FG on Fox: Some More Advice for Ned Yost
Two weeks ago, before the Royals squared off against the A’s in the Wild Card game, I offered a few pieces of advice for Ned Yost, focusing on limiting the amount of outs he asked James Shields to get, not bunting too too early in the game, and using their speed to steal a bunch of bases. Using something not too far off that blueprint, the Royals staged a miraculous comeback and won the Wild Card, and they haven’t lost since. So now, one game away from a trip to the World Series, I have one more suggestion for Ned Yost.
You have two starters available on full rest: use both of them.
Due to the Monday rainout, the Royals have both Jason Vargas and James Shields available to start today’s Game Four, though the Royals have already announced that they’re going to stick with Vargas, while Shields remains on deck to pitch a Game Five, if it proves necessary. The easiest way to make that game unnecessary, though, would be to use both of them on Wednesday.
Dave is the Managing Editor of FanGraphs.
I like the idea – I’d suggest even doing this on a regular basis during the regular season. It seems like a great way to ‘acclimate’ young pitchers to MLB, while limiting their innings during their first year or two. Depending on the team’s personnel, you could have your 5 and 6 starting pitchers basically tandem start each time out for about 4 innings apiece, alternating who starts each game (KC could have done this perhaps with Ventura/Duffy for instance). Not only would it be good for their development at the big league level, but it would increase the team’s win probability in each game, since you’re pulling the young starter before the lineup can face him for a third or forth time.
Excellent piece. I seriously doubt Yost would follow through, but I really enjoyed the article.
This kind of idea will never take off as long as starting pitchers are required to pitch at least 5 innings to be eligible for a “Win”.
That only matters if the team is more concerned about pitcher wins than team wins.
I completely disagree. Unless a team’s entire rotation is a bunch of young pitchers, who don’t have much of a say in the matter, teams have to be concerned with giving their pitchers a chance to earn wins. So long as compensation is, at least partly, tied into traditional stats like wins and saves, managers are gonna keep managing starters and closers the way they do, even if it’s a suboptimal use of there talents. There is a reason even the the very best managers almost always allow their starters to try get through the fifth inning if they have the lead even when they are struggling. Just like every manager manages their bullpen to the save rule. Keeping veterans happy is part of the equation. The quickest way to upset them (and their union) is to do things that will hurt their future earning potential.
I assumed the post was about the first comment in this thread:
“you could have your 5 and 6 starting pitchers basically tandem start each time out for about 4 innings apiece, alternating who starts each game”
That is generally taking away wins from just the bottom end of the staff.
And the poster suggested alternating, so most of those wins arelikely ended up with the second starter and if they are alternating, they probably even end up with a larger share of the wins from those starts [since they will presumably combine to take away wins that the regular bullpen had been collecting].
In any case… some teams are already sticking talented RPs in as setup men rather than closers to save money in arbitration [less saves].
A solution in search of a problem.
http://www.baseball-reference.com/players/split.cgi?id=vargaja01&year=2014&t=p#innng::none
For 2014, with the KC defense behind him, the ERA is fine until the fifth, and is actually stellar in the sixth.
In addition, opponents OPS first time through the order in ’14 is .680, while the second time is only .699. He should be fine to get through 5, if not 6 this afternoon.
…you’re using one seasons worth of data, split by inning. Come on
A pitcher’s own TTO splits have no predictive power anyway. Cameron was using a strawman when he showed you Vargas’ own TTO numbers. What if they had been the reverse – he had gotten better each time through the order in his career? That would presumably negate his whole thesis, right? No. Because those personal, unique numbers are meaningless. We must make the assumption that ALL pitchers have a similar TTO penalty regardless of what their own numbers have been in one season or their career.
What if a manager is going to bring in his closer – his ace – to face the opposing team’s heart of the order in a 1 run game in the 9th. But, wait, he sees that with this pitcher, he has done poorly against team’s 3-4-5 hitters this season or in his career, even though he is a great pitcher overall. Would he not bring him in? No, because those numbers have no predictive value. By chance alone, SOME pitcher is going to have particularly bad numbers against 3-4-5 batters. SOME pitcher is going to have especially bad or good numbers against ANY subset of hitters or situations by chance alone.
What if you are thinking about throwing your best pitcher in an elimination game on Sunday, but wait, he has not so great numbers during the day, or on Sunday’s. Do you hold him out? No, same thing.
I’m going to keep harping on Cameron and other sabermetric writers until they stop implying and disseminating false and misleading information.
And as long as your comments are regularly of the form “I know the Absolute Truth and this article is a failure because it doesn’t include every single detail that I consider important”, I would imagine that the writers will continue to ignore you.
