Mateo League Leader in Defensive WAR

[DrungoHazewood]

1 hour ago, wildcard said:

Your eyes maybe.

All eyes.

[DrungoHazewood]

1 hour ago, wildcard said:

For the most part defensive metrics suck.   They said Markakis was a below average fielder which was not true.  They didn't like Jones even when he was healthy.    What it comes down to is the defensive metrics are inaccurate.   Offensively stats are much better.

Frobby looks at a bunch of defensively stats because they are so inaccurate he doesn't trust any one of them.

- Statcast OAA measures where a ball is hit and its hang time, how far the defender was from that at the start, whether he made the play, and how often an average MLB player would make the play on that type of hit when he was that far from the ball.  That certainly doesn't suck, it's exactly what you want.

- The other metrics are somewhat more crude, but they are at least systematic and account for a number of biases.

- Your eyes only see part of a subset of plays in the game you are currently watching, and cannot account for how other players would do or any number of biases.

What you're doing is watching a series of runners in the 100m dash one at at time, you've decided that your stopwatch isn't accurate enough so you're going to fall back on watching some of the runners and subjectively deciding who wins by who looks the fastest.

[DrungoHazewood]

1 hour ago, Aristotelian said:

That makes sense. I would say however you could quantify how much of a player's overall WAR comes from their offensive versus defensive value. oWAR/dWAR isn't quite the right name but it's something. I agree runs/outs is better.

A huge problem is that both dWAR and oWAR include position adjustment so they aren't additive.  oWAR + dWAR doesn't equal WAR, but many or even most people don't get that and use the metrics incorrectly.

The second problem is that they encourage people to think about WAR incorrectly, to assume that WAR applies to components rather than total production.

I understand why Sean Foreman did what he did, but it has caused far more confusion than clarity.

[Tony-OH]

3 hours ago, wildcard said:

For the most part defensive metrics suck.   They said Markakis was a below average fielder which was not true.  They didn't like Jones even when he was healthy.    What it comes down to is the defensive metrics are inaccurate.   Offensively stats are much better.

Frobby looks at a bunch of defensively stats because they are so inaccurate he doesn't trust any one of them.

So defensive measure suck because YOUR eyes tell you that the players on your team are better than the metrics tell you?

My eyes told me for years that Markakis was below average in RF and definitely that jones should not be playing CF towards the end of his time in Baltimore. Both of my "eyes assessments" were back up by most defensive metrics.

While defensive metrics like range factor and things like that are just based off raw stats, the new ability to track player movements and produce OAA is pretty darn accurate. Like Frobby posted, you really have to look at all of them.

I mean you can just ignore them all an go off your fan eye balls, which is certainly your right, but the metrics now give you much, much more data to add to your eye balls.

[NCRaven]

3 hours ago, DrungoHazewood said:

- Statcast OAA measures where a ball is hit and its hang time, how far the defender was from that at the start, whether he made the play, and how often an average MLB player would make the play on that type of hit when he was that far from the ball.  That certainly doesn't suck, it's exactly what you want.

- The other metrics are somewhat more crude, but they are at least systematic and account for a number of biases.

- Your eyes only see part of a subset of plays in the game you are currently watching, and cannot account for how other players would do or any number of biases.

What you're doing is watching a series of runners in the 100m dash one at at time, you've decided that your stopwatch isn't accurate enough so you're going to fall back on watching some of the runners and subjectively deciding who wins by who looks the fastest.

All good stuff.  Another factor is that as an Orioles fan, the majority of our baseball viewing is centered on the Orioles.  So, it is difficult to put any specific player into context with the rest of the league.  We are biased.  Not just because we're fans, but because we simply don't see enough other games to truly make an accurate judgment.  Stats help to compare players and put player performance into context.

[NCRaven]

You don't need metrics to enjoy the game.  But, they help you understand it better.

[wildcard]

3 hours ago, Tony-OH said:

So defensive measure suck because YOUR eyes tell you that the players on your team are better than the metrics tell you?

My eyes told me for years that Markakis was below average in RF and definitely that jones should not be playing CF towards the end of his time in Baltimore. Both of my "eyes assessments" were back up by most defensive metrics.

While defensive metrics like range factor and things like that are just based off raw stats, the new ability to track player movements and produce OAA is pretty darn accurate. Like Frobby posted, you really have to look at all of them.

I mean you can just ignore them all an go off your fan eye balls, which is certainly your right, but the metrics now give you much, much more data to add to your eye balls.

Markakis may not have been the fastest OFer but he was way better than average but they said he was below average.   Defensive metrics were horrible to him.  Jones was not has good later in his career but he was really good when healthy in his prime but you would not know it from the defensive metrics.   

[Frobby]

6 hours ago, wildcard said:

For the most part defensive metrics suck.   They said Markakis was a below average fielder which was not true.  They didn't like Jones even when he was healthy.    What it comes down to is the defensive metrics are inaccurate.   Offensively stats are much better.

Frobby looks at a bunch of defensively stats because they are so inaccurate he doesn't trust any one of them.

I think OAA is probably the best thing out there.   It didn’t exist when Markakis and Jones were in their prime.  For outfielders, though, it only accounts for catching the ball, not throwing it.  I suspect that Fangraphs’ defense stat, which now uses OAA but also factors in arm, is probably pretty good.   But, we’ll need a few years to know.   

