Twins 2019 Defensive Metrics

[ashbury]

SABR's mid-season defensive stats are out, and some Twins are looking really good in them.

 

Byron Buxton leads all American Leaguers in the aggregate SABR Defensive Index (SDI) score. Max Kepler isn't too far behind. Right with him, and causing me to eat some crow, Marwin is shining in his efforts at third base. A little further down, Jorge Polanco ranks among the league leaders also. Martin Perez is showing well, among pitchers.

 

The entire listing, both tops by league and also rankings by position, are here: https://sabr.org/sdi/2019-07-07

 

I'm here this morning to celebrate our guys - or critique them if need be. If you are skeptical of current attempts to quantify defensive contributions, please start your own thread - this is a thread about defense so I'm entitled to be defensive.

 

[Darius]

We need Buxton back. Our record with vs without him is pretty incredible (under .500 without him, insanely good with him).

 

I’m really impressed with how Poland’s has evolved defensively. His inclusion is a little surprising to me.

[ashbury]

We need Buxton back. Our record with vs without him is pretty incredible (under .500 without him, insanely good with him).

This has been true in 2018 (bad year) and 2017 (good year) as well. The only reason we know this, unfortunately, is that he misses so many games!

 

I’m really impressed with how Poland’s has evolved defensively. His inclusion is a little surprising to me.

 

Isn't Auto-correct fun? I agree completely about our SS. I was very skeptical about his arm, and I still cringe when he throws because he looks a little funny, but it's now the case that he passes the eye-test where it comes to making the plays, and he didn't before, and the SDI number bears out what I think I'm seeing, so it's all good!

[obtusebanter]

They got Garver a couple points better than Castro at catcher?  I only follow gamecasts.  Does that track?

[wavedog]

Very interesting - thanks for posting - my biggest .  My biggest wow so far is Robby Grossman at the top of the leaderboard for Left-Fielders.  

[Danchat]

Very interesting - thanks for posting - my biggest . My biggest wow so far is Robby Grossman at the top of the leaderboard for Left-Fielders.

Lol what?! His 2016 season was one of the worst an OF has ever had IMO.

[ashbury]

They got Garver a couple points better than Castro at catcher?  I only follow gamecasts.  Does that track?

Yeah, I thought that was an eye-opener too. I don't know what the "points" actually translate to^*, but they have Mitch a bit above-average, in whatever it is they measure for catchers, and Jason a bit below-average. I actually think this may be reasonable, if you compare to the catchers on other teams close to them in the rankings (e.g. the Red Sox' two guys). But then I'm not really the biggest woofer in the world for Jason Castro's defense overall, so it's not a big leap for me to think Garver is about on a par and perhaps even a little better by now. I know that Garver still has his detractors, and there is just about one D'oh moment for him every game. But catcher's a demanding position, and a true clown back there would be pretty unbearable, and Garver is beyond that modest hurdle.

 

 

^{* _{Actually I think I do know that it's a mishmash of numbers, somewhat like OPS is, making it "just a number", though I think it's trying to look on a par with "wins".}}

[ashbury]

Very interesting - thanks for posting - my biggest .  My biggest wow so far is Robby Grossman at the top of the leaderboard for Left-Fielders.  

The CF he plays alongside, Ramon Laureano, is ranked very low (a fine play against us the other night notwithstanding). That could still be a flaw in the methodology, even though I know they try to correct for this obvious interaction among fielders. Grossman in his years with us was paired with pretty capable centerfielders.

[jkcarew]

Thanks for posting, Ash. My take...the advanced defensive metrics remained flawed, fatally so, in some cases. They just have a long, long, way to go. Still, probably nothing that more data (way more data) won't be able to solve down the road.

[Doomtints]

SDI is rubbish, but it's nice to see the Twins well represented.

[Doomtints]

Thanks for posting, Ash. My take...the advanced defensive metrics remained flawed, fatally so, in some cases. They just have a long, long, way to go. Still, probably nothing that more data (way more data) won't be able to solve down the road.

 

SDI just adds three existing defensive metrics together to produce an index. This double- or triple- counts fielding events that are represented in more than one metric.

 

SDI is based on a subset of Chris Dial's theories. The development team otherwise included SABR brass and SABR members who have been masterful at collecting baseball data, so in my opinion they missed having a critical eye on it.

 

With the team they had, it would have been possible to create an aggregate score which does not double-count/inflate events which skew the scores (either up or down), simply by modifying some of the metrics to not include the duplicates, but for some reason they chose not to do so.

 

You will note that neither fangraphs nor baseball-reference track SDI.

[ashbury]

You will note that neither fangraphs nor baseball-reference track SDI.

It's a conversation starter.