempt a hit and run?
I will be curios to see that.
Prismo
11/8/2010-12:45pm at 12:45 pm (UTC -4)
Nah, I might be wrong. I guess we’ll just have to wait and see!
TRS86
11/8/2010-12:49pm at 12:49 pm (UTC -4)
Plenty? Why?
Mr North Jersey
11/8/2010-12:54pm at 12:54 pm (UTC -4)
Why did I say plenty you mean?
If that is the case just remove where I said “plenty of” and read what i said without it.
My point is the same.
Prismo
11/8/2010-12:57pm at 12:57 pm (UTC -4)
I think what NJ is saying is that despite what Sandy thinks, ultimately the team does what the manager decides.
kistics
11/8/2010-12:44pm at 12:44 pm (UTC -4)
I posted this up above, but just wanted to post it again. Please correct me if I’m wrong here.
The above table is showing the investment and returns. With a runner at 1st with no out, if you invest an out, your average return is reduced by 0.2 runs. But the probability or scoring cannot be determined just by looking at the above table.
It’s a bit like gambling. The author is saying that given the returns on investment/bet (an out in the case of bunting), it might be smarter NOT to invest/bet (bunt) because the odds are not as attractive. But if you are a conservative gambler (e.g. Jerry), the risk/return may make more sense.
Let me give you another example. Bases loaded with one out will have high probability to score and high returns. But what would you do if you had two choices? (1) Sac Fly – in this case, you are 100% guaranteed to score (let’s assume the runner at 3rd base is Reyes), but your returns (total runs scored at the end of an inning) may not be as much as option #2. (2) Take a chance on swinging for a hit – in this case, the probability of scoring goes down (in case there’s a double play). But the return could be much higher than the option #1 if batter gets a hit.
So I don’t think by looking at the above table, we can determine whether bunting is correct or not. It really depends on how much risk you want to take vs the returns.
Mr North Jersey
11/8/2010-12:51pm at 12:51 pm (UTC -4)
In any event I am glad my views of Metsfan20xx’s chart stirred so much debate and we were able to maintain proper decorum despite some of our differences.
Had I never said what I did I wonder what would of filled this post?
Hat tip for every1 for doing what should not have to even said which is being civil.
Prismo
11/8/2010-12:53pm at 12:53 pm (UTC -4)
+1
Mr North Jersey
11/8/2010-12:56pm at 12:56 pm (UTC -4)
stickguy
11/8/2010-12:55pm at 12:55 pm (UTC -4)
Maybe we can get Metsie over here, to see what his opinion on OBP and sabermetrics is?
Mr North Jersey
11/8/2010-12:57pm at 12:57 pm (UTC -4)
Wow can you imagine that debate here? LoL
TRS86
11/8/2010-12:58pm at 12:58 pm (UTC -4)
http://www.earthstation1.com/WBCartoonFiles/WB_Wavs/ahshedup.wav
GravediggerHebner
11/8/2010-12:58pm at 12:58 pm (UTC -4)
Eff you mother effer!
I mean, well said good sir.
stickguy
11/8/2010-12:55pm at 12:55 pm (UTC -4)
still going I see.
One thing that always bothered me about the sac bunts was that IMO it was managing “scared”, and showed no faith in the team. Instead pf trying for big innings, jerry would give up outs hoping to score 1 run.
And I hate giving up outs, since they are precious. The only thing in a BB game is finite is the # of outs.
TRS86
11/8/2010-12:59pm at 12:59 pm (UTC -4)
Agreed. Now I do see merit at times for bunting a runner to 2B. Pitcher up, terrible hitter good pitcher… Playoffs every thing is magnified and one run could make the difference and you are usually facing the best pitching so it’s up to the manager then to determine the chances of stringing together hits. However, I still think in general as the chart shows it’s a bad move.
kistics
11/8/2010-1:28pm at 1:28 pm (UTC -4)
I think there’s a big difference between managing “scared” vs managing “smart”. Do I agree with all the bunting that Jerry has put on? No. But I do think that you need to consider circumstances before making decisions.
