THE BOOK--Playing The Percentages In Baseball

Long time reader guys.  Keep up the good work!

Gameday had the pitches around 5 mph faster than Fox according to the release point.  Fox’s numbers looked like the average of release point velocity and home plate velocity, for all at bats I watched.

I was confused by McCarver on the count for the Pujols HR - should we be surpised?

I am against IBBs, especially early in a game.  I saw a quote attributed to Bob Gibson, I believe, along the lines of, “If I have to walk the #8 hitter to get to the pitcher, I shouldn’t be in this game.”

If you walk Pujols (or anyone else) in the third inning, you probably shouldn’t be in the game.  Unless the batter has a .501 OBP or 1.001 OPS in a given the situation, you always have “50/50” chance or better to retire him.  If you walk him, you have a 0% chance of retiring him and increase the run expectancy.

On the other hand, Leyland said in the in-game interview his intention was to not pitch to Pujols.  Then you have to give the IBB because you are trying to avoid the mistake Verlander made. 

A batter of Pujols caliber usually knows when he is being pitched around and will not swing at a bad pitch in that case - avoid the mistake and just IBB him regardless of the chance to retire him.

Pujols jumped on the first pitch.  I think McCarver was told by Leyland that they would approach Pujols as if he always had an 0-2 count on him.  And then he repeated that and confused the heck out of everybody.

Re: Reyes, I’m not sure I get where you’re coming from, MGL.  He only ended up making 93 pitches, and he’ll get a full four days rest to pitch in game five, so I’m not sure how much you lose from him by letting him go.

I would think that the fewer pitches, the better.  Unless more pitches is better (at that low pitch count) or at least has no effect, what is the point of letting him pitch the 8th and/or 9th?  It does completely save the pen I guess, so if in fact 93 is no worse than 83, then fine.  I just would have thought that in the WS, if you can rest a starter after 70 or 80 (or whatever) pitches, that is an advantage.

Dufman, I think you are right that in that situation it is not the correct thing to do (walk Pujols), but in general, you need to “know” each batter’s projected something (OPS, lwts, etc.) and then use some kind of formula or mathematcial analysis to figure out what the correct thing to do is (which WE is higher - walk or no walk).  I DON’T think you can figure it out with WORDS as you are trying to do.  That is what managers and commentators and fans do and it often leads them to the incorrect decision.  I encourage you to read Andy’s chapter on the IBB in The Book.  He also has shown that pitching around someone does not really accomplish anything, as you astutely point out.  The reason is probably that batters know when you are pitching around them and they tend to lay off the bad pitches.  I don’t think that it is necessarily any worse than pitching normally to them, so I would be careful about saying that the IBB is better than pitching around someone.  If the IBB is wrong, then you probably want to either pitch normally to the batter or indeed pitch around him, which as I said, is probably around the same thing as pitching normally.

About the pitch speed, the last time Fox broadcast a Verlander game, he was throwing in the mid to upper 90’s according to their graphic.  Last night, he was in the low 90’s at best I think.  It appears that he was not nearly at the top of his game, speed-wise.  I don’t think his command is good enough at this point in his career that he can be very successful without a mid to upper 90’s fastball, even though he has an excellent curve and a good changeup.

That was my take on it as well.  McCarver was making the point (repeatedly) that Verlander should have been throwing “0-2 pitches” to Pujols no matter what the count was.  I also agree that managers should be much more willing to use their worst pitchers in playoff games with big leads.  Tigers/Yankees Game 3 comes to mind for sure .. in a 6-0 game, what the hell was Zumaya doing in there?

There may not have been any other key strategic decisions in Game 1, but there have certainly been some questionable strategic decisions made by Leyland before the game started.  Leaving Chris Shelton off the roster was a pretty questionable move to begin with, but doing so when Casey is not able to play first-base borders on inexcusable.  I really can’t understand how someone can prefer Guillen 1B/Santiago SS to Shelton 1B/Guillen SS.  How good on defense would Santiago have to be to justify the loss of offense?

And assuming that Robertson does indeed start Game 3, that looks like a mistake as well.  Speaking as a Tigers fan, Bonderman is the one pitcher I’d feel most comfortable with in an important game.  (I’m too lazy to run numbers, but I’m guessing most projection systems would agree with this.) It’s just a shame that we get a week of rest and he still ends up being the guy who only starts once.

MGL:  what do you think of the decision to start So in game 1, as opposed to having J-Rod or Spiezio in the lineup against a RHP?  Is his defense in RF good enough to justify that decision?

Hypothetically, what would happen if LaRussa filed a protest after the first inning concerning Rogers’ pine tar on his left hand?

The law degree toting LaRussa might have a valid cause of action. I think a protest might be especially valid since Rogers’ worst part of the game is his first 15 pitches (7.48 ERA). Perhaps Juan Encarnacion hits a three-run homer instead of grounding out.

