Yankees Shouldn’t Go All In Yet
Facing elimination tonight, the Yankees are pulling out all the stops by announcing that CC Sabathia will be available out of the bullpen tonight, even though he started Game 5 on Wednesday and threw 112 pitches. For his part, CC says that he “could throw 45 to 50 pitches” on one day’s rest. My question – should he throw any?
Yes, the Yankees have to win tonight or their season ends. But they also have to win tomorrow or their season ends. As often as the “take it one game at a time” cliche is run out there, the Yankees actually have to take it two games at a time, because winning just one doesn’t really do them any good. Winning tonight is critical, of course, but so is winning tomorrow – they are of equal importance. And I’m not sure that using Sabathia in relief tonight is the best use of resources in effort win both games.
The working assumption is that a pitcher’s quality of performance goes up with each off day they get between starts (to a point, anyway). If we accept that premise, then Sabathia should be more effective tomorrow than he would be tonight. We don’t have nearly enough data to quantify the difference between a starter throwing a few relief innings on one or two days rest, but we should be able to safely assume that there is some difference, even if we don’t know how much.
If they use Sabathia tonight, they would essentially be borrowing win probability from tomorrow night’s game in order to increase their chances of winning Game 6, but the potential quality difference could make that a bad trade-off. Using made-up numbers to illustrate the point, let’s say these were the respective probabilities of winning each game if Sabathia was used in relief tonight versus tomorrow.
Sabathia relieves tonight.
Game 6 – 55 percent
Game 7 – 40 percent
Chance of winning both: 22 percent
Sabathia relieves tomorrow night.
Game 6 – 50 percent
Game 7 – 50 percent
Chance of winning both: 25 percent
Under this scenario, the Yankees would increase their odds of winning tonight but decrease their odds of winning the series. Now, these numbers are not real, and in reality using Sabathia in relief for an inning or two is not going to make a difference of this magnitude in either game, so this isn’t some kind of colossal screwup if he pitches tonight. But I do think the win-Game-6-at-all-costs mindset misses the bigger picture.
You have to win Game 6 to force Game 7, absolutely. But you have to keep the overall goal in mind, and that is to win both games. My guess is that the Yankees best chance to win both would be to save Sabathia for tomorrow, when he may be more effective. I understand the temptation to use him tonight, but I think the Yankees may be better off if they resist that urge and try to win tonight’s game with the rest of their bullpen.
Dave is the Managing Editor of FanGraphs.
I disagree. The trick for the Yankees is to decide at what moment is best to use Sabathia over the next (potentially) 18 innings. If the Yankees are up by 1 in the 8th tonight and Josh Hamilton is up, Sabathia might be the best option. Its entirely possible that the Yankees never would enter into such a high leverage situation in game 7.
I like this logic
Maybe, but the possibility of such a situation is low. The leverage of a starter is going to FAR outweigh that of a reliever in all but the most immediately blown out games. Simply put, getting through 6 is almost always worth more than getting through 1.
Plus, it’s not like their bullpen is entirely incompetent. If anything, they should be throwing out Mariano Rivera if they need a high leverage guy. He’s still the best flat out pitcher on that squad, and he can give you two if you need him. You’d be a heck of a lot smarter to blow him on a tight game, pitch him into the ground, and pray that Sabathia can give you 8 or 9 the next day so you don’t need him.
Except Sabathia was never going to start tomorrows game. He was only available from the pen. Not that it matters to the Yankees anymore.
The difference here is that Andy Pettitte is much more likely to hand the ball to the back of the Yankee bullpen than Phil Hughes. Hughes might pitch into the 6th, while Andy might pitch into the 8th. So you might want Sabathia out there in the 6th if Josh Hamilton comes up with a couple of men on base.
Right. I think the notion that Sabathia is available tonight says more about the team’s faith in Hughes vs. its faith in Pettitte. They could save Sabathia for tomorrow, when they are less likely to need him, at the expense of his performance tonight. I do think they should have Sabathia ready tonight and treat tonight, essentially, as a bullpen game and if that costs them tomorrow night, so be it. I have more faith in Pettitte than I do Hughes.
