Don’t Intentionally Walk Teixeira
In the fifth inning of yesterday’s game between the Seattle Mariners and the New York Yankees, the Mariners decided to intentionally walk Mark Teixeira with runners on second and third and one out in a 1-0 game. With Alex Rodriguez out of the lineup, that brought up Robinson Cano with the bases loaded. Cano hit a grand slam, putting the Yankees up 5-0. At that point, the Yankees win expectancy was over 95%, and a formerly close game was effectively over.
The Yankees scoring after an intentional walk of Teixeira isn’t exactly something new. Yesterday marked Teixeira’s fourth intentional walk. Three times, including yesterday, the result was a grand slam. The other resulted in a run-scoring wild pitch, a strikeout, and then a two-RBI single. Overall, Teixeira and both runners in front of him have scored every single time that he’s been intentionally walked.
Obviously, we wouldn’t expect this to continue to happen every single time, but that doesn’t make the intentional walk a good strategy in this case. Here’s what The Book has to say:
If all batters have equal ability, intentionally walking a batter to set up a double play, force, or other situation is at best a break-even move (or insignficantly better than a break-even move). Doing so early in the game is counterproductive, since it increases the odds of a big inning more than it increases the odds of a scoreless inning.
Now, all batters in the Yankees lineup aren’t of equal ability, but when we look at those after Teixeira, they are all very good. Alex Rodriguez and Robinson Cano are typically the two hitting after Teixeira, making them the important two to examine when it comes to analyzing the value of the intentional walk. The third batter also matters, but far less; this season it has been a mix of players such as Jorge Posada, Marcus Thames, Nick Swisher, and Randy Winn. Overall, though, what matters is that the hitters that come to bat after Mark Teixeira are well above average, particularly Rodriguez and Cano.
Using the Markov chain calculator at Tom Tango’s website, we can take a look at the run expectancy for the composite line of Alex Rodriguez and Robinson Cano, which is essentially what managers are looking at after walking Mark Teixeira. The results, unsurprisingly, suggest that the Yankees are probably going to score a lot of runs after walking Teixeira, even in the situation with one out and runners and second and third, the situation most conducive to a productive intentional walk. Here’s what the calculator spits out.
Runners on 2nd, 3rd, 1 out: 1.578 runs
Bases loaded, 1 out: 1.986 runs
So opposing managers are essentially forfeiting .4 runs for the opportunity to say “Well, I didn’t let Mark Teixeira beat me!” at the post game press conference. The odds of scoring at least 1 run in an inning don’t justify the managers’ decisions either.
Runners on 2nd, 3rd, 1 out: .707
Bases loaded, 1 out: .752
Intentionally walking Mark Teixeira is simply not a smart move. The Yankees lineup behind Teixeira, particularly Cano and Rodriguez, is simply too good, which just magnifies the fact that the intentional walk adds to the possibility of a big inning. Four times, opposing managers have attempted to get out from a jam by walking Teixeira. Four times, at least three runs have scored. If a manager wants to win, his best option is to pitch to Teixeira and hope for the best.
29
So do your calculations say exactly how much worse the next batter has to be to make the walk a good play?
I would guess a pitcher in the NL?
According to The Book there is rarely any situation in which an IBB is correct. The only times in which it might be are when the current hitter and the following hitter are separated by about .130 in wOBA AND the situation is just right, i.e. 2B or 2B and 3B with two out. There are some differences when playing in one-run, ninth inning games but the bottom line is that the IBB is usually a bad idea.
The real bottom line is that most times a team is considering an IBB, they are already in a position where they are likely to give up multiple runs, and looking for a way to get out of it with no runs scored.
The IBB may not be a great move in many cases, but lots of times the D is in a bad spot regardless.
I’m not a big fan of an IBB that loads the bases, because I’ve seen it followed by a WP or a BB too often. It also put the P in a position where he “has to throw strikes” and if he falls behind in the count … here comes a meatball.