Here’s to hoping we see another Game 2 tomorrow night!

So cannot hit a lick.  Versus a RHP, he is less than worthless.  No defense that he can provide can make up for that.  J-Rod or whomever is a much better alternative.  Tony has dissed J-Rod for a while now, for some reason I am not aware of.

When it comes to young or “unproven” players (or non-regulars), Tony thinks he can somehow mix and match through intuition or some such thing, such as who is hot or cold or who is 3 for 7 or 0 for 5 against whom.  Of course, all of that is nonsense.  That is his biggest weakness as a manager.

Robertson is clearly the worst pitcher on the team despite having an excellent year (perfect example of a “banner year does not a player make") and should be the last starter if the schedule allows it.  I am not sure, without breaking out my sim, however, how much different the Cards offense is versus LHP and RHP.

I don’t understand why, as in tonight’s game (game 2), when in the ninth inning, and there are two outs, they allowed Scott Rolen to take second on fielder indifference. Why not try throwing him out? Basically, letting him take second is saying that you don’t care about his run (since the team’s up by three). So who cares if you make an error when you try to throw him out? Worst case, the run scores (which you’re basically conceding anyway), best case, game’s over. I don’t understand.

Well, the first completely moronic decision of the series (I’m sure there will be more).  Letting Molina, the worst regular on the team, bat in the 9th, rather than Duncan or J-Rod on the bench.  Completely indefensible.  I’m sure the “reason” was that Molina was the hero in game 7 of the NLCS.  That reason is as moronic as the decision to let him bat…

How can LaRussa let Wilson and Molina hit in the 9th, when a single ties the game and he has Duncan and J-Rod on the bench?  2006 OBP vs. RH:
Duncan .390
JRod .366
Wilson .288
Molina .264

Unbelievable…

PLEASE don’t EVER cite one-year stats to defend or support a sabermetric argument!  If you do, at least regress them the appropriate amount first.  Even then, why would you want to do that when you can (and should) use career stats.  Especially for something lile platoon splits, in which the underlying true talent probably doesn’t change much over the course of a player’s career (so you don’t need to worry much about weighting for recency).

Not only that, but your sample for one year is SO small when you are dealing with platoon splits (becuase no one faces a lefty pitcher much in one year).  To top everything off, it is heresy (in my religion) to cite how someone does versus a particular side pitcher and ignore how he does versus the other side pitcher.  One very much informs the other.  If you are going to talk about platoon splits (or any other splits which not nearly independent), please make sure you cite a player’s TOTAL stats and THEN APPLY A PLATOON RATIO or something like that.

If Player A hits .340 versus LHP and player B hits .300, but player A also hits .250 versus RHP and player B hits .350 versus RHP, who do you think is likely the better hitter versus LHP???

So I appreciate the fact that you agree with my sentiment, but PLEASE don’t site garbage stats on this site.  Someone impressionable might be reading! 

And if Molina is up against a RHP, we don’t have to cite ANY statistics to know that a LHB would be better, even with the pinch hitting penalty.  We have one of the worst hitters in baseball (truly) in Molina.  There is NO LHB in baseball who is going to be worse than Molina versus a RHP!

MGL:  It was a blog post, 5 minutes after the game ended (and posted before I saw yours), not an essay.  I thought it was interesting just to take a quick look at what they had done this year, because that was easily accessible to me.  I thought about adding “Of course, these stats do not reflect the 4 players’ true talent, but it’s still clear LaRussa made a mistake,” but I thought that was unnecessary for this board.

Sheesh.....

Andy - completely agree, I was thinking the same thing.  Especially given it was Pudge behind the plate, and that Rolen may not have been running his hardest, figuring they would let him go anyways.  Maybe Pudge trusts Todd Jones more than I do… but if I’m historically the best catcher arm in the game, I go for the CS to end the ballgame there.

It’s not a question of the catcher/pitcher, but of the middle infielders.  While I was initially thinking the same thing, I don’t think I saw where the two infielders were playing, and therefore, they’d have no reason to leave their position early to try to make a possible play.

MGL - my point about IBB is better than pitching around is that if you IBB someone, he will only get a walk (99.99% of time).  If you try to pitch around him so much more can go wrong - WP, PB, hits, errors, etc - instead of the intention of walking the batter.  This is only valid if you are deciding between IBB or pitch around as Leyland indicated during his in-game interview.  If your choice is between pitch to or pitch around, that is a different analysis and I would tend to agree that there is little difference other than the chance of a walk increasing.

In fact, I just re-read Tango’s excerpt on THT about pitching around and that may have been where I generated most of my thinking on this situation. 

To quote from the post:

“In terms of an alternative to intentional walks, pitching around a hitter accomplishes very little, as it merely increases the ratio of walks to non-walks without significantly affecting how well the hitter performs if he isn’t actually walked. In other words, if you’d rather have a walk than whatever the hitter normally does if he doesn’t walk, then you should walk him. If you’d rather have what he normally does, you should pitch to him normally.”

