FanGraphs Baseball


RSS feed for comments on this post.

  1. Did you just conclude that there’s a 6.1% chance of winning if you don’t advance Verlander, but a 6.8% chance if you do?

    And it’s logical to take that ~10% haircut?

    I mean, if you are going to make up numbers, at least make up good ones :)

    Comment by DNL — October 11, 2011 @ 10:21 am

  2. Where did you get those ‘odds’ from?

    Comment by Mike H — October 11, 2011 @ 10:28 am

  3. It would have been interesting to actually calculate the Tigers’ chance of winning the next few games based on matchups on full rest, and then actually estimate the effect of short rest using some sort of historical precedent, and thus actually draw a real conclusion as to whether or not Leyland is right.

    Comment by Mike H — October 11, 2011 @ 10:33 am

  4. This.

    I’d agree with Leyland if Verlander could only pitch one more game this series. It wouldn’t matter which game it was. But if by moving him up, you can pitch him twice this series, that should be a no-brainer.

    Comment by Yirmiyahu — October 11, 2011 @ 10:33 am

  5. Guys, the odds aren’t real, but meant to illustrate the point. Dave is trying to conceptually explain Leyland’s decision, and while it’s not super easy to follow, it does a good job.

    Comment by Telo — October 11, 2011 @ 10:34 am

  6. I have to question your 5% discount for moving the better starters up one game. Is there a study to support any discount, let alone 5%?

    Comment by GiantHusker — October 11, 2011 @ 10:36 am

  7. Nevermind. I actually looked at the schedule. Not only would you need Verlander/Scherzer/Fister to all go on 3 days rest (there would definitely be a performance hit there), but for Game 7, you’d be pitching Verlander on 3 days rest yet again. Injury-wise and performance-wise, that sounds like a terrible idea.

    Comment by Yirmiyahu — October 11, 2011 @ 10:41 am

  8. I don’t even know what the first comment means.

    And, I’m sorry, but it’s clearly not a “no-brainer”. The upgrade in one start from Porcello to Verlander comes at the expense of reduced performance from the starters of the other three games. You can’t just ignore that.

    Comment by Dave Cameron — October 11, 2011 @ 10:41 am

  9. In these situations when you follow the lines of logic and guessed percentages and attempts at quantifying the situation, you are always left at the exact same point….

    Is it worth to start your ace on short rest when:

    A. he could get injured


    B. he could be less effective

    (C. additional factor of reduced innings pitched during the following series, if you advance)

    The truth is, no one ever knows the true value of A or B (C you can decently quantify), even if the player has some history pitching on short rest. How you feel about the entire situation comes down to A and B, and whether they are big enough factors to prevent you from pushing your ace in the playoffs.

    I concur with Leyland and Dave.

    Comment by Telo — October 11, 2011 @ 10:42 am

  10. Cameron says Leyland is “exactly right” and then supports that evaluation with some sloppy reasoning and made-up numbers.
    He did not say that he was explaining Leyland’s reasoning.
    Camerson clearly did not make a case for Leyland being “exactly right” or even approximately right.
    I’m not saying such a case cannot be made, but I am saying that Cameron did not make it (nor even come close).

    Comment by GiantHusker — October 11, 2011 @ 10:43 am

  11. There’s been a lot of research on short-rest starts. Here’s one from last year.

    There’s no question that starters who pitch on short rest do worse than those who go on full rest.

    Comment by Dave Cameron — October 11, 2011 @ 10:44 am

  12. I’ve never seen a study on pitching on 3 days rest. I’d love to see one. I’d also love to see a study on how effective regular-season starters are when pressed into postseason relief without a history of working out of the ‘pen or a chance to acclimate to the new role.

    But it’s pretty logical (and accepted by every coach in the game) that a tired and sore pitcher isn’t going to be as effective.

