Dumb and Dumber (Farrell and Matheny)

Posted: October 27, 2013 in In-game strategy, Managers, Post-season

My father had this running gag whenever someone in the family would do something stupid. He would say (affectionately of course), “You know, if there were a contest for idiots, you would come in second place!” Invariably the transgressor would reply, “Why second?” to which my father would gleefully  exclaim, “Because your an idiot!”

Don’t worry, we never understood it either.

In last night’s Game 3 of the World Series, the Cardinals’ and Red Sox’ managers, Matheny and Farrell, probably tied for first in my Dad’s idiot contest. As exciting as the game was, it was also painful to watch. It was a managerial comedy of errors. It was like the Keystone Cops meet the Three Stooges, or an episode of Gilligan’s Island where the group is just moments away from being rescued and Gilligan does something stupid at the last moment. We can all probably get together and file a class action lawsuit against those two managers for intentional infliction of emotion distress (although our lawyers would be too busy with the other lawsuit against Tim McCarver for the same thing). I also think that if Bill James were dead, he would be turning over in his grave about now (sorry Bill).

Of course it is not PC to criticize a manager when he wins the game or his particular decision “works out” but I don’t play that stupid game. A mistake is a mistake is a mistake, regardless of how it turns out or who wins the game. All of you would be more successful in life and a lot smarter if you would analyze your decisions independent of the results of those decisions when the connection between the decision and the results is tenuous, which is almost the case in baseball. Imagine this: Your manager has two choices. With one of those choices, his team is supposed to win 80% of the time and with the other, his team is supposed to win 79%. So clearly choice A is the right choice and choice B is the wrong choice. Let’s say that he makes his decision and we don’t know whether he chose the right one or the wrong one. How helpful is the result in us figuring out whether he made the right or wrong decision, assuming that there is an equal chance of him making one or the other?

If his team wins, which is likely whether he makes the right or wrong decision, there is a 50.3% (80/159) chance that he made the right decision and a 49.7% chance that he made the wrong decision. In other words, not very helpful. The outcome of the game barely helps us determine whether he made the right or wrong decision. That is why we don’t use it in our evaluation process. At all. 50.3 to 49.7 is essentially 50/50. Regardless of how the decision “turns out,” if that’s all we know, we have gained virtually no information. If there was a 50/50 chance that he made the right decision before the outcome, there is a still around a 50/50 chance that he made the right decision after the outcome, whether it turns out good or bad. (If he ends up losing the game, there was a 51.2% chance that he made the wrong decision.)

Now, I’m not going to talk about not pinch hitting for the Cards’ starter Kelly with bases loaded and 1 out in the 4th inning and then leaving him in there for a grand total of another 4 outs. I’m also not going to talk about bringing in Choate to face Ortiz and then removing him for the righty Maness to face the righty-killer Nava, rather than bringing in Siegrest (who is actually better than Choate) and then leaving him in there to face Nava (who would have to bat from the right side, or perhaps Gomes would have pinch hit for him). I am not going to mention the foolish IBB of Ortiz in the 8th, or the equally foolish IBB of Molina to face Freese. I am certainly not going to talk about letting Workman bat for himself (what was that all about?) and then taking him out 3 seconds later. Or Beltran’s bunt on a 3-1 count early in the game (although that one is probably on Beltran and not Matheny).

Note: I just read this quote from Farrell:

Boston Red Sox manager John Farrell said “in hindsight” he should have avoided having rookie reliever Brandon Workman bat in the ninth inning of Saturday’s Game 3 of the World Series.

Eh, it’s only the World Series. No big deal.

Anyway, what I really want to talk about is whether or not it was correct to IBB Jay in the bottom of the 9th with 1 out, runners on second and third, tie game, Kozma on deck and Koji on the mound. I don’t think that they had anyone to pinch hit for Kozma and I think that Wong, another light hitting infielder, was batting after Kozma. I think that the IBB was in order there, but honestly, I am not nearly sure. It seemed to me that most managers would have issued the IBB, but that Farrell is pretty stubborn about not doing some things that most managers do, like issuing IBB’s and attempting sacrifice bunts.

Figuring out the exact win expectancies for each alternative is difficult in this case. Instead, I am hoping that the decision turns out not to be close, one way or another. Sometimes when the analysis of one decision over another is difficult to do, we can only hope that a rough analysis results in one alternative being much better than the other. If that is the case, we can generally say that we have identified the “right” decision, even if our analysis is far from perfect. It it ends up being close, even if one decision is slightly favored over the other, we would call it a tossup, again, if our analysis is rough. Sometimes it is simple to evaluate two decisions. In those cases, even when we find a small difference, we can often say which one is right and which one is wrong with a high degree of certainty.