I would argue that you need to analyze a situation with words, or simple terms, to understand what statistical model to apply to the situation or in the least to create a hypothesis for study or research as Tango alluded to in the post from The Book.

My point in using .500/1.000 was to provide a simple guide that you only IBB the best hitters, especially early in a game, and it probably muddled the more relavent point of if you are going to walk Pujols, then just walk him.

I appreciate the comments and the “kick” to read The Book.

Generally when you let the runner go to second, it is that the first baseman is not holding the runner (and correctly so of course) and there is no chance to throw him out.  Of course, if there is any chance to throw the runner out, they should.  Also, there is the “courtesy” aspect of it.  I won’t try to throw you out and you don’t have to run hard.  It is like the courtesy double to left field.  You don’t have to throw to second and I don’t have to run hard or slide into second.

What about Edmonds taking second base?  If he stays on 1B, it’s 1st and 3rd, 2 outs, down by 2.  He decided to go for two.  If I remember right, he went in there standing up, but it was close enough that he should have slid, and close enough that if he overran even for a split second, he would have been tagged out.  But, not a single mention, and the replay didn’t even focus on it.

I don’t have my WE charts with me, but it was probably a great run for him to make (he’s the tying run), but I don’t understand how you don’t slide there.

I noticed that too and thought it odd that there was no mention.

So Taguchi steps to the playe with 1 out in the 6th, with the pitcher, Carpenter, on deck.  He says, “Will they walk Taguchi?” or something like that.  What is he talking about?  (No, he didn’t think there was 2 outs.) Has this guy (McCarver) ever played baseball?  How scary would it be if he were a manager?

Men on first and second, Pujols at the plate, 0 outs.  Grounder back to the pitcher.  Announcers claim the play should be to second for the 1-4-3 DP.  Is there any chance of the runner on second making it all the way home on this play if it goes like that?  If not, there’s man on third, 2 outs.

However, Zumaya chose to throw to third, in an attempt at a 1-5-3 DP, which would leave a man on second with 2 outs.  Assuming he didn’t throw it away, and Inge at 3B was ready to field the ball, tag the bag, and go to first…

What’s the true correct play here?  I’m guessing figuring it out involves some combination of base/out state data, as well as a bit of assumptions regarding the speed of the runners and the 3B’s arm…

There is really nothing to figure out Mike.  Man on 2nd 2 outs is a little better than man on 3rd 2 outs, but not much.  There is almost no chance of the runner from second scoring on the DP.  I have never seen nor heard of that.

There is no particular reason not to go 1-5-3 other than it is rarely done, the fielders are not usually prepared for it, the middle IF’ers are better at turning the back end of the DP and it is generally a shorter and quicker route.

In this case, he would have turned the DP.  He just made a bad throw.  Inge was “ready” for the throw despite what McCarver said about Inge not being ready.

Note to McCarver:  Preston Wilson does NOT bunt, especially with Pujols behind him!  This guy McCarver seems to be obsessed with the bunt and the IBB - at least talking about it.  Other note to McCarver:  If Inge is playing Pujols to pull, then he is CLOSER to the bag.  Man is McCarver bad.

It gets worse.  McCarver:  “I am surprised that Leyland didn’t just walk Pujols.” With a 1-2 count!

Game 3, bottom of the 4th, Edmonds on 2nd, Belliard on 3rd, LaRussa elects to walk Molina to set up the DP and a force at any base, and pitch to Taguchi--is this a good move? Molina’s such an anemic hitter, why not pitch to him with the infield in and then go after Taguchi, leaving Carpenter (potentially) to lead off the next inning?

One out in that situation, BTW.

A lineup move that had no effect on the game, but still thought it curious…

G3, B8, Cards up 4-0, Leyland brings in Rodney to pitch.  He double switches, taking out Inge and putting Neifi Perez in the game to bat in the ninth spot, which is due to lead off in the top of the ninth.

The only possible reasons for this double switch, that I can think of, are:
1) Leyland wants to put in Neifi for defense (unlikely down 4-0) and so may as well double-switch
2) Leyland wants to insure that Rodney will be available to pitch in the B of the 9th, if it gets there
3) Leyland actually wants Neifi Perez to lead off the top of the 9th

The most plausible is situation 2, but in that case, in order to score 4 runs in the top of the ninth, you’re probably going to want someone other then Neifi leading off.  And given that you score 4 in the top of the 9th, there’s not a horrible chance that you’re going to score 5+ and take the lead, in which case you’re not going to bring Rodney back for the ninth anyway, because you have to bring in your “proven closer” Todd Jones.

(Just checking some real numbers to back this up, Tango shows that for 99-02, teams scored 4 runs 1.6% of the time and 5+ 1.3%.  So given that you score 4, there’s a 45% chance that you score 5+.  Of course, I have ignored the possibility that Rodney allows a run, and then you tie with the 5+, but would Leyland really want to bring him back in a tie game in the B9 if he had given up a run in the 8th?)