But, even without the IBB, the D still needs a lot of good luck to get out of the inning without multiple runs scoring.
You still don’t want to walk someone to face the pitcher because late in the game they can bring in a pinch-hitter and you want the pitcher to lead off the next inning early in the game.
I’d guess walking Teixeira is almost universally terrible but I’d be curious how a pitcher/the lineup opposed affects this situation (i.e. if Pujols didn’t have Holliday behind him or if Bonds was in 2001-2004 form and employed would it still be break even).
Does handedness come into play here at all?
If you are a manager stuck in olde-tyme ways of thinking- then yes.
I think it is incorrect to say that everything that has been done for a while is necessarily wrong. Are you suggesting that lefty/righty splits are always random chance?
Say you have Tim Lincicum (generic lefty strikeout pitcher) pitching with first base empty and two outs (either 2nd and 3rd or 2nd only) against the Phillies and Chase Utley were batting. I’d walk him to get to Ryan Howard. You wouldn’t? Ryan Howard sucks against lefties. I don’t think this is coincidence. He sucks against lefties. ARod is very even against lefties and righties (as is Teixeira) so I would assume for this particular situation, you’d be hard-pressed to find a pitcher who’d have much better success verses ARod than Tex.
Tim Lincicum (generic lefty strikeout pitcher)
what?
generic and lefty would probably be the last two words i would use to describe tim lincecum
I think that’s why he spelled the name wrong guys, perhaps he should have gone whic Generic Joe?
To me, this is where the misperception occurs.
[1] The defense is already in a bad spot. 2nd and 3rd, one out … they’re looking at 1.5 runs in a 1-0 game. In short, the game is potentially getting ready to get out of hand.
[2] The defense is trying to get a play where they can get out of the inning with no runs scored on one batted ball. Like I said, they’re already in a tough spot.
[3] The defense is looking at 2 situations: either being down 3+ to 0 or down 1 to 0 if the move pays off.
[4] The D is acknowledging that if they get down 3+ to 0 against the NYY, they aren;t going to win anyway. They are trying to keep it “within reach” (in their opinion).
Teixeira gets a hit, and it doesn;t matter either. They’re down 3-0, with a runner on 1st and Cano coming up.
IMHO, the IBB stuff gets blown out of proportion. The D often is IBB’ing a batter when “they’re screwed” regardless, and are looking for a miracle that might keep them in the game.
Given that the teams that IBB a batter are often trailing a close game (or a lead that is close to putting the game in “probable lose” territory), it would be interesting to look at the times that the IBB did actually lead to an inning-ending DP, and see how many times the trailing team on defense actually came back to take the lead.
The IBB in this situation is really an acknowledgement of “we have to stop them here without giving up any runs, or we’rte not likely to win.”
Sometimes some of the commentary acts as if the defense is in a good position, and they just throw the game away with an IBB. That’s rarely the case.
GIDP’s occur in 11% of the situations where they could occur. Obviously, if you have Tim Hudson facing Derek Jeter with David Ortiz on first base and Elvis Andrus and Chase Utley up the middle, you’d have a significantly better than 11% chance of turning it, but your average hitter makes out 67% of the time, meaning you’d have a 44% chance of retiring the next two hitters – a.k.a being four times as likely to get both outs than if you put the lead runner on. It takes a lot of things breaking towards the IBB to even it up.
Isn’t this why Moore brought this bit of fun up:
“Runners on 2nd, 3rd, 1 out: .707
Bases loaded, 1 out: .752″
Yeah. The IBB in a bad situation for the trailing team on defense is similar (IMO) to a trailing basketball team fouling the opponent and sending them to the free throw line.
The difference there is that one is trying to stop the clock in that situation, and it gives ones team a chance to chip away at a small lead (since the trailing team can shoot threes, but the leading team is limited to two foul shots). There’s no clock to be stopped in baseball. Both are desperation moves, but only one makes much sense.
By “similar” in this context, I am referring to their “odds of leading to victory”. Giving your opponent free foul shots, while you’re gunning for threes (and the D knows it), is a very low percentage strategy … as is walking someone to load the bases.