Say it’s a pitchers dual between a healthy Santana and Halladay. It’s 0-0, bottom 7th inning. Reyes gets a lead-off infield hit. Mets 3rd hit of the night. With Castillo up now, you would not bunt in this situation?
stickguy
11/8/2010-1:37pm at 1:37 pm (UTC -4)
bunting or playing for 1 run late in a game is a totally different situation. You can’t have a pitchers duel in the first couple of innings! And Jerry would do dumb stuff like this in the 2nd inning, vs. a scrubeenie pitcher that was struggling.
work the count, get guys on, wear out the pitchers, leads to big innings.
kistics
11/8/2010-1:52pm at 1:52 pm (UTC -4)
I agree with you on that.
Ceetar
11/8/2010-1:41pm at 1:41 pm (UTC -4)
I do not. I let Reyes think about stealing, keep Halladay distracted. I let Castillo attempt to work the count and get it in his favor. I let the possibility of mistake creap in, overthrow to first base, stolen base with an overthrow by the catcher sending Reyes to third. Youv’e opened up the hole on the right side, which is primed for castillo to slap a single through.
TRS86
11/8/2010-1:52pm at 1:52 pm (UTC -4)
Right, Castillo is a contact hitter. You let Reyes try and find a good pitch to steal on and then factor in that Castillo will most likely not strike out and Reyes should be quick enough to stay out of most DP’s.
Ceetar
11/8/2010-1:57pm at 1:57 pm (UTC -4)
Well, Reyes doesn’t have a lot to do with staying out of dp as the runner, more Castillo, who is at least above average fast, and has a lower LD% so the balls take longer to get to fielders, which gives him more time to get to first. (He had/has led the league in infield hits for a long time)
kistics
11/8/2010-4:56pm at 4:56 pm (UTC -4)
But you also add a risk factor by stealing a base now. I know Reyes is a good base stealer, but IF you are purely playing by the stats, you’re increasing the chances of not scoring. In return, you’ll have higher returns (by statistics measure), but I would not risk getting caught there.
TRS86
11/9/2010-7:39am at 7:39 am (UTC -4)
Depends on the catcher and pitchers stolen base percentage along with Reyes. If Reyes is making it 90% of the time and the pitcher catcher are only averaging 10% then …. as those percentages get closer together the more risky it becomes.
Prismo
11/8/2010-1:59pm at 1:59 pm (UTC -4)
I calculated how effective stealing 2nd is with no outs, but class just ended so I’m gonna eat. Be back to post the (very interesting) results in the next hour.
Prismo
11/8/2010-2:58pm at 2:58 pm (UTC -4)
Okay, so…stealing with a man on 1st, 0 outs.
The average stealing rate for the MLB in 2010 was 72%.
That means that after an attempted steal, there’s a 72% chance for a runner on 2nd, 0 outs…and a 28% for no runners on base, 1 out.
Given those odds, the number of runs that would fit into the chart above is…0.85! The same number of runs as having a runner on 1st with no outs! So it doesn’t even matter, on average, whether the baserunner steals or not!!
BUT, I did this again for Jose Reyes. His lifetime steal rate is better than league average, at 80%. If he’s on 1st with 0 outs, and attempts to steal…the average runs for the inning is 0.92! So, with Reyes on base, it IS worth the attempt to steal!
And furthermore, with Reyes on 1st – if you sac-bunt him to 2nd *instead* of letting him attempt the stolen base…you’re dropping your average runs for the inning from 0.92 to 0.65!!!
GravediggerHebner
11/8/2010-11:19pm at 11:19 pm (UTC -4)
Interesting. It’s great to know that this new front office is going to at least take these things into consideration.
Assuming they do hire a manager who thinks along with them I look forward to the moment early in the season when I realize I’m upset that a move wisely chosen because of it’s probability for success fails to work out in a particular instance instead of being upset because “I can’t believe they just did that, that was stupid.”
TRS86
11/9/2010-7:40am at 7:40 am (UTC -4)
I should have read your post as that was what I was saying without breaking down the stats.
njstuckintx
11/8/2010-10:14pm at 10:14 pm (UTC -4)
So, to sum it all up, Statistics are fun!
TRS86
11/9/2010-7:40am at 7:40 am (UTC -4)
Bingo, I look forward to more and more of these post.