Of course, the whole move ended up blowing up when Rodney was horrible and had to be relieved by Miner, and though Neifi turned a 53 DP to end the inning, was pinch hit for in the top of the ninth anyway.  I’m overanalyzing this simple procedural move that didn’t change the outcome at all, but it still makes no sense to me.

Yes, the IBB to Molina to pitch to Taguchi was terrible.  Didn’t even make any sense from a “manager’s” point of view.  Basically, managers overvalue the force at home or the DP at the expense of adding the extra baserunner.  You hear announcers all the time say something like, “He can pitch to player A or he can walk player A and pitch to player B.  Which one would he rather face?” As if the player A just disappears if you walk him!  It is of course, “You can pitch to player A with a base open, such that the walk is not terrible thing, or you can pitch to player B with first base ocupied, which means that you probably don’t want to walk B and A might even score if you don’t get B out.” IOW, managers completely underestimate the cost of putting another player on base and “taking away” the walk to the next batter, especially when you walk the bases loaded.  Again, if you read The Book, you will see how relatively rare it is to be correct to walk a batter and you will see how most managerrs walk too much and few if any walk too little.  In fact, in the chapter, Andy (or was it Tango) says something like the IBB should be such a rare occurrence that the manager can completely take it out of his arsenal and won’t cost him any significant WE over the course of a season.  That is good advice.  That was a terrible walk of Molina, again, one of the worst hitters in the lineup (about as bad as Taguchi, maybe worse).  If you read The Book, you will see that for starters, the next batter almost always (maybe always) has to be considerably worse than the one you are going to IBB to even CONSIDER the IBB.

One of the reasons I started this thread is that I wanted people to be aware of how many mistakes, often obvious ones (from a sabermetric perspective of course) at that, managers make ON A DAILY BASIS.  I would like someday for there to be a “record” of all those mistakes (on a random sample of games) so that we can estimate the total cost of “poor” managing per season.  My guess has always been in the neighborhood of 2-3 games, not including bullpen managment and 4-5 including that.  I am figuring roughly 100 blatant mistakes a year at 2-3% cost per mistake.

Patriot, Fans consider Inge one of the best fielding 3B around.  I think UZR/Dewan do as well.  It was not a “for fielding purposes” kind of move.  It was a pure “Inge made last out in previous inning, so put the pitcher there, and I’ll put my best available fielder to replace that hitter” move.

Speaking of which, the Win Prob numbers (for the Cards) for Zumaya throwing to 3B instead of 2B, it was bottom of the 7th, no outs, guys on 1B, 2B, and Zumaya got a perfect hit back to him (ala Mariano Rivera in 2001):
1b,2b,0 outs: .922

Go to 3B, and HOPE for a DP to 1B, with Pujols, a good runner, running, but Inge a great thrower.  Still Inge has to charge to 3B, so I’m not sure how well-balanced he’d be for a subsequent throw.  So, you could have a:

1B, 2B, 1 out: .898
2B, 2 outs: .873

Go to 2B, and go for two.  This is the easier play, since the ball was hit back to the pitcher so fast, and the throw from 2B to 1B is much shorter, than from 3B:

1B, 3B, 1 out: .915
3B, 2 outs: .876

As you can see, even if the DP was a sure thing, you are talking about a difference of .003 wins (.873, .876).

When I’d play ball, before every play, I’d scream out where every one should play if the ball is hit back to the pitcher, so that we’re all ready.  I’m assuming professionals know this already.  For Zumaya to go to 3B was crazy.

If it was a tie game, the numbers for Cards winning are:
1B, 2B, 1 out: .649
2B, 2 outs: .564

1B, 3B, 1 out: .705
3B, 2 outs: .574

Now, it gets a bit more interesting.  That’s a .010 win advantage to Zumaya to get the lead runner off 3B (with a 2B), against a .056 win disadvantage for not getting the twin kill.  The stakes are higher here.

Of course, I’m ignoring the possibility of disaster striking, which is huge, as we know.  If the chance for disaster was greater if he threw to 2B (because the throw was longer for him, ala Mo/2001), then Zumaya may be right.  But, I’d have to think the disaster rate must be around the same.

I’m always shocked when a pitcher is the one who can’t hit a target.

In any case, I’d like to hear Zumaya’s reasoning.  I get the feeling it’ll be similar to Laroche’s decision to bunt in the 9th inning earlier this season.

Patriot,

That move by Leyland is about as strange as it gets.  I agree that it had no effect on the game, nor could it have been expected to, it’s not even close to be one of the 2-3% of a win errors that MGL discusses.  Still, it stands out because it’s so obviously indenfensible.  (Maybe you could try to make the argument that Inge is a power-only guy without a great OBP and thus becomes worth much less when you need baserunners ... but he was replaced with Omar Infante, who can’t have a highed expected OBP.)

