In retrospect, how much has losing Prince hurt?

[Ennder]

Yet they have a worse record. Does this mean that relief pitching, particularly in high-leverage situations, is a lot more important than people think?

 

It is more important and less predictable than people think and that is really the problem. Every year there are a few teams that are said to have great bullpens who are just terrible and teams who had terrible bullpens that are amazing because the sample sizes of a single season of RP are so tiny that you have no idea how good anyone but Rivera is going to be pretty much.

[rluzinski]

Does this mean that relief pitching, particularly in high-leverage situations, is a lot more important than people think?

 

By definition, if we've identified that it is a high leverage situation, we've acknowledged that it is more important than average. The "how much" can be quantified by using a metric called "leveraged index" (LI) or something similar. LI looks at the various possible win probability swings for a particular situation and compares it to average. If I remember right, an LI of 2 means that the situation is twice as "important" as average, for example.

 

As such, you will see saberists multiply the expected value of an elite reliever's performance in an average situation by some LI value to acknowledge that the situations he typically finds himself are in higher leverage ones.

[LouisEly]

There’s been a quite a bit of talk about high leverage situations over the years, but how does a situation become high leverage in the first place?

 

If you look at the history of baseball, with the exception of a handful of really bad teams and a handful of really great teams (playing in the same division as some really bad teams) pretty much every team wins 1/3rd of their games and loses 1/3rd of their games. However the other 1/3rd of games are going to be close games where every run matters. In those games the relief pitching can have a significant effect on the outcome of the game where two runs may make the difference between winning and losing. Maybe it's more a case of "high leverage games" than high leverage situations.

 

the sample sizes of a single season of RP are so tiny that you have no idea how good anyone but Rivera is going to be pretty much

 

While the sample size of a single RP is usually going to be small, the aggregate of the entire bullpen and subsequent impact is not. Looking at the Brewers starting pitchers, if you divide the innings pitched by starters by the number of starts you come out to almost exactly 6.0 innings per start. That means that the bullpen pitches about half as many innings as the starters (extra innings games offset the times where they don't have to pitch the bottom of the 9th on the road because they lost). I can't say that relief pitching as an aggregate is a small sample if it is half of the number of innings the starters pitch. I hear what you are saying - you are referring to individual relievers. I am referring to the aggregate.

[Baldkin]

If you look at the history of baseball, with the exception of a handful of really bad teams and a handful of really great teams (playing in the same division as some really bad teams) pretty much every team wins 1/3rd of their games and loses 1/3rd of their games. However the other 1/3rd of games are going to be close games where every run matters. In those games the relief pitching can have a significant effect on the outcome of the game where two runs may make the difference between winning and losing. Maybe it's more a case of "high leverage games" than high leverage situations.

 

Actually, the opposite is true. Over large samples, close games for almost all teams, hover right around .500. What the really good teams do, that the terrible teams don't, is they blow people out. They win 5-0, they win 8-2 way more often than the terrible teams do.

 

To use a current example: Baltimore has the single best record in one run games in the history of baseball this year. Would you call them a dominant team?

[rluzinski]

Actually, the opposite is true. Over large samples, close games for almost all teams, hover right around .500. What the really good teams do, that the terrible teams don't, is they blow people out. They win 5-0, they win 8-2 way more often than the terrible teams do.

 

I don't believe that is true. We use .500 as a generic expectation for all teams only because it is convenient. Good teams score more runs and/or give up less runs than the average team. I don't believe their distribution of runs scored/game is any different than anyone else, however.

 

And I believe Ennder's point was that the low number of innings a reliever typically pitches makes it difficult to estimate how well we should expect them to perform going forward. Therefore, one must be careful not to overpay for a reliever based on an estimate with a low confidence. He was not suggesting that relief performance is less important.

[PrinceEatMeat]

I posted this in the Ramirez thread, but thought it applied to an extent here too. I was curious yesterday about Fielder vs. Ramirez stats wise. They are pretty similar actually. Obviously there are other varying factors, but it's definitely interesting to look at. Obviously Ramirez isn't the only one working to replace Fielder's production either.

 

Fielder: .305/.403/.518 26 HR 98 RBI 29 2B

Ramirez: .296/.361/.529 23 HR 91 RBI 44 2B