Good piece. Yes humans make irrational decisions based on emotion all the time. However, humans who get paid millions of dollars to maximize their chance to win a competition yet fail miserably at that job can justifiably be fired.
Comment by Koko B Ware — October 17, 2012 @ 4:42 pm
I completely agree with you, Dave, but this got me thinking. How many 9th inning or later, go-ahead/game-tying home runs would Ibanez have to hit before it becomes rational to not pinch hit him? That is, at what point do you say, “Hey, I know this decision goes against the numbers, but he has come through too many times to write off as randomness.” For Girardi, clearly that number was around 5. But what should it be?
Or, to think about it another way, how many times in a row would a coin have to come up heads before you started believing that it wasn’t a fair coin? 10? 20? I have no idea what the answer is, but it’s an interesting question at the very least.
I voted for Ibanez in the poll based on how I felt at the time of the at bat: I didn’t have any particular confidence in Ibanez, but I didn’t think Coke was pitching all that well, and (had I been rooting for the Yankees) I would have preferred to give Leyland a reason to keep him in the game.
Absolutely. When I saw the posted lineup before the game I let out a formidable sigh of frustration. Gardner hadn’t had a major league at-bat since April and Nunez has proven to be a defensive liability. ARod also has decent numbers (as decent as you can get) against Verlander; if you ignored recent slumps and just look at the big numbers, the ones that have come over several games, the very obvious choice would have been to include Swisher and ARod, but Girardi felt he had to change it up. He made the wrong choice, and that was a crucial game to make it, possibly the most important of the series.
why is it so hard to believe that girardi (and kevin long, and cashman, and whoever else) saw mechanical flaws in arods recent stretch and decided ibanez was the more likely guy to reach base? they played him in the orioles series over ibanez because they figured hed eventually fix those flaws theyve been telling him to fix, and since he hasnt (because hes either stubborn or physically unable to), they decided he wouldnt play.
arod has somehow managed about 200 check swings in 30 or whatever postseason at bats. hes not seeing the ball well. hes swinging at balls out of the zone, taking balls in the zone, and swinging and missing at almost everything. most of the same can be applied for swisher.
we can all agree, given a large enough sample, arod and swisher would eventually hit better than they have this postseason. i do think its arrogant to think theres nothing more at play here than “small sample size”
Comment by Sleight of Hand Pro — October 17, 2012 @ 4:56 pm
the problem is dave assumes this was a decision on emotion with only quoting girardi’s “felt really good about ibanez line” as the evidence. im not going to blindly believe its an emotional decision cuz of something he says on record to the media.
Comment by Sleight of Hand Pro — October 17, 2012 @ 5:00 pm
Its also interesting that no one gives credit to Girardi for the Nunez move – generating their only run of the game. Everyone on here before the game hated that move, but you were wrong.
Comment by Noodle Arm — October 17, 2012 @ 5:06 pm
The coin example is actually pretty easy. Take a 50/50 coin and flip it 20 times. There is a 2% chance that a fair coin would come up heads 15-or-more times. So if you get 15+ heads, you should probably put that coin in timeout.
Trying to extrapolate coin-flipping to baseball gets iffy since both assumptions of the binomial distribution are in question (independence and identicalness). But the results from these assumptions just relax the requirements, if anything.
If we give Ibanez’s overall numbers the benefit of the doubt, that he would do good things against a lefty about 25% of the time, then he would have to outperform that figure by a fair amount to change our minds. If we assume independence between “clutch” opportunities, and that all opportunities are about the same, then Ibanez would need the following minimums to sway our minds (at the 2.5% level).
10 chances with at least 6 successful plate appearances.
20 chances with at least 10 successful plate appearances.
30 chances with at least 13 successful plate appearances.
Again, these minimums should only go up if we don’t allow ourselves the assumptions to use the binomial distribution. So these really are minimums. I don’t Ibanez has performed that well in the clutch…
This comment is exactly the tragedy Dave is trying to highlight. The run that Nunez generated is what happened most recently, so it’s bright in your mind. But the results of one or two or six games shouldn’t change what the correct decision was.
Saying “look and see what happened in one game” does not prove that a decision was correct.
Comment by Uh Oh Cordero — October 17, 2012 @ 5:33 pm
The right decision is emotionless, based on data that can be relied upon. It should no sense to come to any other conclusion.
The problem with baseball is they are not dealing with 50/50 decisions. They are dealing with highly complex decisions that historically still only “work out” roughly 30% of the time. The difference between 25% and 35% is the HoF or AAA.