My best guess is what Tangotiger was getting at. Leyland was just applying the rule “In an NL game, if you’re switching pitchers, do a double-switch when the pitcher is due up next” without considering if the rule made sense under the circumstances.

The BPro analysts (Sheehan and Silver) are really hammering hitters for a failure to take more pitches—the Cardinal hitters in game 2, Tiger hitters in game 3.  For sure, both Rogers and Carpenter got through on very few pitches.  But Carpenter, at least, was clearly throwing strikes (I’m less sure about Rogers).  Would hitters really have done better by getting to more 0-2 and 1-2 counts before swinging?  Taking pitches has become a mantra in (parts of) the saber community, but against pitchers like these a hitter may do at least as well by swinging at the first semi-hittable pitch he sees.

How do we determine when/whether more patience is called for?

It is not an easy analysis.  The new mantra about being patient, working counts, taking pitches, etc., is laughably simplistic and doesn’t really get you anywhere in terms of an “analysis.” It’s like when Buck or McCarver scream when a batter swings at the first pitch and makes an out (if you knew that you were getting a good pitch to hit on the first pitch you should probably swing at it) or when someone thinks that the batter should take the first pitch if the pitcher “has been” wild or has just walked someone or thrown X amount of balls in a row.  All of that rhetoric is stupid.  I think once you start writing for a mainstream publication you become another talking head.  For example, some of the stuff that Keith Law writes is just plain idiotic in his ESPN columns.  I thought that he was a good sabermetrician once upon a time.  Maybe not.  I even thought about starting a “Keith Law” thread here.

Whether a batter is or should be patient is of course a complicated thing.  What does “patient” or “not patient” even mean?  Obviously it is a continuum.  The ultimate goal for each batter is to optimize his offensive production, NOT to try and make the pitcher throw as many pitches as possible.  Of course if he can do that and still retain most of that optimization, then fine. 

One player’s patience is another player’s impatience.  For whatevre reason, every player has a unique optimal level of patience, again, whatever that even means.  I mean we can criticise AB’s until we are blue in the face and it doesn’t really mean anything.  Batters swing at bad pitches sometime for various reasons.  They also take good pitches some of the time.  It depends on what they were looking for, how good of a pitch it was, along with a myriad of other factors.

And of course you are 100% right in that the more strikes a pitcher throws, the more the batter should be swinging early and often.  And vice versa.  Also the more predictable a pitchre is in terms of his pitches and locations, the more a batter should probably be swinging.  And of course during the game it depends on the inning/score/count, etc.  For example, with a runner on 3rd, less than 2 out, and the run is important you want to put the ball in play early in the count and avoid getting 2 strikes if you can help it.

The only real way to “gauge” or analyze patience I suppose is to look at how often each batter swings or doesn’t swing at each kind of pitch in each location and on every count and in different game situations.  Even then, as I said, every bater is different.  You can’t force a batter to be more patient in many cases.  Some batters are successsful being more patient and some are not.  Obviously the better batters tend to swing at the better pitches and lay off the poorer ones, but it is not that simple.  If you swing hard, you can afford to swing at worse pitches.  If when you swing you tend to be able to adjust to the pitch, you can also swing at worse pitches.

It is such a stupid and meaningless thing to rant about batters being patient or not I am having a hard time explaining what I mean.  I think that you (Guy) and I are on the same page though.  All their ranting on BP is just sports journalistic crap. It is the same thing with pitchers. How often do you hear that pitchers should be aggressive, should throw strikes, should not nibble, etc.  That is all crap too.  Pitching is way more complicated than that and to frame it those ways is just plain stupid.

Well, no major gaffes today in the WS!

I put the end of game 2 in my sim to see the difference between Molina and Duncan (or J-Rod) batting in the top of the 9th with 2 outs and the bases loaded down by 2 runs of course.

With Molina, Cards win 6.8% of the time.  With either Duncan or J-Rod, it is 10.1% for a gain of 3.3% in WE, a fairly significant amount I think.

That increases the Cards chances of winning the series by around 1%.  If winning a WS is worth 10 mil over the next 10 years or so, that “error” was a $100,000 error.  I wrote to the Cards FO and suggested they take that out of LaRussa’s paycheck.  Do you think he would understand?  After all, he is a “numbers guy,” right?

MGL:  What happens if you put in J-Rod and Duncan for Wilson and Molina (2nd and 3rd, 2 out)?

I posted this on another forum regarding Kenny Rogers:

It wouldn’t surprise me if the pitcher has a choice between a rosin bag or a pine tar rag in the near-future.

This is as much cheating or “law breaking” as the third base coach being half-way down the line, instead of being in his box, or me spitting on the sidewalk.  You get a slap at the wrist at most, which is what happened to Kenny.