Here is what I am going to do with this one. I am going to look at situations late in the game with a very good pitcher on the mound, runners on second and third, 1 out, and the infield likely playing up. The infield should be playing in with any tie game or one in which the fielding team is losing in the 8th or 9th innings. Let’s see how often the defense escapes the inning without allowing a run.

Then we’ll do the same thing with the bases loaded and a couple of weak hitters due up. I think it is reasonable to assume that Jay represents a somewhat average batter and that Kozma and Wong represent very weak hitters. Perhaps I’ll look at bases loaded situations and the #8 hitter due up. With runners on second and third, maybe I’ll look only at situations where the #7 batter is due up. I might look at all pitchers rather than just very good ones like Koji. I don’t want to have tiny sample sizes. In most of these situations, there is likely to be a very good pitcher on the mound, and in any case, we are mostly interested in the difference between runners on second and third, and the bases loaded, so the exact quality of the pitcher is not that important. I might, however, only look at pitchers with low walk rates. If you are going to walk the bases loaded, obviously you want your pitcher to have a low walk rate. You can’t get a much better pitcher in that regard than Uehara! Let’s see what the data says:

Let’s start out simple. We’ll look at all situations as I describes above, either runners on second and third or bases loaded, in the 8th or later, with any pitcher and any batter (other than a pitcher) at the plate.

Runners on second and third, 1 out, no IBB, 1998-2012

No runs score 37.2% of the time (plus or minus 2.5%). N=1621.

The batting pool had a .332 wOBA and the pitching pool, .327.

The average batter in the league for these seasons was .340, and the average pitcher, .339.

Now let’s compare that to the bases loaded, again, presumably with the infield playing up or for the DP – in any case, trying not to let any runs  score at all.

Bases loaded, 1 out

No runs score 33.6% of the time (plus or minus 1.1%). N=3452.

The batting pool had a .338 wOBA and the pitching pool, .329.

So we actually do have more scoring with the bases loaded, although the batting pool is slightly better than with runners on second and third (which is probably to be expected since you would tend to not IBB the batter if he is a weak batter).

Let’s see what happens if we restrict the bases loaded batter to a RH batter with a RHP pitcher on the mound.

Bases loaded, 1 out, RHB and RHP

No runs score 35.4% of the time (plus or minus 2.4%). N=1750.

The batting pool had a .338 wOBA and the pitching pool, .330.

I presume there are more GDP with a RHB and lesser offense with the pitcher having the platoon advantage. But still not as good as pitching with runners on second and third.

Let’s look at the #7 and #8 batters only with the 8th batter being RH and the pitcher RH:

Runners on second and third, 1 out,  #7 batter at the plate, #8 batter is RH

No runs score 32.9% of the time (plus or minus 11.1%). N=85 (oops).

The batting pool had a .313 wOBA and the pitching pool, .325.

Now we have such a small sample, the number is unreliable.

How about bases loaded with the #8 hitter, a RHB, due up?

Bases loaded, 1 out, #8 hitter, a RHB, and a RHP

No runs score 39.9% of the time (plus or minus 4.8%). N=421.

The batting pool had a .315 wOBA and the pitching pool, .333.

This is probably closer to the situation we had in the game. A weak #8 and #9 hitters and the pitcher having the platoon advantage on that #8 hitter.  This is actually the highest “no score” situation I found so far. The sample size is still fairly small, so we are not very certain of that 40% no score numbers (it is 35-45% at the 95% confidence level).

Let’s try one more thing. Let’s limit the pitcher to one who has a very low walk rate. I think that is critical in deciding whether to issue the IBB or not for obvious reasons. I only looked at pitchers with a below average walk rate for that season. Otherwise I just limited my sample to RHP and RHB batting with the bases loaded or next with runners on second and third.

Runners on second and third, 1 out, low walk RH pitcher

No runs score 38.3% of the time (plus or minus 4.2%). N=561.

The batting pool had a .330 wOBA and the pitching pool, .325.

Now let’s compare that to the bases loaded, again, presumably with the infield playing up or for the DP – in any case, trying not to let any runs  score at all.

Bases loaded, 1 out, low walk RH pitcher, RHB at the plate

No runs score 35.3% of the time (plus or minus 3.6%). N=780.

The batting pool had a .338 wOBA and the pitching pool, .330.