This is one of those “I’m so smart” articles that drive people crazy. Of course letting Ibanez hit in that situation was the right thing to do. It wouldn’t have been the right thing to do prior to the post season or prior to the last few games, but it was the right thing to do in that situation.
To think that Arod would have had a better chance of getting a hit against Benoit is absolutely insane (or even against Coke, frankly). Arod couldn’t get a hit against me right now if the stands were full.
To seriously believe that Swisher or Arod were preferable in that situation, you would have to believe that at bats are actually independent from the previous at bats. To believe that at bats are actually independent you would have to believe that baseball players are not human. To believe that baseball players are not human, you would have to be autistic (or, perhaps, not understand statistics well enough to know that statistical models are simplifications of reality that often ignore difficult details like non-independence, player emotions, etc.).
Stopped reading after you said that you could get A-Rod out.
Comment by Rowboat Ralphus — October 17, 2012 @ 5:48 pm
Here, let me try:
Right now the NLCS Game 3 is on. Carlos Beltran was in the starting lineup, Matt Carpenter was not. Beltran went 0-1 with a GIDP. unfortunately for him, he’s hurt to some degree and will not play any more today. So in comes Matt Carpenter to take his place on the field and in the lineup. He hits a 2-run HR. OBVIOUSLY Carpenter should have started the game over Beltran!??
“To seriously believe that Swisher or Arod were preferable in that situation, you would have to believe that at bats are actually independent from the previous at bats.”
In fact, there are many, many studies about the role of psychology and personal momentum in sports. As far as I am aware, every single one has come to the same conclusion: events in sports are largely independent.
Dave was even kind enough to provide you with a link citing such evidence.
Agree 100%. There’s a growing arrogance in the SABR community in general that folks have access to enough information where they can assess smart/stupid, good/bad on strategic decisions, trades, signings, lineup orders, etc.
This entire article is based on public information and quotes from a press conference. Call me crazy but managers tend to not let on about everything in a press conference and will often offer platitudes and generalizations. Like Leyland saying “Valverde’s still my closer” after game 1. I guess you could assume the manager is being 100% honest and that he’s truly driven by these generalizations, but I guess I’m folish enough to think we may not have been given the whole story on why he made the decision.
We have no idea if ARod is 100% – some proprietary info like batted ball speed, bat speed, performance against pitch type may be factoring in to some of these decisions. What was the players health that day? Was Swisher given the day off just to give him the day off? Did he have a nagging injury that a manager may not let on about in a public press conference?
It is a bit arrogant to assume that it had to be recency that made this decision especially without full access to the info that may have fed into the decision or being able to talk frankly with the person that made the decision.
What was wrong with the decision to start Gardner? Great glove, LH bat and a solid history against Verlander: 5 H, 3 BB, 2 SH in 16 PA before last night. He also registered two of the toughest AB’s of the night. Not pinch hitting for Gardner late in the game is tough to defend, but starting him? I didn;t think that was a bad decision.
I’m sorry, but this is not a good example of what Fangraphs endeavors to accomplish. The essence of Fangraphs is not to lecture on the logical fallacy, operant conditioning, or whatever. Fangraphs distills quantifiable data into a form that deciphers truths about baseball. There have been something like 1000 words presented on Jerardi’s decision in the last 24 hours on this site. Alarmingly few have dealt with raw data. We know what the L/R split data is for the principle players involved. We know the degree to which PHs receive a punishment. In concert, these data points should inform our view on whether Jerardi made the correct move, and if he didn’t, the degree to which he was wrong. Where is this data presented?
For Game 162, Cameron made the claim that removing Griffin/Dempster after their first pass of the hitters would be advantageous because of the expanded bullpens and relief pitchers BAA. Whether you agreed him or the data, he laid it out and you couldn’t help but to see his logic. Here, I see him arguing against some radio call-in show premise, designed to distract half-wits, not get to the bottom of an interesting, reasonably quantifiable question.
Comment by Freud-by-Numbers — October 17, 2012 @ 5:59 pm
I agree to a point, but if we just always assume that the team has some inside information that makes every move they make the right one, there there really isn’t much to talk about, is there?