I think if you’re looking at using CC over Joba, Robertson, plus some smattering of Logan tonight, you save CC for tomorrow. The time to use CC tonight is if he’ll be pitching in place of Mitre or Mosely I think, which would either happen in a) extra innings or b) if it is a close came early but Hughes can’t pitch. Extras you’ve probably already used your best pitchers to preserve the tie, so get 2IP out of CC while you can. Early close game w/ Hughes out…I’d still probably try to get 6IP out of Joba/Robertson/Logan/Wood/Rivera, and save CC for 3IP tomorrow. But I can see trying to get 2IP out of CC and still be able to get a few innings out of the others tomorrow.
If Sabathia really recovers that fast, one wonders why he wouldn’t just pitch 3 innings every other game during the regular season.
Because there is a big difference between asking someone to do something once with an entire offseason to recover vs. asking them to do something 80 times in 6 months?
Besides which, the Yanks would still basically be getting the same number of innings out of him as they are now. There’s no upside to it.
Unless you could make sure those 3IP are higher than the average starter’s leverage. But no, I don’t think this is realistic anyway.
Didnt LaRussa try something wierd like that over when he was managing Oakland, back when ? Didnt he try to have his starters work 3 innings each day (or two), then sub in the next “starter” for innings 4-6…
You’d also think that their odds of winning tomorrow night are already lower, what with Cliff Lee taking the hill. If it were me, I’d want to save my bag of tricks for a potential Game 7.
Of course, they could just be bluffing. Hard to know who the audience for that kind of play would be, though.
You can’t save any tricks for a Game 7 that might never happen. Win Game 6 and then worry about Game 7.
Your analysis is flawed on the basis that you’re only using CC if you’re already losing the game, or have a highly leveraged situation to get out of an early jam in the ball game. the analysis has to look something more like
Game 6 with CC – 30%
Game 6 without CC – 10%
You can’t realize the probability of winning tomorrow UNTIL you win today. So lets say you use CC and win today, you get a
Game 6 with CC – 30%
Game 7 without CC – 40%
versus
Game 6 without CC – 10%
Game 7 without having won game 6 – 0%.
This is how Joe Torre lost the 2003 WS to Florida by saving Mariano/Hammond/Nelson and/or starters, while Jose Contreras/Jeff Weaver pitched in 5 or the 6 games, losing 2 of them.
55,40 vs 50,50? Sure, you are just making up numbers – but you are making up numbers to support your premise. You have Sabathia relieving in G6 adding 5% to WP, but you have Sabathia relieving adding 10% in G7. One extra day of rest for Sabathia who would likely throw 3 IP is not going to raise the Yankees WP by another 5%.
Not trying to nitpick, but you are adjusting the WP too much for Sabathia being used in relief. The numbers you use are important, because the larger spread you use the more your premise looks correct. A better spread would be the following.
50,40 vs 47,44
Which would give you… 0.20 vs 0.2068
Your 3% difference now becomes a more likely 0.68% difference. My estimates are likely not 100% correct – but easily more correct than yours. I do this stuff every day. Even if a fully rested Sabathia were to start G7 instead of Pettitte you are looking at “at most” an 8% difference in WP. You have a 10% with Sabathia relieving on short rest. If you make up numbers you can falsely prove any premise.
I agree with this. What the very first poster said makes the most sense. You use Sabathia in the best spot over the next 18 innings that you can possibly find. You dont say “Well, I’m definitely saving him for game 7” or “I’m definitely using him tonight”…. you use him when you think you need him most, balancing all the factors in hand.
Yes. I agree with this too.
“If you make up numbers you can falsely prove any premise.”
I was going to say the same thing, but you presented a much better argument than I could have.
I agree. If a situation were to arise in a game 6 situation like this where your SP on little rest was needed in relief, you take it. Because you don’t know if a similar situation would ever come up in game 7.
So while you’d theoretically (and realistically) expect a better performance on another day’s rest, you can’t assume you’d have an equal need in each game. So if you need him in game 6, you use him. Game 7 could end up being a 10-0 blowout for all you know sitting on the bench in game 6.
Using CC in game 6 just for the sake of using him doesn’t make sense. But if the need is there (ie extra innings, or Hughes gets yanked early), then Girardi should absolutely bring him in.