In this very specific situation, I am not concerned with overall IBB stats, only the odds of the defensive team getting out of the inning with zero runs scored, and secondly, how often teams that accomplish that eventually take the lead.
Either way both teams are screwed and we’re basically arguing whether it’s better to “lose by 3″ or “lose by 5″.
Did you actually READ the article? He addresses this point and proves it wrong.
Really?
What the article says is basically …
Runners on 2nd, 3rd, 1 out: 1.578 runs
Bases loaded, 1 out: 1.986 runs
The defense is screwed regardless (which is kinda my point).
The numbers are averages, so sometimes with the bases loaded you give up more than 2 runs (usually), sometimes less. I acknowledge that, and understand it very well.
What the D is hoping for, is one of the few times where a ground ball leads to a DP and a scoreless inning.
Now, what we need to know are the odds of a scoreless inning with [1] 2nd and 3rd, 1-out, versus [2] bases loaded one out.
The D is essentially putting itself into an “all or nothing” situation where they view a 2-run single being just as damaging as a grand slam in their view on “winning the game”. Losing by 2 or losing by 4 still counts the same in the standings.
I’m not saying walking a guy to load the bases is a “good” idea, I actually stated a preference to the opposite.
What I am saying is “both situations suck for the defense” and the IBB to load em up is a desperate attempt at a scoreless inning, to preserve a smaller deficit, in hopes of possibly taking the lead later.
IMO, the avreages no more “prove it worng” than Joe Maddon IBBing MCabrera twice (with a small lead) on back to bakc nights and having Boesch end the inning both nights.
It’s a high gamble play in a situation where there are no “good” options for a defense trying like heck to keep the lead at 1 or so runs.
I’m not a fan of looking at average stats and applying them as absolutes to all situations.
Let’s face it, when you’re trying to prevent the team from scoring a run, facing Teixiera with 2 in scoring position and 1 out sucks, as does facing Cano with the bases juiced and 1 out. Cano’s a 48% GB guy for his career. Like I said, both situations suck for the defense, and the odds are that the D will give up close to 2 runs regardless of the situation. Yeah, I think we’re making too big of a deal out of the IBB.
Not trying to beat a dead horse or be repetitive, but trying to get more information … information highly relevant to what the defense is trying to do.
The D, presumably, is taking the position of “we can’t give up any more runs, or this game is over” (Given that it’s the worst offensive team in the bigs, this may be a reasonable assumption). So, they’re looking at how they can possibly get out of the inning without any more runs scoring. They are not concerned with whether they give up 1.5 or 1.9 runs, because (likely) in their mind, any runs means they are likely going to lose (we can debate whether the manager should have that view or not).
What I would like to see is what situation gives you the greatest odds (even though they’ll still be small) of getting out of the inning with zero runs scored.
A) 2nd and 3rd 1-out
B) 1st, 2nd, 3rd, 1-out
There has to be stats for this sort of this, right? If someone can point me where to find them, I’ll gladly look it up.
I am not arguing that the IBB is a “good” choice, only which situation gives you the best chance of a scoreless inning, if preserving the 1-0 deficit is your only goal.
Circle did you not read the last couple paragraphs?
This is what you say:
“I am not arguing that the IBB is a “good” choice, only which situation gives you the best chance of a scoreless inning, if preserving the 1-0 deficit is your only goal.”
From the post:
“The odds of scoring at least 1 run in an inning don’t justify the managers’ decisions either.
Runners on 2nd, 3rd, 1 out: .707
Bases loaded, 1 out: .752″
So, you just added almost a 5% chance of scoring one run or more, presumably this is from giving up BB or HBPs, or maybe some odd ROE situation since now the runner at third has to go home.
So, lets do a little math, before the walk you had almost a 30% chance of getting out of the inning with ZERO runs scored, after the walk that chance stood at about 25%.