I think that with the publicly available information, Dave’s point stands. I’m sure he could be swayed from his stance if new information were to come out though.
the entire article revolves around the premise that girardi let his emotions get the best of him. thats a bold accusation, and to infer it based on the 3 words quoted in the article seems irresponsible to me.
i personally believe girardi thought ibanez gave them the best chance to not make an out. i believe he reached this conclusion based on the swings arod is currently taking, the at bats hes currently having, the fact that coke is probably the inferior pitcher and had thrown a lot of high stress pitches in the inning, etc etc….
of course none of us know the real reason girardi kept ibanez in, but to me, to say it was based on irrational emotion with little to no evidence…. is just wrong.
Comment by Sleight of Hand Pro — October 17, 2012 @ 6:18 pm
Players are not static. Some improve, some get worse. Statisticians would wait til a player’s struggles have been going on for a statistically significant amount of time before saying “he’s gotten worse”, when in reality, he had become worse much before it became statistically significant. Perhaps Girardi had seen signs in A-Rod/Swisher that let him know something was wrong, while still in the early stages of a not yet statistically significant decline.
I’m not saying we should assume it is the case and therefore every decision is correct and can’t be challenged; I’m saying we shouldn’t assume that we have the complete story.
Perhaps some of the limitations of evaluating the decision and the actual ability to conclude “recency” should at least be considered and/or mentioned?
Instead it becomes another arrogant “people are stupid and here’s why…” article. These type of articles seem to be increasing in frequency on this site. It is a very one sided and borderline arrogant article – I’m guessing if Joe Maddon made this decision the article would be quite a bit different and there would at least be mention of some of the things we might not know.
Not to mention, this probably would only change the outcome of the game like 3% of the time. The expected wOBA difference between Ibanez and Swisher/A-Rod is probably .050 or so, and even if they come through it doesn’t guarantee a win. A tied game would still leave the Tigers as favorites, and the Tigers would still have a chance to overcome a deficit if they fell behind.
Another glaring issue with a post hoc justification of the decision is that you don’t know what the other options might have resulted in. This is especially the case here where the Yankees lost, not to mention the fact that Nunez hit a home run and didn’t make any glaring defensive errors, but didn’t have the type of game where you could say that it is extraordinarily unlikely that anyone could have done better.
I like this article. I can appreciate that it is very, very difficult to ignore recent observations and go with what the data say.
Say you’re Joe Girardi. You’re the manager of the NY Yankees. You know A LOT about baseball and your paycheck and everybody around you confirms this fact everyday.
You watched and worked personally with these players every day for the last 7 months. How can you not believe that you are supremely qualified and capable of noticing subtle cues about players that give you unique perspectives on probabilities? Your job is based on the idea that you know things other people don’t know — even if those things you think you know are inconsequential or fleeting or not even real at all.
Like Dave said, this an ordinary, yet extremely powerful human tendency to believe what is right in front of you. It’s very very hard to overcome it.
I think even the most rigorous and unsentimental statistical analyst would have a hard time fighting off the urge to do what Girardi has done if the analyst were in his shoes.
At the same time logic and emotion, based on Raul’s recent success, would suggest he wouldn’t come through again, as we all knew that the odds of repeating an improbable event, another game-tying HR in the ninth inning, was highly unlikely. But there is also the desire to see an improbable event happen, and that was the opportuntiy to have it happen. Swing away, Rauuul!
Further, because we are dealing with actual humans, there is no reason to believe that the dependence between at bats is constant. In fact, I guarantee it is not (it is likely to be stronger when facing the same pitcher consecutively for example).
There are all sorts of reasons to believe that Arod’s recent performance is a better predictor of his current probability of success. First of all, there is no evidence he is physically healthy since he hasn’t performed well since returning from a significant injury. The only at bats relevant to Arod’s physical health are recent ones. You don’t even have to accept that player’s have brains to recognize autocorrelation between at bats due to changes in physical health over the course of a season or career.
Comment by Freud-by-Numbers — October 17, 2012 @ 7:40 pm
That’s so retarded on so many levels. What is relevant to Girardi’s decision is not whether the effect exists at all, it’s whether it’s enough to overcome the massive paper advantage for A-Rod. You can scream until you’re blue in the face that the effect exists, it’s real, blah blah blah, but if it’s real and worth .002 wOBA, nobody (but you, evidently) gives a crap because on the rare occasions it would change a decision, the value of that change would be minimal anyway. A “being in a slump” effect on the order of .040 wOBA to the next PA would stand out like a sore thumb over all of the slumps in MLB, but it’s just not there.