Cliff Lee could wrench his back getting out of bed on Saturday. Pettitte could pitch the game of his life. You have no idea how game 7 will play out until the first pitch has been thrown. When you’re in the midst of game 6, you have to play to win that game. Worry about game 7 when that game begins.
Bad move.
CC should be available for relief Saturday, but not tonight. Worked too hard on Thursday.
“50,40 vs 47,44
Which would give you… 0.20 vs 0.2068:”
For the mathematically disinclined, which numbers are you squaring to solve this probability? .40 sq +?
Multiply the two win probabilities together.
ie: 0.5 * 0.4 = 0.2
and 0.47 * 0.44 = 0.2068
Proper handicapping is very important in this exercise. If you are just going to make numbers up, you can support a wide range of premises. I am not sure any of us can handicap this with certainty, but there is no way that using Sabathia in three innings of relief on “short rest” improves the Yankees WP by 10%. Like others before me stated, pick what you think is the highest leverage couple of innings and use him there. It will involve some guess work on the managers part.
Many good points brought up in the comments. Missing some context though.
Logan was designated the LOOGY, but it’s become pretty clear that he doesn’t have much that he can offer Hamilton, hurting as the slugger may be, to get him out consistently. Girardi sees that and accordingly is looking to CC as a potential lefty arm he can use against Hamilton.
In Game 7, by contrast, CC’s role as a BP lefty would be minimized since Andy is starting. Between Andy, Mo, and whoever he trusts most between D-Rob and Wood, Hamilton is not as much as concern in Game 7.
If Hamilton comes up in a big spot, late, with Hughes tiring (especially given Hughes’ trouble with lefties), expect to see CC.
Dave, one of the few times I’ve really disagreed with you here. Here are the issues I have with your philosophy here..
1) if they lose game 6, there is no game 7, making the probability of a game 7 victory 0%. This should factor in.
2) playing for a potential game 7 is nonsensical if there is a potential to lose that game
3) it’s all based on situations. If the Yankees are up by 3 or 5, pitch Robertson/Joba/Logan/etc. If they’re tied or up by 1, go with your best pitcher (Sabathia) in a high leverage situation. We always talk about throwing the best pitcher out there in high leverage situations- why should this scenario be any different?
My issue that CC on 1 day’s rest is you don’t have any idea if he is your best pitcher, much less how good he will be. We know he’s not better than Mo out of the pen. And he’s probably not better than Joba either. What about Wood? I’d say probably not. Logan? Maybe? Robertson? Maybe? Just seems silly. You have a pen full of pretty good relievers that should be well rested for tonight’s game. Wood might be the only guy I’d be half worried about having thrown 2 IP on Wednesday though it was *only* 28 pitches (basically 2 pretty easy innings).
So unless Hughes gets pulled in the 3rd inning, I just don’t get it. And its not like I’d pull Hughes from a high leverage early jam and put in CC, who I have no idea how well he’s going to pitch. I’d think if CC is to pitch, you have to start an inning with him and pull him at the first sign of trouble. You don’t want to throw him in there with the score 3-2 Rangers with men on 2nd and 3rd with 1 out in the 4th. He’s just too big of a question mark.
Sabathia has never pitched on less than 3 days rest in his major league career, and is not particularly effective against LHB. Nonetheless, if it’s 4-3 after 4 innings, I can see that you might try to get 2 innings out of him. The ethics of this are dubious, and it’s a real crapshoot whether Sabathia would in fact be effective.
I cannot however imagine circumstances where it would be preferable to run him out there in the eighth with a 1 run lead to face Hamilton rather to bring on Rivera for the long save.
I wish Bob Brenly had ignored his “ethics” and used BYK to pitch the 9th of game 7 in 2001 instead of Randy Johnson who started game 6.
It’s a little different. Johnson pitched Game 6 on 5 days rest, and the D-Backs gave him a 15-0 lead after 4 innings. You might imagine that while he threw 104 pitches that day, he took it very easy and told Brenly that he was ready to go the next day. It was also Game 7 of the World Series, so that was it.
Using made up numbers, what’s the point?