Sometimes it seems like you make some sense, but then the next time I wonder if you can actually read, and somehow skipped straight to typing bullshit on your keyboard.
Folks are looking at this from only one point of view, and merely looking at the average run expectance – what you need to look at is a histogram to understand the %’s on scoring 0 runs vs 1 run vs 2 runs vs….
If bases loaded 1 out yields a higher probability of 0 runs scored but is worse in terms of 4+ runs, in this situation that is an easy (and justifiable) tradeoff… what matters here is the spread and probabilities of outcomes not simply the average outcome. In other words if the main difference between the 2 scenarious is 3 or 4 or 5 run events…. who cares in this sitaution?
This isn’t a simulation you’re running 10,000 times – the situation and context does matter….
In other words look at the win expectancies of the M’s with different scenarios and there is diminishing returns once the score gets 3-0 and beyond. They were down 1-0 at the time with 4 innings left against CC Sabathia (and potentially Mo in the 9th). Sure the IBB may yield more 4-0 or 5-0 outcomes than the 2nd and 3rd but does it matter at that point?
In other words are the 1-0 and 2-0 score outcomes the same? (if it is better for bases loaded – it might be worth trading off an increase chance of 4-0 or 5-0)
Wally, I don’t remember if I just missed that part of the post altogether, or if it just didn’t register as I read it.
I try not to be the “think about what you’re going to say instead of absorbing the information” guy, but that’s not to say that it doesn’t happen at times.
In this particular situation, we’re making too big of a deal about the IBB in my opinion … which is pretty much my overall point.
The M’s are already losing, the odds are against them not giving up any runs, and the odds are against them for not giving up multiple runs. They’re up against a great team and good starter. They’re screwed. The difference between really screwed and totally screwed just doesn’t matter to me.
I think the problem is that people are looking at this like these are probabilities and their not, their average runs. So the average run expectancy goes up from around 1.5 to 2, all that means is that on average a team will score 2 runs there. That says nothing of the probability that something else happens.
Case in point the probability of a double play with guys on 2nd and 3rd with one out is minimal, there just aren’t that many ways that can happen. But with the bases loaded there’s an 11% chance that a double play is turned, so their odds of getting through without scoring go up dramatically with the bases loaded. The off-set occurs because I’d imagine teams score more runs when they do score with the bases loaded.
Dead this part was talking about probabilities:
“The odds of scoring at least 1 run in an inning don’t justify the managers’ decisions either.
Runners on 2nd, 3rd, 1 out: .707
Bases loaded, 1 out: .752″
I don’t know about most of the time, but many times the IBB makes no sense what so ever however you slice it. Whether you used expected runs, win probability or chances of 1 run or zero runs, or 2 or less runs, or what ever. The conditions for making it a good play to issue an IBB (especially if this would load the bases) are very specific. Sometime managers stumble into a correct move, but many of the times its just an awful choice.
I don’t think managers use the IBB haphazardly or just for fun. I’ve already explained WHY they would use it in certain situations, even though it often means that if things go badly it’ll be worse than the pre-IBB situation. But, then we’re just talkig about a team that lost by 3 instead of a team that lost by 4, and who really cares?
Now, if teams were IBB’ing the opponent while winning 4-2 and the IBB batter puts the go ahead run at first and the winning run at the plate, then we could start talking about stupid.
The manager is presented with a situation, losing by a close margin, where there’s 70% chance that they give up runs on a batted ball or a situation where there’s 11% chance that a ground ball gets them out of the inning with no runs scored. He chooses the latter because in his mind, if they give up runs, they’re likely going to lose anyway.
We also see managers walk a great hitter with a runner in scoring position, first base open, and 2-outs to pitch to a lesser hitter. The trail runner doesn;t really add much extra threat, and you get to pitch to a lesser hitter.
We act as if managers issuing a IBB are putting the go ahead run on base, or bringing the winning run to the plate or something. They’re not. They’re going “all in” to try and keep the game close, and neither situation (IBB or no IBB) really offers them good chances at escaping the inning with zero runs allowed.