My main problem with the advanced stats is when they point in opposite directions.   Then I’m uncertain which to believe.  But when they all point the same way I am very confident in them.  
 

[wildcard]

9 minutes ago, Frobby said:

I think OAA is probably the best thing out there.   It didn’t exist when Markakis and Jones were in their prime.  For outfielders, though, it only accounts for catching the ball, not throwing it.  I suspect that Fangraphs’ defense stat, which now uses OAA but also factors in arm, is probably pretty good.   But, we’ll need a few years to know.   

My main problem with the advanced stats is when they point in opposite directions.   Then I’m uncertain which to believe.  But when they all point the same way I am very confident in them.  
 

Did you read what you just wrote. OAA does not factor in the arm of outfielders.  And you suspect  Fangraph does.   But it does not sound like you know what is factored in.  And you are knowledgable on this stuff.   If you don't know how it is calculated  you  are just guessing and maybe so are they.

[DrungoHazewood]

1 hour ago, NCRaven said:

All good stuff.  Another factor is that as an Orioles fan, the majority of our baseball viewing is centered on the Orioles.  So, it is difficult to put any specific player into context with the rest of the league.  We are biased.  Not just because we're fans, but because we simply don't see enough other games to truly make an accurate judgment.  Stats help to compare players and put player performance into context.

I say it's utterly and completely impossible for anyone to accurately judge the defensive performances of all 1000ish major league players and even bucket them into letter grade categories, much less anything more granular. On TV the camera isn't paying attention to any particular fielder before the pitch, so you can't tell positioning. You can't tell jumps. And no human can watch a substantial part of every teams games in person and pay attention to individual defensive efforts.

Without the metrics your hardly more informed than old Gold Glove voters being asked to free-associate the best fielders at the end of the season. First base... Palmeiro!  Uh, sir, he DH'd all year... Well then, assistant coach, fill out this form, I have to go... do... stuff...

[DrungoHazewood]

53 minutes ago, wildcard said:

Markakis may not have been the fastest OFer but he was way better than average but they said he was below average.    

Markakis was really good but the metrics say he was below average.  How do you know he was really good?  Because I said so. Why then do the metrics say he was below average?  Because they suck, because I said so.

[DrungoHazewood]

18 minutes ago, wildcard said:

Did you read what you just wrote. OAA does not factor in the arm of outfielders.  And you suspect  Fangraph does.   But it does not sound like you know what is factored in.  And you are knowledgable on this stuff.   If you don't know how it is calculated  you  are just guessing and maybe so are they.

But you definitely know all the metrics suck, and you know they suck because you knew how good everyone was before looking at the metrics and sometimes the metrics are different than your opinion, so they suck. 

The best way to validate anything is to compare it to what you currently believe, and if it's different it's wrong.

[Hallas]

8 hours ago, Aristotelian said:

Why can't dWAR/oWAR be isolated? Aren't wins just a certain number of runs?

From a practical standpoint you can't really have a replacement level fielder for a couple reasons: 1: replacement-level fielders are usually demoted to less important/less challenging defensive positions until they find one that they are close to average at. 2: defense is almost never a primary reason a player gets kicked off the team for a AAAA player.  It's pretty much exclusively offense, or at the very least offense in conjunction with the player's positional value, that dictates player replacement.

[Hallas]

4 hours ago, Tony-OH said:

So defensive measure suck because YOUR eyes tell you that the players on your team are better than the metrics tell you?

My eyes told me for years that Markakis was below average in RF and definitely that jones should not be playing CF towards the end of his time in Baltimore. Both of my "eyes assessments" were back up by most defensive metrics.

While defensive metrics like range factor and things like that are just based off raw stats, the new ability to track player movements and produce OAA is pretty darn accurate. Like Frobby posted, you really have to look at all of them.

I mean you can just ignore them all an go off your fan eye balls, which is certainly your right, but the metrics now give you much, much more data to add to your eye balls.

Was that your assessment of Markakis early career?  I thought he was slightly above average for at least the first 4 or so years of his career and kind of fell off a cliff defensively after that because his range went to basically zero, and he no longer had the arm to make up for it, and his arm wasn't really being challenged anymore.  I kind of wonder if the herniated disc was bothering him before he actually got surgery on it.

[NelsonCruuuuuz]

13 hours ago, Frobby said:

I don’t trust any one advanced defensive metric, so I like to look at all of them:

Rtot: +4

Rdts: +9 (note: this stat feeds into dWAR)

UZR: +2.1

OAA: +4 (translates to +3 RAA)

Fangraphs def. +5.5 (7th in MLB).

To explain just a little, dWAR is a combination of positional adjustment, and how many defensive runs saved at a position.   Mateo gets a positional adjustment of +2 and Rdrs (also known as Rfield) of +9.   That gives him +11 runs which translates to +1.2 wins (dWAR).   

Over the winter, Fangraphs changed its Fangraphs defense metric to use Statcast’s OAA to measure range.   Fangraphs defense, like dWAR, has a positional adjustment.  It also has a separate component for infielders for efficiency in turning double plays.   He’s 7th in defense as measured by Fangraphs, 4th among shortstops.

So, there is some variance in how good Mateo’s defense has been depending on what stats you prefer.   However, the good news is that all these stats are solidly in the plus column.   Often, you’ll see them pointing in different directions.   So, you can have a pretty high degree of confidence that Mateo has been a plus defender to this point of the season.  And I for one wasn’t really expecting that.
 

 

Thanks Frobby.