Of course, Calvin, the effect size is what is important. The problem is that you have no idea what the effect size is. You just made up that it is too small to matter. Joe Girardi, seemed to think the fact that Arod has had trouble even making contact was meaningful. Maybe Girardi is just retarded on many levels…..
What was bad about the decision to start Gardner? That you’re basing it on 16 PAs.
Verlander is easy to steal on and I think that Girardi felt, if Gardner reaches, he could steal a base and make something happen. He’s also, as you said, a good defensive OF. It wasn’t a horrible decision and is eminently defensible but not because of 16 PAs.
Credit for right or wrong moves should be based on the available knowledge at the time, not how they turn out. Dave points this out in his post. Managers can make the correct moves all the time and they can not work out and sometimes they make the wrong decision and a player hits a homer. Baseball’s funny that way.
I have no dog in this fight. I don’t know now and I didn’t know then whether or not it was the right decision to play Nunez, though I have no faith whatsoever in Jayson Nix. But the decision to start Nunez is right or wrong based on what was known when the decision was made, not based on what happened during the game, since Girardi had no idea that Nunez would hit the homer.
No, I didn’t make up that it was too small to matter. I stated that research into slumps had shown that whatever the effect of being in a slump could be, if anything, was far too small to override the paper difference between ARod or Swisher compared to Ibanez.
And of course managers make completely idiotic decisions out of ignorance or stupidity (or both) all the time. Just look at tonight’s game where Bochy sac-bunted Cain on every pitch with 1st and 3rd, 1 out, down 1 in the 6th. That’s a horrible blunder that anybody who can handle +-*/ can show is terrible. You score more runs and you’re 1.5-2x as likely to score at least one if you just let the pitcher swing away.
The study pertinent to baseball that I recall looked at bins of as few as 10 PAs and as many as 50 PAs, I believe. It aimed to find autocorrelation from one bin to the next, and the results were not statistically significant. I am well aware of the issue of type-II errors, and without actually running the test myself, I cannot estimate those for various effect sizes. Conventional hypothesis testing definitely puts a lot of emphasis on type-I errors, but too often ignores type-II error rates.
As for success against a given pitcher, it’s easy to show that as the game wears on, the average batter gains the advantage over the average pitcher (relatively, at least). So there is dependence there, but the question becomes “does it exist game-to-game?” I don’t know, but the overarching evidence from all sports suggests that there is no effect size large enough to be statisticized.
As for A-Rod’s injury, and its potential effect on his hitting, that should really be the main point of your argument here. Dividing up his PAs over his DL stint is hardly arbitrary, so I did it for our benefit:
His OBP, Slugging, ISO, K% and BB% all got worse.
As for plate discipline, most stats were pretty close, but the glaring difference was his in-the-zone contact rate which dropped about 8% (80 to 72 ish). So the injury and the statistical evidence add up to say he’s probably worse.
Injured A-rod had a triple slash of .261/.341/.369, but unfortunately I can’t double-split him over Sept/Oct AND against righties. We can assume he was doing worse against righties, though.
Ibanez had a season slash of .240/.308/.453, but slashed .197/.246/.246 against lefties.
Ibanez was so terrible against lefties, and his advantage in power over the injured A-rod does not exist. Ibanez sucks against lefties. i’m not sure the “heating up” effect size from a dinger, a grounder, two strikeouts and a popup is enough to bring Raul back up to A-rod injured level.
No, it’s just a lot of one-sided verbiage to support a personal opinion. Here are four reasons why:
1. I’ve read the original hot-hand paper, and the methodology is flawed, failing to take into account game situations, specifically how a defense may react to a player’s recent success. Congratulations to Dr. Reifman if he’s managed to build a career around it.
2. We don’t know why Joe Girardi made his decision. He has information we don’t.
3. A-Rod and Swisher hadn’t just had a lack of recent success; they’d also lacked a decent approach. A-Rod in particular (okay, and Granderson) seem to be having 3-pitch at-bats where they 1) take a strike, 2) swing at a pitch out of the zone, and 3) strike out.
4. Another recent long article on this site shared your viewpoint, but the statistical breakdown made it look like an extremely close call, even purely mathematically. That’s if you don’t believe that streaks and slumps exist – which is a ridiculous point of view, if only because you have to admit that a player could be injured and hitting poorly for exactly that reason.
You are misunderstanding the meaning of the research that you cite. According to you they failed to detect any effect using their methods. This does not mean there is no effect, and it certainly does not mean that if there is an effect it’s importance is negligible.