I agree that there are many unwritten rules in baseball and many rules which are routinely and innocuously broken or not followed and I have no problem with that.  However, I have never heard of pitchers using pine tar on their hands, in cold weather or otherwise, I have never seen a smudge on a pitcher’s hand like that, and apparently neither have many of the players (otherwise why would they have even brought it to the attention of LaRussa and why would LaRussa even mention it to the umpire?).  So I don’t know that this is the same thing as coaches being out of the box, catchers blocking the plate without the ball, the vicinity play at second base, etc.  I really don’t.  I also didn’t like the idea of Rogers lying through his teeth.  I really think that LaRussa should have insisted that the hand be checked and if it were pine tar he should have been ejected.  What is to stop a pitcher from using the pine tar to “juice” the ball?  I don’t think many pitchers routinely use pine tar on their hands.  I really don’t.  I think your sugestion that they will eventually allow it is preposterous.

From the little I’ve read, it seems to be a common occurrence to get *something*, be it pine tar, or whatnot, to get a better grip on the ball in cold weather.  Rogers’ was more egregious in that he had it right on his hand, rather than greasing up between innings.  It seems that the unwritten code of conduct is “if it’s cold, do what you need to get a better grip, but don’t make it obvious”. 

And the reason we don’t hear about it is because ballplayers know what a bunch of tightasses the media is, that will blow everything into a story.

As I said, I really don’t think that many ballplayers are even aware of this.  Otherwise, why would the STL players and LaRussa have even complained if it is “commen practice.” In fact, why would the Cards hitting coach be screaming bloody murder?  Whatever Rogers was doing, I don’t think it was a common practice and I don’t think it should have been ignored.  You don’t see opposing managers complaining when third base coaches are out of the box, although you do occasionally see managers complain on other “commen practices” like batters out of the box and vicinity plays at second…

Yes, Guy, I think that Wilson should have been pinch hit for as well, although that was not nearly as egregious as not hitting for Molina.  Wilson is a much better hitter than Yadier.

With Wilson batting with runners on second and third (and then Molina if Wilson reaches), the Cards win 7.2% of the time (in a sim of 100,000 games).  With Duncan or J-Rod (they are essentially the same) betting for Wilson, they win 8.2% of the time.

With Duncan batting and then J-Rod if Duncan reaches, the Cards win 8.8%.

With Wilson and then Duncan if Wilson reaches (such as what happened), it is also 7.2%.  That should be higher than Wilson/Molina of course, but we have sample error, even in 100,000 games.  One SD is .0008 in WE or .08%.

LaRussa puts on the hit and run in the 4th with 1 out, the pitcher, Suppan up, and Eckstein on first.

First of all, Suppan hitting away is not a bad idea.  As I explain in The Book, a sac bunt by a pitcher WITH ONE OUT is such a marginal play that an average hitting pitcher or bettre should swing away half the time.  So I have no problem with Supan swinging away.  In fact, more managers should do that - they don’t.

This was a hit and run though.  I think that hit and runs are almost always wrong, but I don’t know that for sure.  I have never done nor have I ever seen an analysis.  Basically you are forcing the batter to swing at a pitch he might not swing at, he usually does not swing as hard, you get a few extra LD and FB DP’s, and the runner gets thrown out a high % of time if the batter does not make contact with the pitch.  In return, you stay out of the GDP DP and you get a few extra base hits on ground balls.  I think it is almost a net negative, but as I said, I am not sure.

What I am pretty sure of is that if a hit and run IS going to be a net positive, two things have to be present.  One, the batter has to be able to make decent contact at any pitch.  IOW, he should not be the type of batter that swings and misses a lot and he should be able to make contact with a bad pitch.  With Suppan that is definitely not the case.

Two, is that the runner should have a decent chance of reaching safely if the hit and run is bustes, maybe 60% or more.  Again, definitely not the case in this situation with Eckstein, not a good basestealer, and Pudge, still the best throwing catcher in baseball, behind the plate.

So I don’t think that this can possibly be a net positive (the hit and run).  In fact, I can’t even imagine what LaRussa had in mind, since he knows all of the things I have just mentioned.  I think it is just another example of him doing unconventional yet wrong, things, in order to look like a genius.  That is the kind of managrer he is.  How much that hit and run cost in terms of WE, assuming that it was wrong, I have no idea.  Probably less than 1%.

Wilson is obviously much better than Molina, but he ain’t great, especially against RHs.  I wonder why he’s in lineup tonight, instead of J-Rod.  Did J-Rod sleep with Tony’s wife or something?

LaRussa has now twice sacrificed with man on 2B and none out.  I can see this in the 9th, or maybe the 8th (when down 1 at home), but hard to see how this is a good move in the 6th and 7th innings.

Guy, read (or re-read) the chapter on sac bunts in The Book (anyone on this site who has not read it should be banned!).  There are many variables which make it right or wrong to sac bunt, regardless of the inning, and in many cases the manager must mix it up (bunt and non-bunt some percentage of the time, even in precisely the same situations) so that bunting or not-bunting can both be “right” (and one would have to look at the ratio in the long run to determine whether the manager was acting properly).