    Comment by Yirmiyahu — October 11, 2011 @ 10:44 am

  13. in the book, tango/mgl (i forget who did that chapter) show that pitchers perform best on 5 days of rest, with slightly lower performance for each fewer day.

    they conclude that 4 days of rest (ie 5 man rotation) is optimal in terms of maximizing overall performance (balancing performance of each pitcher with the face that adding an extra day of rest adds an inferior pitcher to the rotation)

    Comment by Jono411 — October 11, 2011 @ 10:46 am

  14. That’s what I did. If you want to quibble with the numbers, go ahead, but the reality is that the conclusion is going to be the same. You have to drastically overestimate the difference between Verlander and Porcello or underestimate the effects of pitching on short rest to think that there’s a huge advantage to switching to a three man rotation.

    Comment by Dave Cameron — October 11, 2011 @ 10:46 am

  15. Leyland nails it when he says that the Tigers have to win four games.

    He’s a crafty one.


    I’m not a big fan of just riding Verlander like a rented mule, and I’m glad to see leyland take the heat off of JV.

    But, it’s hard not to like Verlander replacing Porcello in the remaining starts. Verlander-Scherzer-Fister-Verlander doesn’t sound too bad, even down 2-0. Each one of those guys has the stuff to hold down TEX through 6 IP. I don’t think Porcello can do it. He’s not enough a K or movement type pitcher.

    One thing I do like about Leyland speaking so declaratively about DET not piggy-backing Verlander all the way is that it sends a message to rest of the pitchers, [1] It’s on you, and [2] I’m confident you’ll get the job done.

    It’s V-E-R-Y tempting to have Verlander replace Porcello in the playoff rotation. But then what do you do in the next series? The point of the 7-game series is that the 4th starter will need to start 1-game.

    Comment by CircleChange11 — October 11, 2011 @ 10:48 am

  16. Thanks for that, Dave. Never seen those numbers before. Though there’s a huge flaw in that study by lumping 1969-2009 all together. Pitching on 3 days rest used to be the norm.

    I’d like to see a study using only, say, 1994-2011, and using ERA+ or OPS+ instead of ERA. Would also be interesting to look at innings per start.

    Comment by Yirmiyahu — October 11, 2011 @ 10:49 am

  17. Listen, if I’m not criticizing Dave, there’s a good chance it’s not worth criticizing. It’s not a perfect reflection of reality, but assuming the numbers are close, they represent the situation well.

    Comment by Telo — October 11, 2011 @ 10:54 am

  18. The problem with Leyland is that he is not open to other ideas. He is very stubborn and not very bright; he doesn’t embrace sabermetrics and is content managing as he always has. I don’t have a problem with him starting Verlander on normal rest but I do have an issue with him not even seriously considering alternatives.

    Comment by well — October 11, 2011 @ 10:55 am

  19. An obvious move would be to replace Porcello early — no more than twice through the lineup, regardless of pitch count.

    Comment by Lex Logan — October 11, 2011 @ 10:56 am

  20. The most logical measure is to fire Gene Lamont last year. Alas,

    Comment by buddy — October 11, 2011 @ 11:06 am

  21. Of course I agree that pitching going back out there on short rest is generally a bad idea, but look at his last start. JV only thre 82 pitches. They showed a graphic that said he averaged something like 115 pitches an outing this year. That’s a massive difference. If Leyland were to start JV tonight, I would have no gripe with it. Turning him around and sending him up for Game 7 if he pitched today? Yeah, that would be a problem.

    Comment by wiersNRAF — October 11, 2011 @ 11:17 am

  22. I see what you’re saying, I just don’t get how you decided that Verlander had a 60% chance of winning on full rest and a 57% chance on short rest. How did you settle on those numbers? What if you’re wrong about a 3% chance penalty for short rest? The factors here are clear (ace on full rest once vs. possible ace on short rest twice, effect of momentum of not going down 3-0, etc.), but isn’t the point to try to test them with actual data to draw the conclusion rather than just guess based on intuition? Wouldn’t there be a way to look at winning percentages of pitchers on short rest vs. rest of their careers or something?

    Comment by Mike H — October 11, 2011 @ 11:17 am

  23. Thanks to the rain out, there’s four straight games (2-5), no off days. Dropping Porcello doesn’t get you Scherzer on 3 days rest.