It may be the case that ON AVERAGE the effect size of “slumps” is very small or not statistically detectable. Perhaps the vast majority of players do not suffer from slumps at all. Perhaps, though some do. Perhaps some do to huge effect. Perhaps, also, some players don’t suffer from slumps for the vast majority of the careers, until they do.
Look, we all know this happens to players. Anyone that has ever watched Chuck Knoblauch, Mackey Sasser, Chase Utley, Steve Sax, Donrelle Willis, Mark Wohlers, Steve Blass, Rick Ankiel, etc. throw a ball knows that psychological issues can pop up from time to time and vastly alter the performance of the player. Chuck Knoblauch never threw a ball five feet wide of Tino Martinez during practice. It was the situation that effected him.
I watch just about every Yankees game so i got to see Arod quite a bit. The more he has failed this postseason, the less he has looked like even the hitter he was a few weeks ago. He is currently struggling to even make contact. He is swinging through fastballs down the middle of the plate right now. Joe Torre certainly thought Arod had a psychological problem. Torre thought that when Arod was confident he allowed his talent to work, but when he started pressing and when the pressure mounted he was not able to perform.
I can’t understand how people can seriously doubt that psychological effects can have large effects on player performance on rare occasions?
Perhaps it was in his mind that all options were equally poor and he would rather have Ibanez take the fall than have Rodriquez or Swisher wear another failure.
Comment by james wilson — October 18, 2012 @ 12:58 am
To keep his job, he can’t make decisions that people will second guess as much. Regardless of who is batting, chances are they won’t get a hit in any situation because most people don’t at over .500. Batting ibanez is what everyone wanted to see. In all likelihood, a-rod, swisher, and ibanez would all have failed. Ibanez is one that’s easier for the average fan to cope with because they believe in mystique and destiny more than the analytical types that make up the minority.
Comment by Antonio bananas — October 18, 2012 @ 1:16 am
And if you carry that line of idiocy far enough, you wind up arguing that it’s impossible for an outsider to conclude that putting 8 pitchers in the field and Granderson on the mound is a bad idea. I mean Swisher might have just turned and smashed Girardi’s skull in with his bat if he’d been asked to PH right there, so, well, I guess it’s possible that Girardi not PHing with Swisher there was the smartest thing he’d ever done in his life and we have absolutely no business concluding otherwise. Hey, you never know, right?
I believe that a strictly statistical of this situation would reveal that, as some wise country hardball guy once said, you should dance with who brung you.
Comment by frugalscott — October 18, 2012 @ 1:30 am
And didn’t replace him on the roster just so he could keep chasing front-row tail during the games? Oh, wait, DL’d players can still be in the dugout, so… he’s still on the roster why again if that’s true?
The general misconception here is that stats, advanced stats, can give a manager enough information, and the right information, to set a lineup, decide what batters should face certain pitchers et al.
The problem with that reasoning is that it does not take into account that how the performance of a player varies during a season, or a career.
The point is, if a player who’s previously batted .314 .422 .645, runs into significant mechanical problems with his swing or quite simply loses his timing, or messes up his routine at the plate, this player could become a below-average player.
Likewise, if a player who’s historically bad versus left-handers, is showing in batting practice that his timing, plate approach and what not, is better than that of the aforementioned “superstar”, a manager should obviously pick the guy who looks to be most likely to hit against Verlander.
It matter to some extent that A-Rod had some success against Verlander this year, and his season stats against a lefty matter somewhat, but a manager should only use the statistics as some kind of guidance. He also needs to make a call based on what form the players are in today.
Stat guys are looking at the data, but I am assuming many of them never did any sports themselves, so they are not privy to the concept of athletes quite simply being better during certain spurts, reasons perhaps being new strength program, different diet or simply a better general wellbeing.
Factors like these are difficult to value, since they are typically to some extent abstract, and relative.
Instead, stat guys tend to scream “streak!” when a player gets hot. Instead of trying to look at the underlying factors to said athletes recent success or failure at the plate.
Look at Rory McIlroy, or Usain Bolt, or Lionel Messi. Those athletes don’t perform on a constant level. If Usain Bolt runs a 200 meter Olympic final, you’re not gonna argue that he’ll run a 19.5 if he looked slow and awful in the quarter and semis one day earlier.
And what do you base that on? Not his stats, but on his apparance that day, on that track.
What we can say is that the season stats pointed towards batting Swisher and A-rod.