It’s basically a damned if you do, damned if you don;t situation, where the IBB offers the slimmest glimmer of hope that you can get out of there and keep your offense within striking distance.
It makes for good discussion, but I really doubt the IBB decides games often … just maybe makes the deficit look a little worse on occassion.
A-Rod has had an off-year while Teixeira has been getting hotter the past few months. I definitely think it’s a bad idea to walk Tex to get to Cano, but given A-Rod’s low OBP and slow speed, couldn’t the IBB possibly be worth it?
The whole point of this article is, no, it’s not worth it. Putting an extra person on base, even if the person hitting behind Tex is someone not good (like Cervelli, Pena, or anyone else who isn’t a 1st ballot HoF in a down year), is a bad idea. It’s worth pitching to Tex, who you are still 60% likely to get out (to over simplify the thing) than to put an extra run on base and take your chance with someone who you have a 65% chance of getting out. The numbers were crunched for you instead of just saying “well maybe since A-Rod is having a low- OBP year an IBB to Tex is a good idea.”
great piece! one thing i was wondering–the article talks about the expected number of runs, but not the distribution, so it assumes a manager is completely risk neutral about the number of expected runs. let’s assume instead that the manager is risk-seeking (i.e., mgr would prefer a risky chance at giving up no runs versus safer chance at giving up some but fewer total runs). would it possibly make sense for a risk-seeking manager to intentionally walk teixeira? or more generally in other lineups without such great hitters up and down?
As someone who was at the game, it had been raining all day, and big storms were rolling in. The fifth inning may seem “early”, but there was a decent chance the game was going to be called after the top of the 6th (and in fact there was an hour rain delay at that point). So the IBB may have been given with the idea that the Mariners might only have 1 inning to make up a deficit.
It is worth pointing out that Robinson Cano has the worst career RISP / empty hitting splits of any player of the last thirty years, and that such differences are absolutely real (per my presentation at SABR this year). His career OBP with the bases full is still just .290.
Be that as it may, the 1-out IBB is always wrong. Period. The Tigers recently walked Kevin Youkilis in that situation in the 9th inning and Ortiz cleared the bases with a walk-off 2B. Youkilis has the *best* RISP splits of the last thirty years, and walking him was still wrong.
“It is worth pointing out that Robinson Cano has the worst career RISP / empty hitting splits of any player of the last thirty years, and that such differences are absolutely real”
Of course, the most recent information matters more than past information, especially when a hitter is still young and improvement is possible. It’s more likely than not that Cano has gotten past whatever hurdle he had with RISP and now has no trouble in those situations. Walking the bases loaded for Cano may have made sense in years past, but if you look at his numbers this year, it’s one of the worst moves you can make.
Using some retrosheet PBP I have handy [2002-2004], I looked every time, 7th inning onward, 1 out, runner on second with first open, game within 2 runs.
If the batter was IBB’ed [431 times] the team scored 0.95 runs on average. When the batter was not IBB’ed [2607 times], the team scored 0.75 runs.
When the batter was IBB’ed
When he was not IBB’ed:
RS %nIBB %IBB
0 58% 56%
1 23% 19%
2 11% 11%
3 04% 06%
4+ 4% 08%
So looks like the IBB means a slightly smaller chance of getting out of the inning allowing 0 or 1 more runs while increasing the odds of a big inning.
I guess I am voting no IBB except perhaps with a good hitter batting, a pitcher on deck and the opposition has no bench players left…
With only a runner on second, the IBB can make a little more sense, because another BB or HBP doesn’t cost you a run. Its also now a little harder for a run to score on an ROE or fielder’s choice kind of play where something rather strange happens on a ground ball with the runner from third now forced to go home (with one out, you often see the runner from 3rd making sure the ball leaves the infield before breaking for home when the bases are anything but loaded).
Can we not fix this?
Hey self, they fixed it! Kudos FanFolks for getting the annoying ads pulled down.