I appreciate your feedback. Without your comments (and Twitter feed) I’d often be unaware of when posts on this site detail misinterpreted data.
If you have Shields who is on full, regular rest, why would you not just plan to start him this game? If he wins, he’s still rested for game 1 of the World Series. If he loses, then you just try to live with Vargas in the next one. Unless if there’s a reason to think he’d be a better starter tomorrow, I don’t see why you don’t just pitch him this game.
With a 3-0 lead in a best of 7 series, using Vargas makes a lot of sense to me – they’re overwhelmingly likely to win even if they lose today, and value in keeping him fresh as they’d almost-certainly need a 4th starter in the World Series and letting Vargas go 23 days between starts has a high probability of going badly.
Yes. Well said.
Also Shields has been on the regular 4 days rest over his last 8 starts – it might be good to get him an extra day of rest this far into the season. I think usually pitchers get that 5th day of rest more often than that with reg season off days (haven’t looked it up – this is more of a guess/speculation).
As he will only get 1 more start this series, I think this makes sense especially with your point on Vargas – unless they were planning to go to him on 3 days rest in game 7 (had he started game 4)
The O’s on the other hand…in an elimination game, not sure why Tillman isn’t starting on normal rest. Then again I have no idea why Chen was left in for the 6th inning yesterday (If Yost makes that move he gets panned by the SABR community – I assume Buck’s reputation gives him the pass?)
I disagree on this one. I would rather see the Royals go to the bullpen early if needed. You can lean on Duffy, Frasor, Herrera, Davis, Holland, Finnegan for an inning each or more if needed. Save Shields for tomorrow if the game is necessary. That gives Ventura an extra day, at least, to rest a balky shoulder.
Advice for Ned Yost: ignore advice from non-team personnel…you’re one game from making it to the world series.
Well, Yost pulled the trigger once by using a starter in relief, and he got bit (Ventura, Yordano). On the other hand, you have Danny Duffy who has been little-used so far. Use him.
Or, don’t listen to anyone. Hindsight is 20/20.
Billy Butler stole a base. That proved beyond a doubt that there’s some sort of mystical force compelling this team to win via divine intervention. Yost could have Jeff Francouer pitching to Luke Hocheaver and the Royals would still clinch tonight.
You also said the Royals should trade Shields at the deadline because their best hope was a one and done play-in game against the Angels.
I’m just sayin…..
And we’ve gone 48-23 since then.
I think I’m gonna start Vargas TWICE today.
The Royals have to lose 4 straight, let’s not over think this. Maybe this actually would be the right choice, but its just not going to happen. Shields will pitch game 5 and as long as Yost is willing to go to the bullpen early, the difference between Vargas for 3 and Shields for 3, vs Vargas twice through the order then the bullpen would seem minimal.
But if we’re going here:
Remember, 3 IP is probably seeing the top of the order twice anyway. So, I don’t really understand the 3IP limit. Clearly the times-through-the-order is what to care about. So why make suggestion that probably leaves Vargas going through 12 batters or so, seeing 1-3/4 twice, then Shields seeing another 12 batters, but probably only seeing batters 1-3/4 once? Would we really want to see Vargas for 1 trip through the order, then Shields for 1 trip through the order? That might only get you into the 5th inning, and you’ll probably have to remove a starter, to be replaced by a starter, in the middle of the inning, which always seems dubious since starters are used to starting an inning clear of base runners and knowing exactly when they will come into pitch.
If this is going to happen, I’d much rather see Vargas for 9 batters, a reliever that could maybe get 1.1-1.2 IP (thinking this reliever would get you through batters 1-5 their second time up, at least), so that both Shields could start a fresh inning (probably 5th, maybe the 4th), and you don’t have to burn a bullpen arm for 1-2 outs just to get that fresh inning. Then leave Shields in to see the top of the order twice (pulling him potentially at the DH spot if he doesn’t have the platoon advantage, if he does pull him after the DH spot). But since the whole point of this is to make him available for a game 7 on short rest, let the relievers handle that bottom of the order (and depending on when Shields came in, this might actually be the 3rd time through the order for the 6-9 guys anyway).
I can see the appeal of this, because it could let Shields face the order twice in game 4 and at least once in a possible game 7. Rather than probably 3 times in game 5 alone, so you’re maximizing his effectiveness over the remaining innings in this series. But this trade off is probably very small compared to aggressive bullpen usage with Vargas/Shields starting game 4/5 respectively, and a “all hands on deck” game 7, if needed.
Advice for Yost. More proof Dave hates Baltimore. 😉
This advice is rooted in logic and extremely sound. Are the Royals fans here trying to overcome the O’s fans for most bitter fans of teams that are winning? I really think bitterness should be reserved for the rest of us.
