Question for the stats guys on closers

[danzig6767]

Even though i don't read up much on advanced stats gathering, i understand the basic premise being used when they say so and so position player or starting pitcher was worth X number of wins based on that players stats. When it comes to closers though, does a different criteria need to be used because these guys so often pitch in a situation where their inning pitched will decide which team ends up winning the game or not?

 

Say you have two closers that both had 45 save opportunities and the bulk of their other appearances came in lower leverage situations where both were just in games to get work.

 

Closer A saves 41 of the 45 opportunities and closer B only saves 36 of the 45 save opportunities. Now i understand there are also a fair number of games where closers enter in the 9th of a tie game which could influence quite a bit whether their team win that given night. There are also tougher saves than easier ones, depending on the score. Obviously it's tougher to get that three out save if your team is winning by three runs instead of only one run, i'm sure by fluke that each year some closers face more "tougher saves" than other closers.

 

All that variance stuff said, when comparing closers to others in a season, is there a need to look beyond just stats like ERA/WHIP/K-9 etc and mainly look at the end result, how many games did that team end up winning when their closer entered a game in a save situation in the 9th inning? Let's pretend you're comparing Lidge to Wood, Lidge to Torres, or Torres to Wood. When trying to roughly assess how many more wins one closer was worth vs the other, is anything computed in that say the Phillies never lost a game that Lidge entered with a save opportunity, but the Brewers lost 5-6-7 games that instead of would have been a win had Torres finished it in the 9th like Lidge did for the Phillies? Maybe when computing their respective ERA's to innings pitched, it only equals a 2-3 game advantage, but in actuality, the Brewers lost more than 2-3 games by having Torres compared to Lidge.

 

I'm not exactly sure if i got the point/question out that i was trying to get at. I ask because often on here i'll see when comparing two starting pitchers, a stat guy will say that so and so starter was worth X number of games based on their final stats over many innings. Closers though pitch many less innings, yet, they can have a very dramatic effect on who actually win a game because they so often enter with a small lead and it's the 9th inning. It would seem like a top notch closer who rarely blows a game vs a struggling closer could end up swinging that teams record by 4-5-6 games, maybe even more. Am i wrong?

 

Like many here, i agree that the save stat not only seems kinda silly given the 8th inning guy very often has a nearly equally important job, but it also has so many managers refusing to use their best reliever in other vital late game situations simply because it's not the 9th inning yet. Then again, if one closer blows an extra 5-6-7-8-9 games compared to another one that results in say 3-6 or more losses, those losses count in the standings of playoff races that so often come down to only one or two games. It gets me conflicted on the importance/value of closers.

[kramnoj]

http://www.fangraphs.com/...type=3&season=2008&month= is a fun page to play with to look at relievers and value. It looks like the top relievers are usually worth 2.5-3.5 Wins based on the formulas used there.

[logan82]

If they changed the save stat to 2 runs and eliminated blown saves prior to the 9th inning, it would be much more meaningful. The main reason I don't like putting much money into a closer is because they are not used properly. I doubt that is going to change anytime soon. I am sure somebody can point you to a leverage index that will give you a good idea which points of a game are more important.

 

Quickly glancing back through "The Book", they say that it would be better to bring in your closer in the 8th with a 2 run lead than in the 9th with a 3 run lead. The difference between an ace reliever and an average reliever in the 9th with a 3 run lead is about 2%.

[danzig6767]

Quickly glancing back through "The Book", they say that it would be better to bring in your closer in the 8th with a 2 run lead than in the 9th with a 3 run lead. The difference between an ace reliever and an average reliever in the 9th with a 3 run lead is about 2%.

Is there a place to find out the percentage of 1,2,or 3 run lead saves a closer had on a season?

[Ennder]

The stats that try to assign a win value to players are generally based on future win value more than past win value. Something like WPA is probably what you would want for your closer question.

 

Lidge led all RP with a 5.37 WPA, Rivera had a 4.47. Marmol was the highest non closer at 3.77. The highest Brewer was Torres at 1.13.

 

Gagne had a -.77.

 

So in theory if we had signed Lidge instead of Gagne last year we would have won around 6 extra games. I haven't studied WPA a ton so don't know how accurate it is, I dont' really care what a players value was in the past, I try to focus on what they are worth in the future usually since the past always contains a ton of random noise.

[TooLiveBrew]

Lidge led all RP with a 5.37 WPA, Rivera had a 4.47. Marmol was the highest non closer at 3.77. The highest Brewer was Torres at 1.13.

 

Gagne had a -.77.

 

And just to further clarify (in case it wasn't clear) a 0 value is league-avg. for WPA

[logan82]

Correct me if I am wrong, but wouldn't WPA be more a function of when a pitcher pitched as much as if not more than how well they pitched?

[sbrylski]

Correct me if I am wrong, but wouldn't WPA be more a function of when a pitcher pitched as much as if not more than how well they pitched?