So, again, the base loaded is worse  as far as preventing any runs from scoring, but we have a better pool of batters in the inning. That is because, as I stated before, historically managers will tend to pitch to the batter with first base open if he is a poor hitter. In our situation with Jay at the plate, he is not a poor hitter and he has the platoon advantage (although Koji has virtually no platoon splits).

I guess the final verdict is that it is inconclusive, but I lean towards thinking that that not walking Jay was the correct move. Certainly in that spot you are trying to strike him out and you don’t mind the unintentional walk (although Jay is trying to do the exact opposite). As I said at the outset, and I am a man of my word, if an incomplete analysis, which this surely is, yields results that are close or even ambiguous, and I think that is true as well, we can’t really conclude anything one way or the other. I guess we can give the benefit of the doubt to the manager, although I don’t think that either one has demonstrated that he is worthy of that!

Finally, I want to say a few words about the  obstruction call. Not that it hasn’t been discussed already a million times on the web and elsewhere. There really is no controversy, or at least there shouldn’t be. The call was 100% correct according to the rule book and there would be no reason not to call it according to the rule book. If the obstruction call had not been made, it would simply be a bad, missed call and the Cardinals would have had a right to be furious and perhaps been able to file a protest, since there really is no judgment involved with that call in that situation (although they would probably lose a protest on the grounds that is was a judgment call) . The rule clearly states that a fielder when not in the act of fielding a ball or receiving a throw, and I am paraphrasing, may not impede a runner in any way shape or form. There is no intent necessary. In other words, it could be by complete accident, for example, the fielder could be lying dead on the field, or it can be an intentional act by the fielder. The umpire, thankfully does not and did not have to make that judgment. All that was necessary was that the fielder was not in the act of fielding a batted ball, which Middlebrooks wasn’t, and that he was not in the act of receiving a throw (which requires that the throw be on the way, by the way), which he was not, and that the runner be impeded in any way shape or form, which he was. Obstruction. Q.E.D.

A few people including Middlebrooks himself, were barking about “the baseline.” The baseline has nothing to do with this call. Neither the runner nor the fielder must be or not be anywhere in particular. The assumption of course is that the runner does not completely alter his direction in order to “throw himself” in front of a fielder, but clearly that was not the case here. If you want to invoke some kind of “baseline” argument (which, as I said, is a strawman argument since the rule has absolutely no “baseline” requirement one way or the other), the generally accepted definition of a baseline is that which the runner creates, not some straight line between the bases. If the obstruction rule required that a runner stay withing some pre-defined baseline like a straight line between the bases, imagine this play: A runner rounds third base trying to score. He is around 4 or 5 feet outside of “the baseline” between third and home as he rounds third, the normal position for a runner trying to score. At that point, a fielder steps in front of the runner and the runner does a flip over the fielder lands on his back and the throw beats him home. The fielder is not guilty of obstruction because the runner was “outside of the baseline,” right? No, I don’t think so. Anyway, the baseline has nothing to do with this rule. Read it. It is in the definition of terms in Rule 2.00, and it is in rule 7.06, under the runner. The intent of the rule as it is written is obvious. A runner can never be out just because he trips over, bumps into, or is impeded in any way by a fielder, regardless of whether the fielder intended to impede him or not. The runner must allow a fielder to field a ball and sometimes to catch a thrown ball, but absent that the runner has the absolute right to advance or return to any base without being impeded by a fielder (presumably without the runner deliberately veering off his own base line in order to create obstruction). Period. End of discussion.

  1. Brian says:

    This is a brilliant piece – thanks. The whole game I was so frustrated by nearly every decision Matheny made (I’m a Cards fan), then usually about 5-10 minutes later Farrell would trump him with an equally bad, and sometimes worse, call of his own. It really was a game of follies, by both managers and players. It was a game of follies for the umps too, but oddly not by Joyce (who clearly got his big call correct), but by Demuth’s horrible strike zone. Now that’s something Sox fans really do have a right to gripe about.

    • MGL says:


      Yes, I agree that that Demuth’s strike zone was particularly bad – probably horrendous – in it’s inconsistency. We can excuse umpires for their “personal” strike zone – many of them have their own (wait until Miller umps game 5 – he has a huge zone). But, we do not excuse inconsistency like Demuth’s last night. I have no idea whether he is normally that bad or last night was just a terrible night for him.