    Comment by joseph — October 11, 2011 @ 11:25 am

  24. The logic in this article is solid. Sure, the numbers are swags but they all pass the sniff test.
    vr, Xei

    Comment by xeifrank — October 11, 2011 @ 11:30 am

  25. Just to use your percentages, what about this:

    Game Four w/Verlander: 57 percent
    Game Five w/Porcello: 50 percent
    Game Six w/Scherzer: 45 percent
    Game Seven w/Verlander: 53 percent

    Odds of winning all four games: 8.0%

    And additionally you have Fister available on regular rest in Game 7 to throw as many innings as needed.

    There are people who believe Verlander is a hoss and the percentage play says that the Tigers are most likely to win when Verlander is on the mound. Therefore, these people would rationally say that Verlander needs to be on the mound as much as possible, and all the other starters can fill in on normal rest when necessary.

    Comment by The Nicker — October 11, 2011 @ 11:35 am

  26. In Baseball Between The Numbers: Why Everything You Know About Baseball Is Wrong”, the chapter on this topic showed that pitchers in 4-man or 5-man rotations pretty much pitch at the smae performance level, with the conclusion being that teams should stop giving so many starts to their 5th, 6th, 7th best starting pitchers.

    So, we seem to have studies that tell us oppossing things, which is not the end of the world. We’d need to look at methodology, samples, selection, etc.

    Comment by CircleChange11 — October 11, 2011 @ 11:38 am

  27. lol, Xei and his swags again.

    Comment by Telo — October 11, 2011 @ 11:39 am

  28. How do you know he did not consider the alternatives. All you know is that he did not communicate his considerations to the media. He made a decision, and he only communicated the final decision to the media. To assume that there was no consideration may be faulty.

    It is possible that going back to last series he decided, for whatever reason(s) that Verlander won’t pitch on short rest for the remainder of the post-season.

    Comment by CircleChange11 — October 11, 2011 @ 11:42 am

  29. According to the article you linked to, Dave, the difference between 3 days rest and 4 days rest was considerable less than 1% based on ERA and 1.6% based on Game Score, a very dubuious statistic. 4 days rest was also better than 5-plus days rest.
    This study was inconclusive at best, probably worthless, and even it does not support your assumptions.

    Comment by GiantHusker — October 11, 2011 @ 11:42 am

  30. By what percentage? Cameron is claiming a 5% difference.

    Comment by GiantHusker — October 11, 2011 @ 11:44 am

  31. You are making a lot of assumptions.

    Leyland, his coaches, and the front office probably discuss and consider a lot of things we don’t and will never know about. And Leyland, in his own mind, probably weighs out all kinds of options.

    That said, it is probably unlikely he uses many “advanced” metrics. But the point is we don’t know either way, so basing your criticism this way is a little crazy and unfair.

    Comment by Nathan — October 11, 2011 @ 11:45 am

  32. Look, quibble abt numbers all you want. I’ll give u a 60% chance of every Tiger pitcher winning in each remaining game (4-7). Law of compounding probability still means only 12% chance of winning 4 straight.
    Author’s last line about not going down 3-0 is the most important point.

    Comment by OrigRangersFan — October 11, 2011 @ 11:45 am

  33. Who says he didn’t consider alternatives? Managers don’t bring the media into the strategy room to walk them through every step and scenario. I think it’s clear from Leyland’s comments that he has considered it and immediately ruled it out. As Dave explains above, Leyland’s chosen strategy makes sense.

    Comment by Dave — October 11, 2011 @ 11:48 am

  34. The problem is, the point being illustrated changes with the numbers.

    The analysis is worthless if the numbers are wrong.

    Right now, Switching the lineup increases your chances of winning by about 12%.

    But if move Verlander to 65-62 instead of 60-57, we’re up to 7.5%, which is 22% improvement.

    If you look at the actual results, IE, that the Tigers went 25-9, in games that Verlander started (73%) and the numbers get drastically higher. On the other hand, they went 21-10 in the games Porcello started, (67%), so his numbers should probably be raised.

    Essentially, the results depend completely on what numbers you make up.