 

Situation is important in WPA. WPA/LI neutralized for situation, however, by valuing WPA+ performances more highly if they were accomplished in low leverage, or unimportant, situations. For example, if a player adds 50% chance of winning during his AB in a average situation, he gets .5 of WPA/LI points. If he adds 50% in a twice as important as situation, he only gets .25 points. So it doesn't give credit to players who appear in high leverage situations, as the player doesn't really control that.

[Ennder]

Yeah I guess Lidge would have stolen some of the WPA from Torres so he would have only been like 5 Wins better than Gagne. Of course who knows what Lidge would have done under a different situation etc. The stats I enjoy the most are the ones that try to isolate how good a player really was to predict what he will do in the future rather than those that just decide how valuable they were in the past which has all the noise left in.

[logan82]

You mean like ERA? I think it will tell you about results, but not talent.

[rluzinski]

"Even though i don't read up much on advanced stats gathering, i understand the basic premise being used when they say so and so position player or starting pitcher was worth X number of wins based on that players stats. When it comes to closers though, does a different criteria need to be used because these guys so often pitch in a situation where their inning pitched will decide which team ends up winning the game or not? "

 

You can't use the 10 runs = 1 win rule of thumb for elite relievers. That only works for players who play average situations over a season (which has an LI of 1). Positional players and starting pitchers. Closers don't face average situations on balance. The average LI for a closer is something like 1.8. I think the average fan presumes something much higher than 1.8, though. It feels like closers are a lot more important than they really are.

[danzig6767]

"Even though i don't read up much on advanced stats gathering, i understand the basic premise being used when they say so and so position player or starting pitcher was worth X number of wins based on that players stats. When it comes to closers though, does a different criteria need to be used because these guys so often pitch in a situation where their inning pitched will decide which team ends up winning the game or not? "

 

You can't use the 10 runs = 1 win rule of thumb for elite relievers. That only works for players who play average situations over a season (which has an LI of 1). Positional players and starting pitchers. Closers don't face average situations on balance. The average LI for a closer is something like 1.8. I think the average fan presumes something much higher than 1.8, though. It feels like closers are a lot more important than they really are.

I can grasp what you're trying to get at. In the end though, if say two closers in the same division take part in roughly the same number of save opportunities and one closer blows 7-8-9 more games than the elite closer and thus their team loses another 4-6 more games, didn't that closer alone cost that team 4-6 wins to their competitor?

 

I'm in agreement with most baseball fans who hate the rigid manner in which nearly every manager uses their best relief pitcher and feel it's silly that in say a 4-2 game that the 8th inning reliever could face Hardy/Braun/Fielder and the closer faces Hart/Hall/Kendall, yet the closer gets the socalled "save". That said, i just see when legit playoff contending teams have a shaky closer situation and they blow quite a few save opportunities, while another team or two they are closely fighting for a playoff berth has a much better closer, that position alone can cause a 4-5-6-7 game swing and easily be the difference between why one team makes the playoffs and the other stays home given how close so many playoff races are decided by. These factors all together conflict me on how to value closers.

 

Let me ask you a hypothetical from a stats perspective. For pretend lets say the Phillies were in our division and we were fighting them for a playoff berth, Torres had been our closer all season, and Torres/Lidge had basically the same number of save chances. Lets also say Pujols was their firstbaseman instead of Howard and Pujols was healthy all year. Which player matchup would have swung more games in the standings to the Phillies, Lidge vs Torres or Pujols vs Fielder?

 

BTW, in case you planned on telling me this, i understand that Lidge will likely never have another season where he doesn't blow a save. I'm only asking this strictly from how a stats cruncher would view this hypothetical and because i can't help but think that the closer role is on one hand way over valued/over paid, managers are dumb to only use their best reliever in the 9th inning, the save stat itself is silly, and yet, the difference between two teams fighting for a playoff berth can so easily be swayed by 4-5-6 games simply by having a better closer than your competitor, thus that closer role was hugely important.

[Fatter than Joey]

I think we continually trip over the save stat/closer label -- I think the thing to try to assess -- is "who is the better BP arm", Lidge or Torres -- If they are both entering games in high leverage situations. Upgrading a BP arm to be used in high leverage situations is probably going to net you some wins, I don't think that there is any doubt about that, of course the caveat though is ensuring that the "upgrade arm" is used in the right spots.

 

I seem to recall, that upgrading a league average 1b, (e.g. Mike Jacobs) to Ryan Howard the year he won the MVP, was like a 4 win upgrade.

 

I am pretty sure if a manager used his best arms in the highest leverage situations they could swing 4 wins a season pretty easily. For arguments sake, let's say that Brad Lidge is the best BP arm available -- if a manager used him 70 times a season in the highest leveraged places instead of a Brian Shouse (like Yost used him, not as an exclusive LOOGY), or Mota, I would bet you could get 4-5 wins -- at least until the actual scenarios unfolded.

