Impact of Defense

[jay]

Because we can be absolutely positive it's not the measurement system?

Sure, in the most literal sense especially for HRs. Offensive counting stats reflect what happened, but even most of that measurement system has some degree of interpretation. Was it a strike or a ball? Was it a hit or an error? Was it a double or advance on throw? Most of them are easy calls, but then some aren't. And we accept that.

 

My point is that defensive stats are doing the same thing. They are telling us what happened. Just as with the stats we are familiar with from yesteryear, there's going to be variation in those numbers and we shouldn't unduly dismiss defensive metrics based on that. Admittedly, there's some higher degree of interpretation, but I don't think it's as high as some skeptics currently believe. Most of the calls are easy, but then some aren't.

 

However, the measurement system is designed to inherently capture and minimize that. We factually know whether or not the ball was caught at the same level of certainty we have on those offensive stats. The interpretation becomes how hard it was to catch and, again, the majority are easy. Even the hard ones though, we can pretty easily recognize it was a hard catch, so whether it's categorized one way and the guy gets a +.7 run value or another way and he gets a +.6 run value, we aren't talking about some giant gulf. If he missed that ball, we know it was a tough play and he's not getting any significant negative run value assigned.

 

The most important factor is it's all grounded in that we factually know what happened in reality.

[jimmer]

 

It is not that big or impossible. I don't have the exact number handy, but it is less than a run. Maybe 0.7 runs for a hit to the outfield? 

A typical outfield hit is worth around .56 runs and any batted ball out is worth around -.27 runs, so the difference between a hit and an out is worth around .83 runs.

[jimmer]

I often wonder why OPS is so widely accepted as fact over other metrics, not that it didn't take a long time for OPS to stop being discarded.

 

OPS is a stat that is widely accepted as a way to value a player's offensive contributions, but it has assumptions as well by the way it values slg % in comparison to OBP.  It assumes they are equal when OBP has been proven to be roughly twice as important as SLG% in scoring runs.

 

The calculation of slg% also assumes that a double is twice as important as a single, a triple three 3 times as important as a single and 50% more valuable than a double, and so on.  It's an easy assumption to accept, the math is so simple and hey, a guy is on 2B instead of 1B, but that doesn't make the assumption that a double is twice as valuable as a single true.

[Riverbrian]

 

That is an argument for defensive WPA Win Probability Added. A diving catch with the bases loaded and two outs is more valuable than an identical catch with no one on. Interesting, but not really a good way to estimate fielding performance or ability.

 

Its hard to quantify the true result of the identical catch with no one on.

 

Its hard to quantify what happens with the bases loaded and two outs as well... Its easy to say "3 runs saved by that catch"  ... but we don't know what the guy hitting behind him would have done. He now gets to reach into that raffle drum. A raffle drum that has 3 "reach base" cards and 7 "out" cards.

 

When the catch happens with zero outs... or the bad jump happens with zero outs... its harder to quantify because there is more baseball left in an inning.

 

You can go through history and come up with an average... 0.4 runs saved on average... for X outs... X base runners... all games... all time.

 

But baseball games are decided when the game blows past the average.  

 

In my opinion... when a team scores 5 runs in an inning on you... Its usually game over.

 

When you look back at those big innings... You can usually find a play where a defender could have contributed to stopping it... preventing it. By moving his feet... by not laying up on a soft liner... by not double clutching on a DP chance.

 

These moments happen with zero outs and two outs. Most times there is no effect to the game but some times its part of a landslide and those landslides lose games.

 

Its about outs... Give them an extra one and you risk the landslide... Steal an out from them... Its a lot harder for them to string something together.

 

These moments are huge and very hard to quantify. Each landslide is hell on a pitchers ERA.  

 

Apart from Dozier... The Twins don't have players who steal outs defensively. This is a handicap our average pitchers have to overcome and something the Twins need to fix or we will be complaining about pitching for awhile.

[USAFChief]

My point is that defensive stats are doing the same thing. They are telling us what happened.

And therein lies the crux of our discussion. You and many others take your statement as factual. You might very well be proven correct.

 

To me, your statement requires a leap of faith that I'm not ready to make yet. I don't trust the data collection, and I don't fully agree with the design of the measurement systems. I'm troubled by the fact there are multiple measurement systems out there, and they often don't agree. I believe there is more to defense than just range. I doubt the impact on runs from a corner outfielder is as large as a middle infielder at the major league level.