      Lately, I have been clamoring for using a robot strike zone. Pitch trax or pitch f/x (I think they are one and the same) is nearly perfect. I have never seen it make a mistake – not even an occasional glitch. Obviously it is not perfect, but it is far better than a human umpire can ever be. I think it would take about a week to get used to. Most of us are already used to seeing the pitch trax on the screen and I think most of us secretly wish that the result of the pitch trax was what the umpire actually called. I can’t imagine too many people thinking, when pitch trax shows a pitch misses by 3 inches (as with Demuth last night on some of his calls), “Nice job by the umpire! Man, I love that human element!” I think most people mumble, “That was a really bad call,” to themselves or out loud.

      I predict that withing 20 years we will have computers calling balls and strikes.

      • DavidJ says:

        Well said. Has anyone ever watched a well-umpired game and felt disappointed that no egregious mistakes were made? That’s what I always wonder when I hear people wax poetic about “the human element.”

      • Brian says:

        Rob Neyer had a great retort to anyone who says they want to make “the human element” a permanent part of the game: just what exactly IS the right amount of human element? If the umps goof up 7 strike zone calls a game, why not 10, or 15, or 20? Why is 7 better than, say, 2? How many screwups do we like on the base paths – will 3 per game do it? I’d honestly like to hear someone answer these.

  2. Florko says:

    Where strategy is more important than leadership!!

    • MGL says:

      Well, I don’t know which is MORE important, but like the commercial says, “I’d rather have sweet AND sour chicken, or nuts AND bolts…” Why not good strategy AND leadership. All it would take is for a manager to on his own, or at the bequest of the front office, sit down with their analytics department (if a team legitimately has one – they all CLAIM to have one) and learn some rules of thumb as to what is correct and what is not correct and to disabuse them of the faulty notions and belief systems that virtually all managers hold that cause them to make poor decisions.

      I am available, for a price of course!

      • Brian says:

        I agree, MGL – it’ll probably take about a generation for umps and the public to come around to it, but computers calling balls and strikes seems to me an inevitability.

        Another prediction within 20 years – the dugout will be structured more like the NFL, where you have a manager (much like a head coach) who makes final decisions and whose responsibilities lean toward the interpersonal, and a coach (much like an offensive or defensive coordinator) who specializes in strategy. He/she will probably have a tablet in front of him/her that spells out probabilities according to The Book – managers can then see these probabilities the same way we see probabilities while watching poker on TV and make decisions accordingly.

  3. tangotiger says:

    I agree that the robo-ump would do a far better job, and be more consistent. You could have a human ump override in “obvious” scenarios, but that should be limited to at most one per game. I’d even let the manager be the one to challenge the call!

    • MGL says:

      Sure, you could have a human ump or someone overrule an obvious bad call, But, it would have to be a glitch (IOW, you can’t overrule because the pitch trax said the pitch nicked the strike zone, but you claim that it was 2 inches outside the zone). I have NEVER seen pitch trax have a glitch. The pitch f/x guys would have to chime in on this, but I would think that it is less than 1 pitch in a thousand. Maybe more like 1 in 10 thousand.

      Do you realize how often a human umpire actually calls a pitch right down the center a ball? Or a pitch that is 6 inches out of the zone a strike? A lot more than a pitch trax glitch would occur! And those are not reviewable! So I am not even 100% convinced that we would need an overrule system, unless the computers just went haywire or something like that.

      I think that everyone is in agreement that a robo-ump would do a much better job (I mean, 100 times better). The main point, I think, is that, in my opinion, almost everyone would get used to it quickly and like it. We have precedent in other sports, don’t we? Tennis?

  4. Florko says:

    I would love to know what Bill and his team were saying/thinking/breaking throughout all those questionable (wrong) choices by Farrell..

    • MGL says:

      That is a good question. I am thinking that it doesn’t bother them too much. If it did, they would have done something about it a long time ago. These mistakes happen in virtually every game during the regular season and thus even James and his minions must expect them to happen during the post-season. You can’t do what you don’t know. And clearly managers, including Farrell, don’t know what to do in many situations. As I always say, no one would, unless they did their analytical homework, or someone did it for them (I don’t expect a manager to be able to figure out whether an IBB in that situation was correct or not) beforehand. For example, as I said on Twitter last night, I had no idea whether an IBB was correct or not in the 9th. I didn’t know whether the IBB to Papi in the 8th was correct either (Tango said it wasn’t).

      You know the conventional wisdom about, “The team that is best prepared for a World Series game will have the best chance of winning?” Well, they forgot about being prepared strategically!

  5. Florko says:

    A manager’s job is simple. For one hundred sixty-two games you try not to screw up all that smart stuff your organization did last December-Earl Weaver

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s