Of course the average is higher in the scenario you paint. You rigged it–CC improves game 6 outcome by 5 % and game 7 by 10%.
Right, it is rigged. As mentioned above.
Also factored in should be the benefit of the series going to 7 games – in terms of revenue, fan enjoyment, being able to say you stretched it to 7 instead of falling in 6 and so on – versus 6.
I estimate this to be worth a small amount, a little less than a 1% difference in overall series WP – about .68%, in fact.
Winning is important, but going down “swinging” is worth a smidge of something.
Guys, it’s not “rigged”. He SAID that this is following the premise that pitchers are better the more rest they get. That’s why using him tonight adds less percentage tonight than it would add tomorrow. The numbers are made up, but IF you accept that pitchers are better on three days’ rest than two (which you should), it’s not the magnitude of difference that matters, it’s just that there IS a difference.
That said, I agree with what everyone’s saying about both leverage and using him in a game where the starter’s likely to give you fewer innings.
The magnitude of the difference is what matters. If there’s a difference, but it is as small as Xeifrank estimates, it makes more sense to use CC in game 6 because you can’t get to game 7 without winning game 6. If the difference is greater like Dave estimates, saving him for a potential game 7 becomes more relevant in the decision whether to use him in game 6 or not.
~$200M and yet, everything seems to come down to CC Sabathia.
That, to me, is the most amazing thing about the whole situation.
it also came down to the rest of that 200mill. Burnett was in there too !! Dont forget his “achievements” in the playoffs.
On full rest, CC gave up 11 hits over 6 innings and 110+ pitches two days ago. Is the assumption that he will not be as “effective” as his last outing, but better than his first start on full rest? What kind of value is that? Is there really value in the possibility of him pitching tonight?
I think that I understand the psychological impact of a guy saying, “I’m there if you need me.” But if that guy has a wing that’s hanging, isn’t he only more valuable that Mark Teixeira because Tex isn’t eligible (and possibly Sergio Mitre at “full strength”)?
I just don’t see the scenario where CC pitching tonight is a good thing for the Yankees, except in the 17th inning with an empty pen. But then again, I’ve missed things before.
For what it’s worth, according to “Vegas,” the Yanks are about 41% to win Game 7 given Game 6 is a win. It’ll be interesting to see what the actual odds on the game are given Sabathia’s involvement (or uninvolvement) in the game tonight.
Sure, but if you are going to use that as a measuring stick (G7 movement) you need to take into account that some of that is already priced in. How much? Who knows?? And there will/could be other stuff that move that 41% line. Good post.
If there is a situation where using CC in tonight’s game makes sense, they should use him.
My concern isn’t “burning” him tonight and not having him tomorrow. My concern is that he’s kinda sucked so far this post-season.
I like the idea of using your best pitcher in the highest leverage situation. What many people are missing is that their best pitcher is not left handed. Their best pitcher is beyond a doubt Mariano ‘The Closer’ Rivera. He pitched in Game 5, but they had an off day, so he should presumably be at full strength, especially not pitching in the consecutive losses. Now if it’s the eight inning with men on base and Hamilton up, I want Mo brought it, not CC. If it’s the 10th inning of a tied game, sure bring in CC for a couple innings to preserve the tie. If it’s the 4th inning and Hughes is knocked out but they’re tied or within a run on either side, then sure bring in CC for a couple innings to get to Joba, Robertson, Wood, Logan, Mo. But in the 8th inning up by a run with men on base and Hamilton up to bat, I want to hear Enter Sandman on the speakers and see number 42 jogging across the outfield grass.
I like the idea of using made up numbers to support a self-made claim.
Then you must be a fan of “WAR,” which uses the exact same methodology.
Question is moot. 🙂
Bottom line, in an elimination game, all pitchers are available, all batters are as well, even fielders (playing out of position, if need be). Its the classic “play like there is no tomorrow”. If they had won 6, they get to hit the reboot for game 7. Yes, they would have been in a lesser position due to fatigue, mental as well as physical. BUT, they would have been there (assuming that they played as if there was no tomorrow), if everybody was available and played, and won. Now, its a moot point whether CC gets to pitch on 3 days rest or not.. Now he gets to pitch on 150 days rest (or so) !