But those stats can only be a general guide for a manager. He manages real players, not numbers. So he has to check how Swisher, Arod, Ibanez and Gardner looks day to day.
We may not agree with his decision, but the stat line, and Arods success during this summer against Verlander is only one of many valid arguments for and against batting him.
Some guys needs to realize that perhaps Joe Girardi just took a look at Arod, and with his knowledge of baseball, perhaps even discussing with a hitting coach, decided that Arod today is worse against lefties than Ibanez.
Comment by Smallie Biggs — October 18, 2012 @ 2:06 am
Stats tell you what happened. They don’t tell you why and how they happened.
Right now, Eric Chavez and Raul Ibanez are having better AB’s than Curtis Granderson, Nick Swisher, and especially A-Rod.
That’s a big factor in baseball. The stat sheet could say a batter hits this pitcher great. Except that batter’s mechanics and confidence are just shot. Another batter doesn’t hit that pitcher so great, but right now, he’s locked in.
There are so many things we’re not privy to. For example, how do we know A-Rod’s hand is OK. He broke his hand in late July. Maybe his hand just never fully healed, and is affecting his swing.
Right now, Raul Ibanez at least gives you a professional AB, and A-Rod doesn’t.
Comment by waynetolleson — October 18, 2012 @ 2:19 am
This isn’t nearly as clear cut as you’re making it out to be.
You have a choice between Ibanez facing Phil Coke, who had a 1.63 WHIP this season and had just given up a pair of hits, or Nick Swisher pinch hitting against Octavio Dotel and his 1.07 WHIP. Swisher has put up a .120/.214/.240 slash line in 28 plate appearances as a pinch hitter over his career, and a .167/.284/.300 line in 177 postseason PAs.
Also, even though Ibanez has struggled in 61 ABs against lefties this season, he’s managed to put up a respectable .262/.315/.420 against them over his career.
All this is before you start talking about slumps, or clutchness, or players psychology, or any other intangibles.
Yes, Calvin, that is a wonderful analogy. Surely, the fantastic situations you dreamed up are exactly the same as common anxieties. Because, just like your fantasy world, MLB players never suffer from anxieties. ….just don’t ask Rick Ankiel why he’s an outfielder these days….
Perhaps it is just my idiocy – which is retarded on many levels – that makes me thinks baseball players are normal people with normal problems.
Joe Torre unequivocally stated that Arod let his head get the best of him. Joe Girardi seems to believe it too. …..but you know better, because you are without idiocy, and are not retarded on any level.
It’s pretty clear that the decision to sit A-rod is not entirely a baseball decision. It looks to me like they are done with him. I doubt that the decision was up to the manager. I think hitting on chicks in the late innings of October baseball was the last straw. While the decision to pinch hit him was defensible, and the decision to bench him slightly less so, based on the fact that he is clearly unable to get around on right handed fast balls, there is no explanation for not allowing him to at least swing the bat once in a game where one run could mean the series.
Comment by rockymountainhigh — October 18, 2012 @ 9:59 am
Yes! You have a lineup full of lawn jockeys and two beer league softball players at the corners. I don’t know why he didn’t put some balls on the ground and try to scratch out a run. The tv guy (i think Smoltz) pointed out that he should have had Ichiro steal second when they put the shift on against Texeira. If they misplay the throw he could be standing at third.
Comment by rockymountainhigh — October 18, 2012 @ 10:03 am
Arod clearly cannot hit a right-handed fastball. His bat is not fast enough currently. The day after Ibanez pinch hit for him against Baltimore, he swung through two straight 85 mph fastballs down the middle.
Comment by rockymountainhigh — October 18, 2012 @ 10:07 am
“team Fangraphs(aka Team WAR) wouldn’t be sitting at home empty handed. Again.”
I like your comment except for labelling decisions “right,” “wrong” or “correct.”
Baseball decisions are not any of those things, but simply have better or worse probable outcomes.
Eventually the manager or player who goes with the higher probabilities will be more successful than if he had gone with the lesser, but that certainly doesn’t mean there was one decision that was the right or correct one in a certain situation.
Uh… how many at bats have the Yankees pinch-hit options in question had against Coke (or whatever Tigers reliever)?
Comment by The Real Neal — October 18, 2012 @ 2:11 pm
“Also, even though Ibanez has struggled in 61 ABs against lefties this season, he’s managed to put up a respectable .262/.315/.420 against them over his career.”
