Say you have 2 pitchers who had the same RA9 in a season, but pitcher A had a 1.00 ERA in shutdown innings and pitcher B had a 4.00 ERA. Which pitcher likely had the best season in terms of wins and which pitcher likely benefited his team the most?

Psychologicially benefited his team? Pitcher A. Whether that translates into actual benefit is unknown. I suppose TFA's implication is that "unknown" = "either nonexistent or irrelevant".

Both have the same RA9, so I'm assuming the actual benefit of both pitchers is equivalent, all other things being equal. The inequality of run distribution, one pitcher giving up seldom after his team has scored but regularly when his team hasn't, the other giving up regularly after his team has scored but seldom when they haven't, is a potential difference maker. I don't know if it produces an actual difference, but I know in a 5-5 tie there's a psychological difference between having had a 5-0 lead and lost it vs. having been down 5-0 and coming back to tie vs. a constant back-and-forth battle. Whether that manifests in the results, I don't know.

I'm sure mathematically it doesn't matter but as a fan it sure as hell feels like being able to keep your opponent off the board when you've just scored is important. The Red Sox this year seemed horrible at this. It seemed like money in the bank that if the Sox scored twice in the top of the fifth they were giving up runs in the bottom of the inning. Of course the Sox would often give up runs even when they didn't score them so that's not the best example perhaps.

Basically, the half inning following a half inning in which your team scores at least 1 run, you are supposed to “shut down” your opponent and not allow any runs.

So...if your team doesn't score any runs, then during the next half-inning it's OK to allow your opponent to score runs??? Would this be termed a "shut up" inning???

Both have the same RA9, so I'm assuming the actual benefit of both pitchers is equivalent

Haven't RTFA, but I believe pitcher B is more valuable. Preventing runs has more value in a lower run environment. Absent all other information, the fact that your team just scored implies a higher run environment than if they didn't score.

I may be missing something, but it seems obvious to me that MGL is right: when you are in a silly "shutdown inning," you know your offense has just scored. So it's more likely that your team will be able to sustain run(s) allowed by you during the shutdown inning.

I may be missing something, but it seems obvious to me that MGL is right: when you are in a silly "shutdown inning," you know your offense has just scored. So it's more likely that your team will be able to sustain run(s) allowed by you during the shutdown inning.

Just ask Jack Morris!

I may be missing something, but it seems obvious to me that MGL is right: when you are in a silly "shutdown inning," you know your offense has just scored. So it's more likely that your team will be able to sustain run(s) allowed by you during the shutdown inning.

Exactly, especially b/c it can be multiple runs.

If your team puts up a four-spot, and you give one or two right back, you're still in great shape, ceteris paribus.

Haven't RTFA, but I believe pitcher B is more valuable. Preventing runs has more value in a lower run environment. Absent all other information, the fact that your team just scored implies a higher run environment than if they didn't score.

Oh, come on. What's the difference going to be, like less than a percent?

Oh, come on. What's the difference going to be, like less than a percent?

Very small, I'm sure, within the actual variation seen. But that's not the point.

The stat in question is saying the value goes the other direction.

Ray got this right. MGL's point is that if you have to give up runs, the best time to do it is when you know that your team has already scored. Basically Pitcher B is Jack Morris, he pitches to the score. I don't know where you're going with this "run environment" thing, unless that's an awkward way of referring to the numbers on the scoreboard.

It's not even a matter of "knowing," is it? I mean, the best time to give up runs is if you also score some. You lose all the games in which you give up runs while being shut out, etc.

So it's more likely that your team will be able to sustain run(s) allowed by you during the shutdown inning.

They're also, at that moment, more likely to sustain runs allowed by you in subsequent innings. There's nothing special about that particular half-inning, is there? If you give up N more runs in the game, does it matter if some happen in that particular half-inning?

I don't know where you're going with this "run environment" thing, unless that's an awkward way of referring to the numbers on the scoreboard.

The fact that X runs being scored in an inning means the expected runs for the game is higher than if zero were scored.

If your offense averages 0.5 R/inning, or 4.5/G, the fact that you scored in one inning (vs. taking a zero) means your expected scoring is at least 5 R/G, vs. 4 R/G.

If your team puts up a four-spot, and you give one or two right back, you're still in great shape, ceteris paribus.

But if your team scores once, and you then give one or two back (a more common occurrence), you are in worse shape.

