The Matt Williams Effect

[Ray D]

A bit more on Monday's decision to not bunt Frandsen to 3d in the 14th and to send up Lobaton to hit (or "hit," depending on your view of Lobaton).  The hitter due up 3d in the inning was Span.  Span is not a deep fly hitter, generally (OK, leadoff homer yesterday , so let's say not a deep in-play fly hitter).  70% of his contact this year has been ground balls (~50%) or liners (~20%). Runner on 3d, 1 out, defense probably plays the outfield in to where they could throw out a tagging runner.  By bunting, you are diminishing the chance for Span to get a hit on a bloop or a liner in front of the OF, without getting much of a bump on him hitting a ball deep enough to score a run.  The infield also may play in to cut a run at the plate if there is a ground ball.  Infield in increases the likelihood of a ball going through. Pitcher I think would work Span high to make it tougher to hit a grounder or get down a successful squeeze.  I can see how someone smarter than me might not like the odds as much as having Lobaton and Span hit away.  There's also a slightly better chance of Rendon coming up with < 2 outs if you don't give one away bunting (wlaks and infield hits). That's the best way to drive up the odds of scoring the runner.

Do you really think that anyone on the staff actually thought this through as deeply as you have.

[Slateman]

No. Matt just made a terrible decision.

[HalfSmokes]

Do you really think that anyone on the staff actually thought this through as deeply as you have.

one would hope- football coaches have 'go for two' cheat sheets, this seems like the kind of thing an intern could assemble in a database- you'd need score, inning and current lineup (pick your stat) and possibly opposing pitcher plugged in, but that doesn't seem like too much

[Ray D]

one would hope- football coaches have 'go for two' cheat sheets, this seems like the kind of thing an intern could assemble in a database- you'd need score, inning and current lineup (pick your stat) and possibly opposing pitcher plugged in, but that doesn't seem like too much

I would think it would be very hard to factor in something like "with the runner on third the outfield will play shallower; Span doesn't hit deep balls often, if you force the outfield to play shallower you reduce Span's chance of getting a hit"  into an automated algorithm.

[HalfSmokes]

I would think it would be very hard to factor in something like "with the runner on third the outfield will play shallower; Span doesn't hit deep balls often, if you force the outfield to play shallower you reduce Span's chance of getting a hit"  into an automated algorithm.

Not really, you could use fly ball rate or slugging of the guys hitting after the bunter. Odds are someone at BP already has a formula, and I like to think teams are a step ahead of the interwebz in terms of resources

[spidernat]

JCA, you're wasting your time. You might as well be talking to a brick wall.

[JCA-CrystalCity]

Do you really think that anyone on the staff actually thought this through as deeply as you have.

No. I think Matt Williams has been around baseball long enough and works at it enough that this comes as instinctive to him and his staff as driving a car deciding whether to cross a street comes to me.  I don't discuss speed of approaching cars, light and weather conditions, presence of absence of police or cameras , evaluation of whether there are folks i don't want to hang around with nearby, etc... when I make that decision.  For Williams, he probably has a good sense of all this stuff without spelling it out.

[JCA-CrystalCity]

No. Matt just made a terrible decision.

no you did

[JCA-CrystalCity]

JCA, you're wasting your time. You might as well be talking to a brick wall.

Hey, MorsetheHorse asked, I answered.  As long as it is civil, I'm happy to explain my view.  I learn a lot from talking with folks who disagree with me. For example, thinking things through with Span coming up 3d, if it had been Rendon coming up after a better bunter, it might be a closer call.  I like him coming up with more holes in the defense and when a deep fly is more likely.

[aspenbubba]

Do you really think that anyone on the staff actually thought this through as deeply as you have.

This is baseball and it really hasn't changed that much over the past 50 years even with sabremetrics being introduced. I am sure the odds have been calculated and ingrained in every managers head along with the bench coach, defensive metrics guy, pitching coach all chiming in.

[Ray D]

This is baseball and it really hasn't changed that much over the past 50 years even with sabremetrics being introduced. I am sure the odds have been calculated and ingrained in every managers head along with the bench coach, defensive metrics guy, pitching coach all chiming in.

That makes it sound like there is a straightforward  formula which would produce a deterministic answer for a given situation.  However, while Matt Williams made the decision he made, a different manager, with the exact same inputs, might have made a different decision.  So I don't think it's quite that straightforward,

[Slateman]

no you did

Indeed. I stopped drinking at 10 am this morning

[JCA-CrystalCity]

Ray - I'm thinking it was more or less a conventional decision. It is as simple as "no need to bunt him over if he's already in scoring position."  That is conventional wisdom.  You can add complexity to it, but you get to the same answer if you think about it.  What has to be explained is the decision to do something different.