This seems like over thinking things. You are up 3-0 and you are talking about manipulating your starting staff for game four based on who might be available if it miraculously gets to game seven? That seems like managing for the sake of managing.
Besides, if you were going to be uber-aggresive to try and close out the series today, why not just start Shields? Wouldnt that be the more logical strategy? Why bother with the Vargas/Shields platoon?
This is excellent and sound advice, and…
A pitcher’s own TTO splits have no predictive power. Cameron was using a strawman when he shows you Vargas’ own TTO numbers. What if they had been the reverse – he had gotten better each time through the order in his career? That would presumably negate his whole thesis, right? No. Because those personal, unique numbers are meaningless. We must make the assumption that ALL pitchers have a similar TTO penalty regardless of what their own numbers have been in one season or their career (pitchers with more or fewer pitches in the repertoire can be assumed to have a slightly less or more penalty).
What if a manager is going to bring in his closer – his ace – to face the opposing team’s heart of the order in a 1-run game in the 9th. But wait, he sees that with this pitcher he has done poorly against teams’ 3-4-5 hitters this season or in his career, even though he is a great pitcher overall. Would he not bring him in? No, because those numbers have no predictive value. By chance alone, SOME pitchers are going to have particularly bad numbers against 3-4-5 batters. SOME pitchers are going to have especially bad or good numbers against ANY subset of hitters or situations by chance alone.
What if you are thinking about throwing your best pitcher in an elimination game on Sunday, but wait, he has not so great numbers during the day, or on Sundays. Do you hold him out? No, same thing.
I’m going to keep harping on Cameron and other sabermetric writers until they stop implying and disseminating false and misleading information.
“We must make the assumption that ALL pitchers have a similar TTO penalty regardless of what their own numbers have been in one season or their career ”
No we don’t. As you stated in your next line, pitchers with more pitches(I would amend that to say effective pitches) would likely have less of a TTO penalty. That makes logical sense since a pitcher would be able to have a better chance at mixing pitches and keeping hitters offbalance even after facing them once or twice in the game.
Regression would be a factor, but once you get to a large enough sample size, why would you assume that James Shields TTO penalty would be similar to Vargas, when their career numbers state otherwise?
Youre next three examples are just small sample size garbage with no basis in reality.
Yeah, I’m not sure why he has to make that statement with such arrogance, but he’s been spewing it for the better part of a year. The beauty of his statement is that he can never be proven wrong though. If you identify pitchers who have a less-than-average penalty, he’ll just say it’s due to small sample size and that you can’t take anything away from the data when there are so few innings.
That’s the fangraphs special. Process is everything! And if something turns out wrong in the results, its just SS. Circular reasoning 101.
John.
No we don’t. As you stated in your next line, pitchers with more pitches(I would amend that to say effective pitches) would likely have less of a TTO penalty. That makes logical sense since a pitcher would be able to have a better chance at mixing pitches and keeping hitters offbalance even after facing them once or twice in the game.
If this were true, we could certainly find it. Simply get a list of all pitchers with 3+ effective pitches, maybe set some minimum thrown percentage cut off. Then see if that group has a less TTO penalty than those that don’t make that cutoff. Now, you’ll have to be sure to normalize things by each groups overall talent, since as you point out, those with more pitches are likely also better pitchers.
Alas,
MGL has done this: http://www.baseballprospectus.com/article.php?articleid=22235
The distinction there is mainly those that are mainly 1 pitch pitchers, or those with at least a 2nd pitch thrown 15% of the time. If you scroll down you’ll see a table with 1 pitches > 15%, 2 > 15%, 3 > 15%. And you’ll notice the biggest single TTO effect is for 1 pitch pitchers after 1 time through the order. While for 2 and 3 pitch pitchers, the TTO penalty is about the same each time through.
Anyway, its nice to bring up there is more too it than just TTO. Different pitchers do better or worse, but MGL’s main point here is that single season TTO numbers are meaningless. I have no problem stating that strongly, because we see single pitcher (both career and single season) TTO numbers come up fairly regularly. When instead we should either use the league average penalty or if a little more is done, pitcher type penalty. Also, both Vargas and Shields throw more than one pitch. So, I’m not sure it really matters in this particular discussion.
Let’s say Yost did this, and the Royals lost. That would mean Ventura (who left his last start with a stiff throwing shoulder) going tomorrow on normal rest. If the Royals lost that game, who pitches Friday? Guthrie on two days rest? Duffy, even though the article argues you don’t want him pitching game seven?
I think, even if Yost wanted to follow this strategy (which is very non-Yosty), the lack of the travel day due to Monday’s rainout makes it foolish.