And when your team scores four and the other team comes back with two, even though you are still in relatively good shape for winning the game, the dynamics of the game have changed - you are less likely to be able to get out of this game without using your top relievers, for example. And that can affect your ability to win games down the road.

If you have a choice of when to give up runs, of course, you would prefer to give them up when you are already behind, because those runs have a lower impact on your ability to win the game.

-- MWE

Ok snapper. I thought you were making a different point, one that was more convoluted and stupid.

yeah yeah, but show some respect to John SMoltz.

Both Smoltz and Jack Morris are hall of famers, so I don't think MGL or anyone here has any credentials to argue with them.

That means they are clearly right, even though their philosophies are polar opposites. A good pitcher needs to either shut the opposition down after their team scores, or pitch to the score and give the opposition something, but not the lead.

Both the Smoltz or Morris approach works, it's mediocre pitchers who try to split the difference and can't commit to either of the Smoltz/Morris approaches who fail, and are often over-rated in our "saberistic" modern age. Like Felix Hernandez in Seattle last year, who was inexplicably given the Cy Young by a bunch of math obsessed basement dwelling bloggers despite his inability to either throw shut down innings or pitch to the score.

Now that we've explained that to all you viewers out there, it's time for a commercial break.

Ah, cool, cold, Budweiser, nothing makes for a better tasting breakfast, Budweiser with eggs, Budweiser with bacon, even Budweiser on your cereal. Pound a cold bud this morning! And every morning, it's the choice of ex-jocks everywhere. Don't you want to grow up and be like us, kids?

I'm pretty sure John Smoltz has forgotten more about baseball than MGL will ever know.

I think the tone of the piece could be a little more respectful.

I'm pretty sure John Smoltz has forgotten more about baseball than MGL will ever know

Based on this example, it does indeed appear he has forgotten a great deal. (Unless he never knew much to begin with, which I suppose is also possible.)

To turn a Smoltz bash into (perhaps) an interesting question:

Is there any correlation inning-to-inning between runs scored between teams? There should clearly be a correlation for a given teams scoring in inning-X and inning-X+1 because the opposing pitcher is "similar" (correlated). But if Team A scores in inning X - what is the probability that team B scores in bottom X (or top of X+1)?

I'm pretty sure John Smoltz has forgotten more about baseball than MGL will ever know.

I think the tone of the piece could be a little more respectful

I really loved John Smoltz the player, and I think he makes some valuable contributions in the booth at times. But for all his talk of how he evolved into a "pitcher" instead of a "thrower" I never saw it. Even in the twilight of his career he would just throw his breaking balls at the middle of the plate and let them break either to the bottom of the zone or off of it. The man had one of the best sliders in the history of the game, and later on he picked up a very good splitter as well, and he relied on those and 96 mph heat through the majority of his career. All this talk about knowledge and disecting the opponents should really be left to Glavine.

All this to say, I don't doubt that John Smoltz has very little to offer when it comes to explaining the side of pitching that is more game theory oriented than mechanical.

This says "Pitchers who give up fewer runs than their team scored in the previous frame" are valuable. IOW, this is "Pitch to the score" in different clothing.

I think the tone of the piece could be a little more respectful.

I surmise you are not familiar with the writing style of mgl when he is discussing the Great and the Good.

Based on this example, it does indeed appear he has forgotten a great deal. (Unless he never knew much to begin with, which I suppose is also possible.)

I know that you are mgl's best friend around here, but even you have to admit that mgl's obsession with 'just the stats' leaves unaddressed a problem here with 'what do players think'. It may be that players feel better about their chances of winning the game if the opponents are actually shut down during the shut-down inning. And it may be that optimistic players win more games than pessimistic ones. Do we have research for that? I have no idea, but the issue of 'intangibles' in the sense of how they affect player morale and how player morale translates into wins may be fertile terrain for a new, touchy-feely sabermetrics.

I should probably look on J-Stor and such for studies of sports psychology myself. But I won't be able to do so today.

EDIT: I ought to make it plain that I have not read the article, and have no intention of clicking through to it. I'd honestly rather read Chass nowadays.

This says "Pitchers who give up fewer runs than their team scored in the previous frame" are valuable. IOW, this is "Pitch to the score" in different clothing.