 

And that's fine. We'll both enjoy watching the coming baseball season, and if we end up at a game together I'll buy the first beer*, where we can argue some more in person.

 

*EDIT: and maybe even a hot dog. But no GD sushi. That doesn't belong at a ballpark...can we at least agree on that?

[jimmer]

 

 

We'll both enjoy watching the coming baseball season, and if we end up at a game together I'll buy the first beer*, where we can argue some more in person.

*EDIT: and maybe even a hot dog. But no GD sushi. That doesn't belong at a ballpark...can we at least agree on that?

 

Well, I can absolutely agree with the sushi part.  :-)

[jay]

And therein lies the crux of our discussion. You and many others take your statement as factual.

Do defensive stats tell us what happened? Yes, I take that as factual. Do they accurately tell us what happened? That's what we debate.

 

And while I love sushi, I don't love sushi while watching baseball. If I was at an NPB game though... it'd get real interesting.

[Squirrel]

Well, I can absolutely agree with the sushi part. :-)

Ahhhhhhh, get through this entire thread, and this is the common ground you all get to?

[jimmer]

 

Ahhhhhhh, get through this entire thread, and this is the common ground you all get to?

Chief and I were both senior NCOs in the AF.  I imagine the list of things we agree on far outweigh the list of things we don't.  Just have to find an advanced metric to test that theory now.  I hate homework ;-)

[jimmer]

If you knew sushi, like I know sushi :-)

[Riverbrian]

 

It is not that big or impossible. I don't have the exact number handy, but it is less than a run. Maybe 0.7 runs for a hit to the outfield? And the fielder only gets debited that full amount from his UZR if the average player at his position coverted it into an out at virtually 100% rate.

That's not why Gutierrez went from +30 to 0 UZR. He would have needed to miss 40 such balls in 2010 over 2009 for that to be true.

 

Gutierrez went from +30 to +5 one year to the next because there isn't enough non routine plays to stabilize the stat in a given year. 

 

Yet the stat is quoted by the year and used to hang players by many year to year... And the stat is folded in to WAR and WAR is used to hang or glorify players.

 

If UZR is the best we got... We ain't got much. My opinion... nothing more... nothing less.

 

I've looked into because I wanted to believe in it. I'm a defensive freak to a fault.  

 

 

 

 

[Brock Beauchamp]

 

Gutierrez went from +30 to +5 one year to the next because there isn't enough non routine plays to stabilize the stat in a given year. 

 

Yet the stat is quoted by the year and used to hang players by many year to year... And the stat is folded in to WAR and WAR is used to hang or glorify players.

 

If UZR is the best we got... We ain't got much. My opinion... nothing more... nothing less.

 

I've looked into because I wanted to believe in it. I'm a defensive freak to a fault.  

Yep, this.

 

Torii Hunter's DRS in 2012-2014 is the following:

 

15 (hey, that's really good), -10 (ugh), -18 (hide the children).

 

It's possible, maybe even likely, that Torii Hunter's defense has declined to the point that he's horrible in the outfield.

 

What's equally possible is that DRS is completely off its rocker and Hunter is closer to an average fielder than any of us realize.

 

Yet I've heard the phrase "worst outfielder in baseball" bandied about on this forum at least 100 times in the past three months. Hey, it's possible that Torii is now the worst fielder in baseball. It's also possible - maybe even likely - that he's not because DRS (and UZR, which is generally pretty close to DRS) cannot be trusted at all over a single season, yet we use those metrics all the time over a single season and fold them into other metrics that are also quoted as single-season numbers.

[Otto von Ballpark]

Gutierrez went from +30 to +5 one year to the next because there isn't enough non routine plays to stabilize the stat in a given year.

 

That still assumes Gutierrez converted notably more nonroutine plays into outs in 2009 than 2010. It might very well be due to sample and opportunity, but I still think it is interesting to know. Maybe more years of this data would suggest which year was more the outlier.

[USAFChief]

Yep, this.

 

Torii Hunter's DRS in 2012-2014 is the following:

 

15 (hey, that's really good), -10 (ugh), -18 (hide the children).

 

It's possible, maybe even likely, that Torii Hunter's defense has declined to the point that he's horrible in the outfield.

 

What's equally possible is that DRS is completely off its rocker and Hunter is closer to an average fielder than any of us realize.