 

If used correctly, I'd gladly pay $10M for a BP stud, if I was ensured he'd not be used in the "by the book" manner.

 

Part of me wonders if teams are hanging on to the whole "save scenario" for marketing reasons. I think teams like a fireballer coming in the 9th to K three guys even though the team has a 2-3 run lead. Brian Shouse "holding down the fort" isn't as exciting as ending the game on 3 K's. I think teams like the fans leaving the game on a high.

 

EDIT: A team having a closer is sort of like the WWF that has the finisher move, to end the match -- same sort of fan reaction.

[TooLiveBrew]

Part of me wonders if teams are hanging on to the whole "save scenario" for marketing reasons. I think teams like a fireballer coming in the 9th to K three guys even though the team has a 2-3 run lead. Brian Shouse "holding down the fort" isn't as exciting as ending the game on 3 K's. I think teams like the fans leaving the game on a high.

 

I think there's some truth to that 'fan experience' angle. However I think it's also largely due to the fact that the best pitchers record more outs with the K, and that managers/coaches/organizations are still stuck on thinking that the 'best' RP should be the CP.

[Fatter than Joey]

that and the salary issues as well... No team wants to have middle RPs making more than their closer.

[rluzinski]

I'm not understanding this facination with the closer and saves. Your best reliever should be in the highest leverage situations possible. The fact that the average LI for a closer is about 1.8 suggests that managers aren't misusing them too terribly. If, more often, they put them in with a 1 run lead in the 8th and the bases loaded and 0 out, their LI might be over 2 though. They'd be even more valuable than they already are. The save stat is just a crappy way to measure WPA. Just use WPA. A 3 run save adds almost no WPA.

 

Yes, if two closers faced the exact same situations over a season and 1 blew 5 more games, the other guy was worth 5 more wins. But we want to know how much a closer IS worth, not WAS worth. Lidge isn't worth whatever his WPA was last year to any team this year. He WAS worth that much to the Phillies last year. The fact that we all know that he won't repeat his perfomance next year shows that we understand some luck was involved. Spead in performance is always greater than spread in true talent.

 

So, a closer might face situations that are 80% more important than a starting pitcher, but the starting pitcher typically has 3 times the IP. That's why they get paid so much more.

[TooLiveBrew]

The save stat is just a crappy way to measure WPA. Just use WPA. A 3 run save adds almost no WPA.

 

WPA is really great, since it works for pitchers & hitters, and it's actually quite simple to understand. The fact that I feel I have a decent grasp of it should tell you all you need to know on how easy it is to 'get'. I agree with Russ that WPA is a great way to value RP in particular, since so much of their assumed value is in how much high-leverage work they get.

[logan82]

TooLiveBrew wrote:

I agree with Russ that WPA is a great way to value RP in particular, since so much of their assumed value is in how much high-leverage work they get.

What is their actual value though? I don't like the save stat either, but I also don't like giving extra value based on a manager's opinion.

[rluzinski]

Their role has more potential value., since the result of their perfomance has a greater influence on the result of the game. Of course, even having those high leverage situations available is also contingent on the rest of the team's performance over the previous 8 innings. That's why we shouldn't get TOO hung up with WPA or saves.

 

To once again compare closers to starting pitchers, it's a heck of a lot easier to ask a pitcher to get 3 outs than 18+. The average pitcher's performance as a reliever is significantly better than as a starting pitcher. An average starting pitcher could be a passable reliever.

[endaround]

See I hate WPA for hitters because the definition of a situation is so out of there hands. For example, a hitter hits a 2 run HR in the top of the 4th in a 4-0 game they're leading. That has no WP value. Now if the other team scores 5 runs in the bottom of the fifth it still has no value. Say instead the team is now the home team and bats second in the bottom of the 5th after a 5 run outburst by the opponents in the top of the inning, that 2 run HR now has about .75 WPA.

[mothershipconnection]

They'd be even more valuable than they already are. The save stat is just a crappy way to measure WPA.

 

If they go with a simple stat, it should be a variation of the hold. Here's my proposal:

 

1) Enters game in the final 3 innings.

2) 2 Run lead or less

3) Records at least three outs

4) Does not allow the tying runner to reach base.

 

Or

 

1) Enters game as a non-starter, regardless of score/lead.

2) Records at least 3 outs.

3) Does not allow a run.

 

This makes a scoreless IP or late-game hold to count the same. I think a hitless 5th should get you at least the same credit as coming in with a 3 run lead in the 9th and allowing 2ER. Yep, every RP would have a bunch of these, but really great middle relievers would look great, and maybe that leads to closers being used in the proper situation to both benefit the team and pad their stats. This also allows a pitcher to get credit for coming in in the 8th, recording 4 outs, then giving way to an effective LOOGY to end the game.

[rluzinski]

I don't hate WPA for hitters. It tells you who happened to hit when the game was on the line. It's interesting. RBIs are interesting as well. As long as you don't abuse either, there's nothing to hate.