But recent performance doesn’t matter..oh, wait, you are saying that Dave C was USING recent performance to tell us that recent performance doesnt’ matter. That’s not very smart.
Comment by The Real Neal — October 18, 2012 @ 2:28 pm
I totally disagree with this.
Everything that Girardi and Cashman has stated has pointed towards it being a pure baseball decision.
And we’re to somehow second guess their statements, “they say one thing but they really mean something else”. I am very reluctant to do that. But facts suggests that Girardi isn’t in “spin mode” either.
Because Yankees number one priority is of course to win. It is for every team.
If they for some reason were more interested in getting rid of Arod, then it’s not exactly a good idea to bench him either. That only brings his trade value down.
Unless you’re able to provide any type of proof of the benching being something more than baseball related, share. Otherwise, it’s just pure speculation, as plausible as a speculation that Joe Girardi will run for NY mayor (guessing without any substance).
I don’t mind people slamming managers and GM’s for bad decisions, not at all, but there is this tendency that the critisism is based on pure guessing, or in David Camerons original article above, doesn’t take into account relevant factors.
David Cameron’s article about Girardi is very well written, and I enjoyed it, but it is, unfortunately, based on false premises. Cameron, it seems, lacks a fundamental understanding of athletic performances. Cameron is one of my favorite writers in baseball, but here he just seems to have lost himself in everything Sabremetics.
A manager, like Girardi, is going to make decisions about players and their current form. When doing that, advanced stats are a very helpful tool, a baseline if you will. But you have to factor in where a player currently is.
Maybe in a few years teams will have developed advanced formulas where they have things like OBP value against lefties and multiply it with current physical/swing mechanical/mental estimated values (much like you rate arm and bat power in a prospect) and get a “current” OBP (eg).
Such a formula could’ve shown that Ibanez just narrowly was a better option than Arod, or the other way around. This is just pure speculation, but baseball seems to move toward more and more analysis.
I do not want to be offensive, and I sincerly hope Cameron see this as somewhat of an attempt at constructive critisism. I am very grateful for Fangraphs and its content. I do not take it for granted.
Comment by Smallie Biggs — October 18, 2012 @ 5:04 pm
How would you apply this kind of reasoning to Jim Leyland keeping Valverde as his closer?
Stats and the current day form doesn’t have to differ.
Valverde looks terrible, perhaps even worse than his “seasonal stats” suggests.
Leyland should just use the relief pitcher who is performing best at the moment.
I am not extremely familiar with the Detroit bullpen, but Dotel looks to be throwing ok, I may be wrong about him being the optimal choice though.
A better example is Phillies and how Manuel kept using Lidge, despite Ryan Madson showing much better form. This was a few years ago. Manuel wouldn’t go with Madson until Lidge was clearly much worse.
Which, obviously, is a mistake.
But the decision on what pitcher to use, is not only gonna have to be based on numbers, it’s also gonna be based on current form. And Valverde, IMO, is in really bad form right now.
Leyland still was quite rational when he stated that he had to discuss Valverde’s ongoing status in the postseason with the pitching coach.
To stat guys this may not seem relevant, and something quite unnecessary. But for someone who’s paid millions to manage a MLB team, they need to evaluate a situation as proper they can.
What would you say if, theoretically, Valverde would come out next time and close like he did 3-4 years ago. And that Leyland said that they discovered a technical flaw that they rectified, and that brought Valverdes velocity up significantly?
A very unlikely scenario, and one I don’t believe in. I actually believe Valverde’s just a guy who doesn’t keep himself fit and now that he’s getting up in age, that hurts him.
But it something that Leyland cannot rule out without discussing internally.
Still, to me as an outsider, both stats and current form points towards sending Valverde packing.
Comment by Smallie Biggs — October 18, 2012 @ 6:30 pm
Unfortunately, he looks totally lost against lefties too.
Against Smyly he swung on pitches that wasn’t even near the strike zone. And Smyly, with a 92 mph fastball, looked like Aroldis Chapman pitching against Arod. He just blew those pitches by him. It was terrible.
Comment by Smallie Biggs — October 18, 2012 @ 6:37 pm
I agree current form needs to play a role in the decision making; however, Valverde’s season stats are not terrible.
I think the general consensus among the “Next Gen” type of fans, such as most of us, is for managers to be much more “active” in managing the bullpen.
Matheny has us all salivating from going with Motte, who I’m not even sure IS the best reliever in that very strong bullpen, but still, to go with him from the 8th, when it mattered the most.
