Jimmy Nelson - Ready for the Challenge?

[Thurston Fluff]

Since there are multiple WARs and they don't always agree with each other trying to use WAR to determine how many games a particular part of a team won or lost is pretty futile.

It also seems to me the wrong way to use WAR. The player with the best WAR in baseball can let an easy ground ball go between his legs for a game winning run. Had he not made that mistake they would have won the game therefore he lost the game for the team. That does not drop is WAR by one game.

[CheezWizHed]

The player with the best WAR in baseball can let an easy ground ball go between his legs for a game winning run

 

Its just as bad as that same player allowing the first run to score and lose the game 9 to 10, right?

 

I think the point is that because the BP is at the end of the game, it highlights losses where they gave up the winning run as solely their fault. But not scoring enough runs earlier or lots of runs given up by the SP put far more emphasis on those last innings where the BP wasn't perfect. Increase the offense and a couple of those losses are still wins even when the BP gives up a few runs (while giving a statistically "average" performance).

[gregmag]

You can't just look at what happens at the moment a relief pitcher melts down and claim that's the cause of the loss. That makes no sense. What about the hitter who struck out with the bases loaded in the fourth inning, or the starting pitcher who gave up the other team's first four runs? Clearly they have to share some of the blame, maybe a lot of it. I understand the complaints about WAR in this setting, but at least WAR reflects a systematic effort to process all the relevant information. Using bullpen W-L record or isolated analysis of the relief pitcher while ignoring everything else that happened in the game is far less useful.

 

If there's an analytic measure out there that shows the Brewers' pen was responsible for four to six losses in the second half, taking the rest of the team's performance into account, I would love to see it. It strikes me as nearly impossible that the claim is true.

[brewcrewdue80]

I just can't commit the time to look it up to argue the individual situation vs WAR stat. All I can say is that when you get to the bullpen generally, you've compartmentalized a baseball game in to 9 outs or less. Yes they can't control that a team went 0-2 with 2Ks and bases loaded back in the 3rd inning where 1 hit would have likely meant 2 more runs. Increasing the odds their 9 out contribution wasn't met with such official outcome of the game. Nor can they do anything about a team's offense when putting up zeroes in the 7th,8th,and 9th innings being behind by 1 or 2runs and the offense doesn't score creating a win from behind. It's just that in a 9out or less game, they are brought in with a lead and do have an impact on whether they keep that or lose it. And I got 21-22 record. 7-10 in the 2nd half.

 

Maybe someone could go through the 162 games with a % chance supposedly the Brewers had of winning games the moment the starter left the game. What % counts for RPs and what doesn't? If it's 66% or less does the outcome not matter? You could tie the % with all the previous PAs and assume at 66% and lower a RP didn't decide the game? Should it be 75%? And then how about the Brewers coming back and winning a game? What % chance of that happening when a starter leaves does RPs have to be considered as a true impact? Same 34% or lower? 25%? That's kinda where I'm looking, is how many games did the Brewers lose when the starter gives the RP the ball and there's a highly likely chance they should win at that point?

[igor67]

For whatever reason they aren't as commonly cited anymore, but baseball prospectus use to publish leverage stats and did a lot of this work on evaluating the relative importance of being able to pitch your best player only when the game is close and late. My recollection was that your best reliever at maximum leverage could basically double his innings. So 70 innings of Mariano Rivera were worth the same as about 140 innings of a starter with the same ERA.