    Comment by RC — October 11, 2011 @ 11:50 am

  35. It might not be a huge factor but I would have to think all of the uncertainty surrounding the rain delays, stops/starts, will I/won’t I pitch today has been an added peril for pencilling in future rotations.

    This article makes sense.

    Comment by tdotsports1 — October 11, 2011 @ 11:56 am

  36. I don’t want to rain on a sabermetric parade or anything, but we shouldn;t be shocked that a baseball lifer in his upper 60s doesn’t use advanced metrics.

    I bet he doesn’t text or tweet either. Probably doesn’t even own an iPad.

    However, it is possible for a manager to avoid advanced metrics and sabermetric research and still have the same conclusions that the research supports.

    There are been managers that would be considered “sabermetric friendly” before the invention of sabermetrics.

    Comment by CircleChange11 — October 11, 2011 @ 12:06 pm

  37. For what it’s worth, I *also* agree with Leyland.

    Comment by DNL — October 11, 2011 @ 12:21 pm


    Comment by Marc — October 11, 2011 @ 12:22 pm

  39. -i*hbar*gradient. duh

    Comment by Drew — October 11, 2011 @ 1:03 pm

  40. “The point of the 7-game series is that the 4th starter will need to start 1-game.”

    I agreed with a lot of what you said, then I read this nugget. Seriously?

    Comment by My echo and bunnymen — October 11, 2011 @ 1:21 pm

  41. Game 4 is tomorrow……

    Comment by My echo and bunnymen — October 11, 2011 @ 1:25 pm

  42. Pray for a rainout.

    Comment by Neuter Your Dogma — October 11, 2011 @ 1:30 pm

  43. I remember a few years ago with the Jays, Halladay pitched on short rest a number of times. He was quoted as saying he liked it – felt he had a better feel for pitching.
    I forget his results, and the year. Halladay’s not a good example because he’s awesome, but I always think of this when short rest gets brought up.

    Comment by lexomatic — October 11, 2011 @ 1:31 pm

  44. Whatever you think of Dave’s fictional “percentages,” his point stands – that whatever choose you make amounts to a coinflip. In other words, should the Tigers be eliminated in a worst case scenario, Leyland’s lynching will be totally unwarranted.

    Comment by minesweeper — October 11, 2011 @ 1:38 pm

  45. You’re point is clear but the conclusion isn’t real. While you might be correct that everyone is worse on 3 days rest it is not certain. In fact they could be better.

    However you do illustrate a scenerio.

    Comment by Joecooldps — October 11, 2011 @ 1:41 pm

  46. Ah, but every game they play in Detroit is another sell-out on the revenue side of the the Tigers’ ledger. If you’re going to lose the series anyway, you’d rather do it in game 6 or 7, once you’re back in Texas and have netted all the gates in Detroit you can. Losing game 4 or losing game 6 may work out the same in terms of (not) moving on, but game 5’s revenue hangs in the balance.

    Comment by joser — October 11, 2011 @ 2:12 pm

  47. “You have to drastically overestimate the difference between Verlander and Porcello or underestimate the effects of pitching on short rest to think that there’s a huge advantage to switching to a three man rotation.”

    What if the argument is not that there is a “huge” advantage to the 3-man rotation. What if the argument is only that there is likely a slight advantage. Aren’t slight advantages worth taking?

    Comment by bc — October 11, 2011 @ 2:32 pm

  48. Ha ha. I demand that Xei uses the word “swag” in every comment he posts from now on. Otherwise, I’m giving it a hundred -1.

    Comment by grape — October 11, 2011 @ 2:38 pm

  49. I some states sniffing someone’s swags is a felony.

    Comment by bc — October 11, 2011 @ 2:44 pm

  50. There’s a pretty clear reason the winning percentage numbers are higher. The easy reason – those are the Tigers numbers against all teams in baseball, not an elite playoff team like the Rangers. You can’t give Verlander a 73% chance of beating an excellent team, even if you trust the small, skewed by other factors samples of single season winning percentage.