[mitlen]

I would think it would be very hard to factor in something like "with the runner on third the outfield will play shallower; Span doesn't hit deep balls often, if you force the outfield to play shallower you reduce Span's chance of getting a hit"  into an automated algorithm.

Werth could do it.

[Ray D]

Werth could do it.

Good one!

[Slateman]

Another poor managing decision.

Treinen walks, Span walks. Two on, nobody out, and a slumping Rendon is up. Obvious solution is bunt Treinen over. You simply have to bunt it to third and beat the short stop. No bunt, Rendon grounds out with the Pirates getting the out at third. Next hitter, Werth, hits a deep fly ball that Treinen would have scored on.


STOP freakING PLAYING FOR A BIG INNING AND START SCRAPPING FOR  A RUN.

[MorseTheHorse]

There's actually stat work on this, and you can chase it down if you like.

I'm well aware of the work.  With zero outs, odds of scoring a run go up bunting a guy from 2nd to 3rd.  They go down bunting a guy from 1st to 2nd.  That's why I find the two decisions in tandem so baffling...

[MorseTheHorse]

I'm well aware of the work.  With zero outs, odds of scoring a run go up bunting a guy from 2nd to 3rd.  They go down bunting a guy from 1st to 2nd.  That's why I find the two decisions in tandem so baffling...

But it's a good point about Span being the guy coming up with a guy on 3rd and 1 out.  That's reasonable. 

[Squab]

I'm well aware of the work.  With zero outs, odds of scoring a run go up bunting a guy from 2nd to 3rd.  They go down bunting a guy from 1st to 2nd.  That's why I find the two decisions in tandem so baffling...

I think there's something like a 4% increase in chance of a run scoring with the guy on third and one out. If there's a 20% chance the bunt doesn't work that advantage goes away. With the random nature of the game that small advantage would probably be hard to see over even a season.

[tomterp]

I think there's something like a 4% increase in chance of a run scoring with the guy on third and one out. If there's a 20% chance the bunt doesn't work that advantage goes away. With the random nature of the game that small advantage would probably be hard to see over even a season.

Right.  While it's helpful and informative to understand the advantage of being on one situation (say man on 2nd no outs) versus another (say man on 3rd 1 out), what isn't visible is the risk of failure on the bunt attempt.  Assuming 100% success on a sacrifice is to overlook a significant occurance of complete failure, like a pop up for example.

Also, the skills of the batter in question also should be considered by a manager, who should be informed by the average probabilities but not make decisions directly based on them.

[HalfSmokes]

Right.  While it's helpful and informative to understand the advantage of being on one situation (say man on 2nd no outs) versus another (say man on 3rd 1 out), what isn't visible is the risk of failure on the bunt attempt.  Assuming 100% success on a sacrifice is to overlook a significant occurance of complete failure, like a pop up for example.

Also, the skills of the batter in question also should be considered by a manager, who should be informed by the average probabilities but not make decisions directly based on them.

the skill of the bunter, the skill of the next three batters (the out you're trading results in the pitcher hitting if the batter gets a hit instead of sacrificing themselves vs. the number 4 hitter makes a huge difference), the speed of your base runners... there are a crap ton of variables at play- it seems like something that a database would be good at, but short of that, you just have to hope the manager gets it right

[Slateman]

You guys are over thinking it.

What he's doing isn't working. Not bunting, not playing small ball . . . it's failing. This lineup cannot wait for a big inning because they're not getting them consistently enough. Matt Williams has to adapt to this and play some small ball for now. If you keep doing the same thing, you can't expect a different result.

[tomterp]

the skill of the bunter, the skill of the next three batters (the out you're trading results in the pitcher hitting if the batter gets a hit instead of sacrificing themselves vs. the number 4 hitter makes a huge difference), the speed of your base runners... there are a crap ton of variables at play- it seems like something that a database would be good at, but short of that, you just have to hope the manager gets it right

Sure, all true.  It's really helpful for a manager to know that in an average situation, here are the probabilities.  Now, assess the skills of the players (without bias if possible, but this is a likely source of failure) and adjust accordingly.

[sportsfan882]

what a crapshow. close up shop, boys.

William Ladson ‏@washingnats  8m
#Nats lineup for Friday vs. #Pirates: Span-CF, Rendon-2B, Werth-RF, Ramos-C, Desmond-SS, Dobbs-1B, Walters-3B, McLouth-LF, Zimmermann-P #MLB

freaking a

[skippy1999]

what a crapshow. close up shop, boys.

freaking a

alright I've been out of the loop a bit, who is Dobbs and where did he come from?  Wasn't Moore doing ok at 1st?  and why oh why is McLouth still on this team, is there truly no one better at AAA that could at least have a chance of getting a hit??