If pitchers actually could pitch to score, it would be very valuable. It's just that they can't.

the problem is nobody actually knows how players think. Who the #### is anyone to say that a shutdown is a morale booster?

edit: except for bob tufts

even you have to admit that mgl's obsession with 'just the stats' leaves unaddressed a problem here with 'what do players think'. It may be that players feel better about their chances of winning the game if the opponents are actually shut down during the shut-down inning. And it may be that optimistic players win more games than pessimistic ones. Do we have research for that?

I don't think it leaves players' feelings "unaddressed," I think it assumes their feelings rarely matter. Research has largely been unable to find evidence that supports the "hot hand" idea, and you would think a recent run of personal success should increase a player's "optimism." And I believe people have researched the similar idea that blown leads late in a game hurt a team the next day, and found no relationship. In general, any theory other than "professional baseball players will perform at their usual talent level" has failed when put to the test. But by all means, review the research and see what you find.

To me, Smoltz's premise is absurd. He wants us to believe that hitters do the reverse of pitching to the score -- i.e. they will hit better when it is least valuable (when they have already scored, and the other team was then shut down). I would expect that a study of this theory -- if anyone wants to waste their time doing it -- will turn up no evidence to support it. But I could be wrong.

He wants us to believe that hitters do the reverse of pitching to the score -- i.e. they will hit better when it is least valuable (when they have already scored, and the other team was then shut down). I would expect that a study of this theory -- if anyone wants to waste their time doing it -- will turn up no evidence to support it. But I could be wrong.

There may be a correlation; i.e. a pitcher who has already allowed runs is more likely to give up more runs, and as a team gets further behind, they'll use worse RP. So, scoring runs may correlate with more future scoring, but it has nothing to do with the hitters. It's 100% the quality of pitching faced.

Everyone who is saying they haven't RTFA: Yes you have. The entire article is pasted above.

I'm not clear on this: Is there research on "shut-down innings"? Is there a correlation between high "shut-down innings percentage," or whatever it is, and wins?

"Shut-down innings," it seems to me, is one of those stats the TV networks give in a vacuum. They'll tell you that a guy's shut-down innings percentage was 63 or something, but they won't tell you what league average is.

I think it assumes their feelings rarely matter. Research has largely been unable to find evidence that supports the "hot hand" idea, and you would think a recent run of personal success should increase a player's "optimism."

I remember reading a discussion of Todd Zeile in this context in Curve Ball, but this isn't exactly what I'm talking about.

What I haven't seen is a study of actual players' psychology before and after certain events, or any sort of discussion with actual players about whether they believe what Smoltz is saying, and how that might relate to actual wins.

But there's another problem here, which is that I think mgl is taking people like Smoltz far more seriously than they would take themselves. Smoltz is paid to offer a player's insight into the game. So if Smoltz is to be regarded as saying 'we players really like a shut-down inning after we score a couple of runs' (we don't have an exact quote of what set mgl off), it kind of makes mgl's deflation of this comment churlish and petty.

Joe Morgan used to emphasize the importance of shut-down innings when he did ESPN games too. The concept has been around a long time and players have really believed in its importance.

Is there any correlation inning-to-inning between runs scored between teams?

There's bound to be at least some minimal correlation due to weather, umpires, park, etc. Wind blowing out at Wrigley will lead to lots of innings where both teams score; wind blowing in at Wrigley will lead to lots of innings where neither team scores. Same with umpires' big/little zones, etc.

As to player psychology -- piffle. Sure, you get a boost when you go up 1-0 and your pitcher holds the lead and a letdown when he gives up a run or two. Of course you get a boost when you didn't score and your pitcher holds the tie and a letdown when he gives up the lead. If there's a pscyhological effect here it's "hey, I can't really put my finger on why, but I feel more confident we can win when our pitcher doesn't give up runs. It's probably the copper bracelet I wear, those things are awesome."

Joe Morgan used to emphasize the importance of shut-down innings when he did ESPN games too. The concept has been around a long time and players have really believed in its importance.

Typical blackjack players really believe they can win based on many things that have zero correlation to winning at blackjack.

We descended from the trees and ascended from being mere monkeys to a great extent because of our ability to discern patterns and draw conclusions from them. We are, in effect, pattern matching monkeys, and this also means when straining for answers to difficult problems, we often discern correlations in patterns that are nonsensical.

Anyone who played baseball full time as a career would naturally strain to find reason in their results, correlations that could explain to how they could just hit or pitch a little bit better. Doesn't make any of those perceived correlations real or accurate.