    Comment by todmod — October 11, 2011 @ 3:03 pm

  51. There is no right answer here. I don’t think a study, no matter how far reaching, is going to help much. Every human body is different and every body responds differently to short rest. Its already been written that JV has never once pitched on short rest in his life. Al Leiter couldn’t get out of the 1st inning against the Braves in Game 6 of the 98′ NLCS. The Yankees went to a 3 man rotation and won the WS pitching all three guys on short rest in consecutive starts. Until we see JV actually start a game on short rest there’s no way of knowing how well he might perform. If Leyland had Sabathia and not JV, its a mortal lock that C.C would be bumped up because we all know he has an amazing track record when pitching on short rest. Leyland probably figures his best shot at getting a dominant JV is on normal rest and I can’t fault his reasoning.

    Comment by jpg — October 11, 2011 @ 3:26 pm

  52. The Tiger’s biggest Problem is not their rotation, it’s their bullpen. If you can pitch Verlander on full rest in one game, you will probably need less innings out of your bullpen than with starting Verlander in two games on short rest where not only his performance, but also his durability may take a hit.

    Let’s just assume that a fully rested Verlander goes 8 innings, while any other starter as well as a short rested Verlander goes 5 innings per game.
    For the two games in question (4 and 7) you have in one case:

    8 innings Verlander
    5 innings random starter
    5 innings bullpen

    in the other case:

    10 innings Verlander
    8 innings bullpen

    So in case B you get 2 more innings out of Verlander than you would in case A. But you would also need 3 more innings from your the bullpen. Especially for the tigers that’s not going to be pretty.

    I know those numbers are kind of random. But the result is always the same. Starting Verlander two times on short rest means more innings from the Tigers bullpen. More precisely: middle relievers (yuck!)

    Comment by grapesoda — October 11, 2011 @ 3:47 pm

  53. Which is why I pointed out that Porcello’s is higher too.

    As opposed to Cameron, who appears to have just made up his odds.

    Comment by RC — October 11, 2011 @ 3:59 pm

  54. “Let’s just assume that a fully rested Verlander goes 8 innings, while any other starter as well as a short rested Verlander goes 5 innings per game.”

    I love arguments that start out with totally unsupported drastic assumptions that make any cogent argument impossible.

    Nice work.

    Comment by RC — October 11, 2011 @ 4:03 pm

  55. Insert any numbers you want. But realistically we have to assume that on average one start from a fully rested Verlander plus one start from any starter the Tigers have available will give you more innings than two starts from Verlander on short rest.

    Comment by grapesoda — October 11, 2011 @ 5:09 pm

  56. “In my heart, it’s a no brainer”. That’s funny!

    Comment by hotwater — October 11, 2011 @ 6:10 pm

  57. He will get bashed for not pitching him twice in the series, but i don’t think it makes much of a difference, let’s face it those Rangers are mashing even if it was against Verlander. The Tigers just don’t have enough.

    Comment by DodgersKings323 — October 11, 2011 @ 6:42 pm

  58. I think i heard on Fangraphs podcast that it’s only a 5% drop in performance, this is the postseason! This is why he pitched all those innings in the regular season, this is what it’s all about so i don’t understand all this nonsense about “protecting his arm” you guys will be sitting at home watching the postseason with the rest of us “protecting his arm”. This is the time you ride your iron horse to glory.

    Comment by DodgersKings323 — October 11, 2011 @ 6:45 pm

  59. “Expected performance”, and other expectations where you take an average of multiple scenarios, is overrated. For example, maybe it would be “expected” for Verlander on short rest to give you (making up numbers here) 7 IP and 3 ER, whereas Porcello would be “expected” to give you 6 IP and 3 ER. But in short series such as there, I think you’re better off looking at the probability of an outlier. So maybe over a population of 10 short-rest starts, you could expect Verlander to be only marginally better than Porcello on full rest, but he might give you a much higher probability of having an outstanding start. When you’re down 2-0 (2-1 now), I think it’s much more worthwhile to look at in that sense.

    And of course you need to worry about an outlier in the other direction.

    Comment by Anthony — October 12, 2011 @ 12:36 am

Leave a comment

Line and paragraph breaks automatic, e-mail address never displayed, HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Close this window.

0.340 Powered by WordPress