That’s “situational” managing.
In other sports, in soccer’s Champions league eg, managers will manipulate a lot with the starters on a daily basis.
I think that’s something that needs to be implemented in baseball as well. If someone like Nunez is the only one really hitting in Yankees lineup, Girardi needs to put him in the 2nd spot.
And that would be based purely on current form.
Then, there’s always the factor of players psyche. Bobby V tried to get everyone on the roster starting in April, it ruffled the feathers of Youk and his pals in the Red Sox club house…
We like to make it out to be very easy. But it’s, of course, not.
But I think the current trend of more “creative” managers is making the game so much better. My own team the Phillies just happen to have the most clueless skipper right now, as well as a megalomaniac for a GM…
Comment by Smallie Biggs — October 18, 2012 @ 8:04 pm
Because he can hit lefties, play third, and is a better option off the bench than their 26th guy?
But really, my only point is that while letting Ibanez bat against Coke is a questionable decision, it’s absurd for any of us to presume to know everything that factored into the Yankees’ reasoning.
I understand that these studies are often based on many players so as to obtain a necessary sample size. At that point, players are often lumped together and described as a whole using the mean.
First off, most statistics are mean reverting. Outliers are guaranteed by randomness, and thus are just as likely to fall back to earth for any mean reverting statistic. If you don’t believe the individuals in baseball are mean reverting in terms of their auto correlation, then…
Second off, I took Alex Rodriguez’s entire career, split it up into 10-game segments, calculated the batting average, OBP and ISO for each 10-game segment, and checked the auto correlation charts (and probably more importantly, the partial auto correlation charts).
If you can believe it, the first lagged partial correlation was actually negative for batting average. For OBP and ISO, they checked in at less that 0.05.
I did the first lag by hand, checking each 10-game segment correlation with the next 10-game segment. There were 123 pairs of segments, each independent under the null hypothesis, and the highest correlation occurred in ISO. It was 0.16, which is not statistically significant, but more importantly, not even ACTUALLY significant. That’s an R-squared of 0.0256.
Finally, I looked at the segments in which he had performed especially badly, measured as segments a full standard deviation below the mean for the 14 surrounding 10-game segments (14 + 1 = 15 x 10-game segments = 150 games = 1 full season). The results:
In 36 10-game segments, A-rod performed badly in batting average, relative to the nearest segments. The following 36 segments showed an average Z-score of -0.012. In other words, he rebounded to his average average.
In 37 10-game segments, A-rod performed badly in OBP, relative to the nearest segments. The following 37 segments showed an average Z-score of 0.127. In other words, he rebounded to do slightly better than his average OBP.
In 33 10-game segments, A-rod performed badly in ISO relative to the nearest segments. The following 33 segments showed an average Z-score of 0.309. In other words, he performed significantly better in the power department after suffering a 10-game slump (p-value = 0.055).
A-rod has shown no history of allowing 10-game slumps to compound.
Agree with you 100% Sleight of Hand… I’ve begun to take this approach with SABR’s/”Old School Baseball” like I approach fox news/msnbc. You know both are going to be extremely biased, telling you everything you already know without any objective opinion. SABR’s think they know all, that stats rule all, well you don’t and that is why you’re not a manager of a professional baseball team.
If you guys were so successful you wouldn’t be only running the numbers behind trades, you would be in the trenches everyday managing a game. Just because STATS tell you that A-Rod will eventually “get it right” does not make it so. They’re so many more variables “mechanical flaws, MENTAL makeup (oh I’m sorry according to STATS, baseball players are robots with no ability to feel emotion, therefore there is no such thing as clutch hitting)
You guys need to take a good hard look at yourselves, because articles like this are what keep me from coming from this website more often, and I suspect other people agree with me as well. Find some humility yourselves and if you can’t hire someone who can pull this website out of the biased fox/msnbc views you adhere to.
Comment by pear1jamten — October 19, 2012 @ 8:16 am
I’d also like to note you don’t include the fact that
A. He was injured this year and that could be a factor and information that Girardi has that we don’t.
B. He has been injured many times over the last few years and his body could be breaking down because of that and AGE.
C. Age.. Just because his stats indicate he will eventually “even out” decline is a major factor.
D. Steroids- He’s off his juice and who knows if in the first half he was receiving some sort of supplement that they now screen for, they’re so many variables that you fail to account for, quite frankly, it’s embarrassing.
Comment by pear1jamten — October 19, 2012 @ 8:53 am