I should probably look on J-Stor and such for studies of sports psychology myself. But I won't be able to do so today.

I've watched enough soccer matches to know that there are teams which often look exceedingly vulnerable to conceding after scoring one themselves, particularly if the team which has just scored is the underdog because they then tend to become more defensive while the opposition steps up a gear.

Is there a correlation between high "shut-down innings percentage," or whatever it is, and wins?

I just gave a quick glance to the entire Retrosheet game log database, and the ERA of pitchers in shutdown innings is 0.00 while the ERA of pitchers in non-shutdown innings is > 0.00.

As such, I conclude they are extremely important to winning games.

It's probably the copper bracelet I wear, those things are awesome

Johnny Oates once told me in complete seriousness that Bucky Dent swore by copper bracelets and attributed all his success in life to them. Well, in seriousness that Dent thought so. I believe Oates thought Dent was nuts.

I don't think it leaves players' feelings "unaddressed," I think it assumes their feelings rarely matter.

Well, then, you're a bad scientist.

Any time players' feelings have been accounted for, it's been through a proxy variable. If the proxy bears a very weak relationship to the desired variable it's only going to pick up at best a very weak signal, and likely a statistically insignificant one at that.

Let's say you want to test the claim that women are more likely to give birth to a baby than men are. Your data has whether someone has given birth, but you don't have gender. "Well," you say, "that's not a problem, because I have their height, and women are generally shorter than men. If I look at short people, they should be more likely to be women, and thus more likely to give birth!" Then when you look at your data you'll find short people are less likely than tall people to give birth*, and you'll conclude there's nothing to the notion that women are more likely than men to give birth. In reality, you just picked a bad proxy, or made a bad assumption. Bad science is worse than no science.

With players' feelings, we have nothing but weak proxies. If you can't sufficiently test something, should you assume anything based on insufficient tests? Or should it be considered unaddressed? Definitely the latter. So, if you want to say there's no measureable evidence for Smoltz's suggestion, that's fine; but if the problem is that we can't accurately measure, we can't conclude it's true or untrue. It's simply unaddressed.

* Are short people less likely to give birth than tall people? Pretty much all pre-pubescent kids are shorter than most women of child-bearing age, and the birth rate among pre-pubescent kids stands solidly at 0%. Likewise, post-menopausal women are generally shorter than pre-menopausal adult women, further skewing the relationship between height and birth rate.

For some reason I think of economic statistics. Feelings definitely drive people's behavior. You buy things, you save, you invest, you start a business, you squander your nest egg, whatever, and emotions, many of them irrational, drive what you do. On aggregate, this doesn't matter, and to some extent, even in looking at individual outcomes (assuming we're looking at the broad middle-class of Americans) it doesn't matter a whole lot, either. Scores of millions of Americans live in relatively similar houses or apartments, drive cars, own appliances, have phones and computers, are better or worse off within some parameters that don't vary too much, cosmically speaking. Sure, some people become dumpster divers, live in ashrams, spin their own wool, become homeless drug addicts, build huge houses on the Gulf Coast, and other outliers. But how you feel, which can take huge swings in a lifetime, manifests itself economically within a range of likely possibilities.

Sports are much, much more limited than that. The best or worst things that can happen to human beings may befall ballplayers during a season – let alone giving up a run in the top of the fifth. And how can their professional reaction manifest itself? They swing at the curveball and either hit or miss it. Baseball statistics just don't allow for a very nuanced expression of emotions, or a broad palette of emotional outcomes.

With players' feelings, we have nothing but weak proxies. If you can't sufficiently test something, should you assume anything based on insufficient tests? Or should it be considered unaddressed? Definitely the latter. So, if you want to say there's no measureable evidence for Smoltz's suggestion, that's fine; but if the problem is that we can't accurately measure, we can't conclude it's true or untrue. It's simply unaddressed.

What a silly thing to say. The shut-down theory could easily be evaluated. We don't need any "proxies." Just see if teams score any more after their pitcher records a shut down inning than we'd otherwise expect them to score(based on their hitters, opposing pitchers, park, time of day). I don't think you will find anything, but if I'm wrong I will happily admit error and acknowledge your discovery.

Players' state of mind clearly can matter. We know that because home field advantage exists. But theories that predict short-term changes in player performance based on their effort or psychological factors generally do not fare well when subjected to objective analysis. I'd be very surprised if this is an exception.