Easy rule of thumb for sac bunting with a runner on first or a runner on second:

Batter must be a good bunter and a fast runner.  Early in a game, if the defense is playing agnostically, then the bunt has about the same WE as a non-bunt.  Late in a close game, if the defense is playing well up in anticipation of a bunt, then a bunt is also around the same as a non-bunt.  If the batter is either slow or a poor bunter then a bunt probably reduces the WE no matter where the defense is playing.

In a situation where the bunt is in fact warranted (see above) the defense MUST bunt sometimes and not bunt sometimes to prevent the defense from paying even more optimally.  Of course you cannot tell if a manager is doing this from observing one situation.

In general, where the defense is playing determines more than anything else whether a bunt increases or decreases WE (in conjunction with the quality of the batter, the game situation, the speed of the runner(s), and the speed and bunting ability of the batter.

If the defense tends to play optimally, you MUST use game theory on offense to mix up your bunts.  If the defense is not playng optimally, then you must either bunt or not bunt 100% of the time until the defense starts to adjust.  That is a huge advantage for the offense - the fact that they get to see where the defense is playing before having to make a decision.  In many “games” that is no the case - each player has to act in the dark and try to anticipate or “guess at” the strategy of the other player. 

As I discuss in The Book, the notion that bunts may be correct late in a game but not early in a game, is simply not true.  It would be true if the defense played the same early and late in the game, but they don’t.  You do and should in fact bunt far less early in a game than late in a game, given the same batter, runner, etc., and the defense knows this which is why they play more in late in the game than early in the game. 

So early in a game, you might bunt (with a good bunter/fast runner, etc.) maybe 20% of the time and late in a game, maybe 50% of the time, again, given the exact same situation, personnel, etc. (other than the inning of course).  So if you observe a manager bunting or not bunting either early or late in a game, you CANNOT determine whether it was correct or a mistake!  This is a difficult concept to get ones arms around.  As I said earlier in this post, you would have to look at many bunt situations in order to determine how optimally a manager is using his bunts and non-bunts.

It is the same thing with observing a certain pitch by a pitchre and saying that is was a bad (wrong) pitch.  Let’s say that in a certain situation, the best pitch is a curveball and it should be thrown 90% of the time and that the fastball is the “worst” pitch to throw but that it should be thrown 10% of the time nonetheless (in almost any situation, game theory dictates that almost every pitch must be thrown SOME percentage of time to keep the batter from being able to predict what the pitcher is going to do).  Well, what if you are watching a game and you see the fastball (and the batter happens to crush it, even though he should not be looking for it of course).  Did the pitchre make a mistake?  Most people would say yes.  We cannot tell however (for sure).  What if the pitcher tells us after the game that was his “10% pitch.” We would have to look at all of those same situations and see how often the pitcher threw the varous pitches to see how optimally this pitcher was pitching.  It is actually more complicated than that and there is actually a significant chance that the pitcher did in fact make a “mistake” (had the wrong % in his head for that batter in that situation) since we did observe the unlikely pitch even in one observation.  It is like if we have two bags of balls, one 75% red and the other 75% white and we tell a person to grab a ball out of one of the bags and he shows us a red ball, what is the chance that he pulled a ball out of the first bag?  Bayesian probability.  Anyway, that is a similar situation to oberving a bunt or non-bunt and trying to figure out whether the manager made a mistake.  Unless he is supposed to bunt 100% of the time (which is rare) or never (which is more often), we cannot really determine whether the manager made a mistake whether he bunts or does not bunt, no matter how thoroughly we analyze the WE’s.

One correction.  If the defense is playing optimally, you do NOT have to bunt or not bunt a certain optimal percentage of time.  That would be the case if you didn’t see the defensive alignment before you make your bunt/non-bunt decision.  But since you always see them first, the onus is ALWAYS on the defense to play optimally.  If they don’t, they are in trouble (you would eithre bunt or not bunt 100% of the time).  So when they are playing optimally, you can do whatever you want and it does not matter.  That is assuming that their optimal position is not either 100% in or 100% back.

Now, since that is the case, once the defense plays optimally, you want to try and bait them into changing their defensive position by either bunting or not-bunting 100% of the time.  If they change their alignment, which they should not, you do the opposite.

I don’t disagree with the game theory implications, but at some point that means you can’t evaluate ANY decision, only a manager’s career (or maybe a season).  I think it’s still valid to ask:  “setting aside the need to sac in some non-optimal situations, did a sac bunt make sense in X situation?” A successful sac in the 6th inning, down 1, reduces WE by about .03.  That’s pretty big.  Obviously, it’s more complicated than that:  sometimes the runner will reach (as So in fact did in the 7th), and sometimes the runner will fail to advance.  And defensive alignment has a big role.  All that said, assuming the Tigers are at least somewhat anticipating a sac in both situations, were these good percentage plays?

