Fun with PECOTA

[Mass Haas]

Via BrewCrewBall:

 

PECOTA Projects 76 Wins for the Milwaukee Brewers in 2017

That’d be an improvement over last year!

by Kyle Lesniewski

 

***

 

Will Leitch has a concise team-by-team easy-to-digest look with his over/under takes.

[adambr2]

None of the projections I've seen seem very bullish on Villar, but these look better than the ones from Fangraphs mostly.

 

They see us being a good defensive team which will be a nice change. Strange to me that Broxton is projected for a 1.7 WAR in only 352 PAs, barely half a season's worth. The extrapolation is nice but if he's healthy and productive I see no reason he won't get 600 PAs this year.

[Ennder]

I think PECOTA struggles to give proper credit to bullpens. That is why it fails on teams like the Royals and Orioles so consistently.

[JohnBriggs12]

If their projections for individual Brewer players were accurate, this team wouldn't win 60 games, much less 76. Totally incompatible. They have virtually every Brewer hitter performing worse in 2017 than they did in 2016 with the exception of Arcia.

 

They particularly don't like Villar and Perez, who I think are two of the best offensive weapons they have.

[adambr2]

If their projections for individual Brewer players were accurate, this team wouldn't win 60 games, much less 76. Totally incompatible. They have virtually every Brewer hitter performing worse in 2017 than they did in 2016 with the exception of Arcia.

 

They particularly don't like Villar and Perez, who I think are two of the best offensive weapons they have.

 

PECOTA doesn't project out every player in the league only to make arbitrary win/loss projections. They're based on player performance. True, the batting average projections aren't pretty, but in a a lot of cases the WARP is over what I would predict, probably in many cases based on defense.

 

I draw encouragement from some of this actually to the thought of if these projections add up to 76 wins, I wonder how much better we'd be if some our pitchers actually do carry a sub 4.00 ERA and we get a breakout or two in the lineup.

[And That]

The most accurate player performance projections have typically been the average of the major systems... PECOTA, Marcel, ZIPS, Steamer, and anything else you'd trust to be relevant. It would be fun to see the Brewers projections in that way, if someone was inclined to put it together.

[JosephC]

If their projections for individual Brewer players were accurate, this team wouldn't win 60 games, much less 76. Totally incompatible. They have virtually every Brewer hitter performing worse in 2017 than they did in 2016 with the exception of Arcia.

 

They particularly don't like Villar and Perez, who I think are two of the best offensive weapons they have.

 

I tend to agree.

 

I realize the methodology here and if the WARP's are added up it would get the team to the area of 75/76 wins.

 

But if the Brewers don't have a single superstar position player, only 3 position players that hit the 2 WARP level, and only 1 pitcher that exceeds 1 WARP...under those circumstances it's hard for me to believe the team could even reach the 70 win plateau.

[MrTPlush]

Ignoring the WAR projections the stats seem pretty accurate to me. Some will likely do better than those numbers, but overall second worst offensive team in average? Yah I believe it. I'm not as bullish on that great defensive prediction though and I think Villar will be great again.

[3and2Fastball]

Not sure which I like more, that Pecota has the Brewers improving to 76 wins, or that the Cardinals will be tied with us at 76 wins!

[FVBrewerFan]

Thing is, it's difficult at best to come up with realistic for any if the Brewers position players other than Braun and Scooter. Thames is coming over from Korea. Arcia will get his 1st full season. Santana hasn't been able to put in a full season. Villar and Broxton burst on the scene, is that the trend or will they fall back significantly?

 

Just really tough to project players with minimal MLB experience, so while I always find this stuff to be interesting reading I don't take it too seriously.

[Thurston Fluff]

The most accurate player performance projections have typically been the average of the major systems... PECOTA, Marcel, ZIPS, Steamer, and anything else you'd trust to be relevant. It would be fun to see the Brewers projections in that way, if someone was inclined to put it together.

 

I thought the same thing. Nate silver seems like a place to start an inquiry. He's a baseball guy who pretty much started the aggregate polling trend in politics. I haven't gone there for a while but Nate Silver's 538 website used to have a place to suggest topics of inquiry and I know they had a sports section. This seems like something right up their ally.

[Ennder]

The most accurate player performance projections have typically been the average of the major systems... PECOTA, Marcel, ZIPS, Steamer, and anything else you'd trust to be relevant. It would be fun to see the Brewers projections in that way, if someone was inclined to put it together.

 

I thought the same thing. Nate silver seems like a place to start an inquiry. He's a baseball guy who pretty much started the aggregate polling trend in politics. I haven't gone there for a while but Nate Silver's 538 website used to have a place to suggest topics of inquiry and I know they had a sports section. This seems like something right up their ally.

 

I don't know if you knew this but Nate Silver actually created PECOTA. He might not be the least biased guy out there to test this.

 

It is also a really hard thing to track. You can nail a players WAR perfectly but come about it from completely different stats than you expected. That isn't really a win for your system. Projections are also a wide range of predictions turned into an aggregate. So you can be 'wrong' yet still be right very easily.

[Thurston Fluff]

The most accurate player performance projections have typically been the average of the major systems... PECOTA, Marcel, ZIPS, Steamer, and anything else you'd trust to be relevant. It would be fun to see the Brewers projections in that way, if someone was inclined to put it together.

 

I thought the same thing. Nate silver seems like a place to start an inquiry. He's a baseball guy who pretty much started the aggregate polling trend in politics. I haven't gone there for a while but Nate Silver's 538 website used to have a place to suggest topics of inquiry and I know they had a sports section. This seems like something right up their ally.

 

I don't know if you knew this but Nate Silver actually created PECOTA. He might not be the least biased guy out there to test this.

 

It is also a really hard thing to track. You can nail a players WAR perfectly but come about it from completely different stats than you expected. That isn't really a win for your system. Projections are also a wide range of predictions turned into an aggregate. So you can be 'wrong' yet still be right very easily.

 

Since stats are stats and he is a stats guy a bias to his system wouldn't really be a problem. All he would be doing is crunching the numbers of different systems and seeing what comes out the other end. It would take a while to see how accurate it is vs single systems but I would think anyone who created a system would like the chance to see how their system compares to an aggregate of all systems. It's more his experience aggregating different means of measuring the same thing that I think would help make sense of all the different systems.

[3and2Fastball]

Teams on the way back up with young talent are the toughest to predict.

 

Every prediction formula I've seen has Trea Turner regressing a lot. .290 to .300 with 15 HR's and 40 SB's. The guy hit .342 with 13 HR's 33 SB's last year in 77 games, and he's gonna bat lead-off playing for a manager (Dusty Baker) who loves to run. At the very least I think the SB predictions are way low. Very tough to predict though. And a lot of the young Brewers are like that.

 

You could predict that Arcia is gonna hit .275 with 20 HR's or hit .225 with 9 HR's .... either would have a credible argument, to me