As you say, the WE after a “successful sac” (out, runner advance) tells you nothing.  You have to know or estmate how many hits and RBOE’s will occur, which is basically a function of the batter’s speed and bunting ability and the position (and defensive ability) of the defense.

That being said, if a team is supposed to bunt sometimes and not bunt other times, which is almost always the case in an obvious potential bunt situation, then yes, it is true that you can NEVER evaluate a “one time” decision.

There is a caveat, however, or more accurately a qualification.  As I think I explained above, since the offense gets to see the defense first, then they DON’T really need to mix up their strategies.  So I am going to do a complete turnaround and say that you are right.  Once we establish where the defense was playing, we can indeed say whether a bunt was correct or not (whether it reduced or increased the WE as opposed to hitting away).  Of course it is even more complicated than that from a game theory perspective, as the manager usually has to make a decision whether to bunt or not BEFORE he sees the defensive alignment (he can try and guess of course how they are going to play).  In that case, then he MUST mix up his strategies.  Of course the batter himself can always switch his strategy after he sees the defensive alignment.

Anyway, in order to do a “one-time” analysis (and it is complicated and not easy), we have to have observed the defense in order to estimate the RE or WE of the bunt attempt (remember that there is a huge difference in the RE or WE of the bunt attempt between the defense playing up and back) and then we have to compare that to the RE or WE of the particular batter hitting away.

From my research however, as explained in The Book, as long as the batter is a good bunter and has good speed, and the defense is playing somewhat optimally (considering their estimation of how often the batter is going to bunt), a bunt attempt early or late in the game usually yields around the same WE as hitting away.

So my guess is that those sac bunt attempts last night were just fine.

About the only time that sac bunt attempts by a non-pitcher are clearly incorrect is when the batter is either slow or a poor bunter, or the defense is playing so sub-optimally (either too far back or too far in) that either a bunt or non-bunt is clearly correct and the manager does the opposite (which happens all the time, although it is not the “norm").

No one cares anymore, but…

100% blunder by Leyland pinch running for Casey in the 9th, for obvious reasons.  No strategy there - just a brain cramp.

LaRussa should have pinch hit for Weaver with runners on 1 and 2 in the 4th and one out, and then again with a runner on 2nd in the 6th with 2 outs.  If you don’t pinch hit in the 4th, then the sac bunt is a TERRIBLE play, especially with Molina on 2nd and Weaver being a decent hitter.  I can’t emphasize how much of a bad play that is, sacrificing with ONE out and a runner on second, even with a runner on first also.

But no one cares how bad LaRussa manages, right?  Count the rings, as Tango likes to say…

Don’t you think pinch hitting for Weaver in the 4th is pretty risky, in terms of wearing out the pen, when you may have to play games the next two days (and you have limited confidence in Reyes in game 6)?  And Weaver had pitched well, other than Casey’s HR.

Not an especially memorable postseason is all I have to say. Yuck.

That is an awful lot of RE to be giving up not pinch hitting in the 4th inning.  I don’t see how any negative in terms of the bullpen is not worth it.  And our research suggests that when a pitcher is pitching well in a game it has no predictive value, not to mention the fact that Weaver rarely pitches more than 6 innings anyway.

I can live with not pinch hitting for him in the 4th.  I can’t live with the sacrifice bunt or not pinch hitting for him in the 6th.

Why is there “limited confidence” in Reyes?  He is the second best pitcher on the team, by far.  That LaRussa and staff do not know that is further evidence of their ignorance.

I mainly meant that you would likely need the bullpen in a Reyes start, as he averaged only 5.0 IP per start.  But it’s interesting that you consider him better than both Wainwright and Suppan “by far”.  His minor league stats are good enough to make you that confident?  (happy to hear it, as a Cards fan)

Actually, I though the Book presented pretty good evidence that a starter’s early performance IS somewhat predictive of what follows.

Reyes’ minor (MLE) and major performance are stellar.  Wainright is of course a reliever for now, but I think that Reyes is better (as a starter), but we’ll see with Wainright.  Suppan is a league average starter.

I think The Book shows that early bad performance is predictive, mostly for young pitchers, and that early good performance is not predictive for anyone.

Starters who did very well against the first 9 hitters did somewhat better than expected after that, though not statistically significant.  And pitchers who do well third time through lineup do very well after that.  And we don’t know if other indicators might be even more predictive:  for example, Weaver’s unusually high K-rate in the early going.  Certainly there is some tendency to pitch well or poorly on a given day, and to me it’s an area worthy of more study.

Sure, I agree to some extent.  As I said, I could live with him batting in the 4th.  Pinch hitting in the 6th was I think a no-brainer and bunting with 1 out and a runner on second is horrendous especially with one of the slowest players in baseball on second base.