 

Yet I've heard the phrase "worst outfielder in baseball" bandied about on this forum at least 100 times in the past three months. Hey, it's possible that Torii is now the worst fielder in baseball. It's also possible - maybe even likely - that he's not because DRS (and UZR, which is generally pretty close to DRS) cannot be trusted at all over a single season, yet we use those metrics all the time over a single season and fold them into other metrics that are also quoted as single-season numbers.

If we choose Inside Edge Fielding as our metric du jour, Hunter is just fine. Maybe even slightly above average in 2014.

[Otto von Ballpark]

Yep, this.

 

Torii Hunter's DRS in 2012-2014 is the following:

 

15 (hey, that's really good), -10 (ugh), -18 (hide the children).

 

It's possible, maybe even likely, that Torii Hunter's defense has declined to the point that he's horrible in the outfield.

 

What's equally possible is that DRS is completely off its rocker and Hunter is closer to an average fielder than any of us realize.

 

Yet I've heard the phrase "worst outfielder in baseball" bandied about on this forum at least 100 times in the past three months. Hey, it's possible that Torii is now the worst fielder in baseball. It's also possible - maybe even likely - that he's not because DRS (and UZR, which is generally pretty close to DRS) cannot be trusted at all over a single season, yet we use those metrics all the time over a single season and fold them into other metrics that are also quoted as single-season numbers.

It is tough to respond to this because I didn't call Hunter the worst outfielder in baseball. If someone does and cites UZR as proof, then that person is obviously misusing/misapplying it.

 

I still don't understand what you want UZR, or people who reasonably consider UZR as one of several inputs for their evaluations, to do about it.

[jimmer]

 

If we choose Inside Edge Fielding as our metric du jour, Hunter is just fine. Maybe even slightly above average in 2014.

How is that? I guess I'm reading it wrong.  I looked at Inside Edge Fielding at Fangraphs (since they've had it there since last year) and for qualified RF he was 16th out of 16.  Then I extended it to OF who had 700 or more innings and he was 21st out of 21. Then I extended it to 600 innings and he was 26th out of 26. To 500, he was 31st out of 31.

[Riverbrian]

 

That still assumes Gutierrez converted notably more nonroutine plays into outs in 2009 than 2010. It might very well be due to sample and opportunity, but I still think it is interesting to know. Maybe more years of this data would suggest which year was more the outlier.

 

I agree... I think both years were probably outliers. I think it all comes down to opportunity to do something out of the norm.

 

BTW... Using eye test data... I've always thought that Gutierrez was one of the few players that lowered Pitchers ERA's by himself.

 

When he had a 30 UZR... I was like... Damn Right... this stat agrees with me... It rules.  

 

I just can't anymore...

 

However... I don't want it thrown away... I want it... it's info...  I just want it tempered... and I don't like players being hung by it year to year knowing the sample size isn't sufficient.

 

Perhaps the biggest problem with UZR is this: The data that makes up the stat isn't transparent for the average fan to recreate and this leads to massive misunderstanding and misuse..

 

 

[jimmer]

In the case of Hunter, it's not hard to believe when a defender gets to his age (and with his type of build combined with that age), his defensive ability starts going down at a pretty good rate.  I'm not sure how he'd be immune to it when others haven't.

 

Dyson is only 30 and some are talking about his best tool, speed (which helps his offense and defense), is going to drop dramatically soon which will affect his value.  His build is way better for the speed skill than Torii's and he's a decade younger.  Why is it hard to believe Hunter just isn't a good defender anymore?  I think too many Twins fans just have bias towards Hunter due to what he used to be able to do on defense and it hurts to hear Spiderman isn't able to avoid the pitfalls of aging everyone goes through.

[Brock Beauchamp]

 

It is tough to respond to this because I didn't call Hunter the worst outfielder in baseball. If someone does and cites UZR as proof, then that person is obviously misusing/misapplying it.

I still don't understand what you want UZR, or people who reasonably consider UZR as one of several inputs for their evaluations, to do about it.

I'm not accusing you of doing any such things, for the record... But that kind of statement is incredibly common, even by some veteran writers around baseball.

 

I'm not sure exactly what I want from UZR. Probably some way for the metric to better use contextual runs saved. If a metric is going to use runs as a measurement, it should find a way to apply those runs into game situations. That should help remove some of the insane outliers and stabilize the metric a bit.

 

For example, if Gutierrez saved a bases loaded double that 90% of center fielders can't reach, he gets 3 runs x .9 (not exactly perfect math, as he also added an out and removed a person from second base, but you get the point). Those were runs he actually saved in a game situation.

 

If people are going to say things like "the Royals defense was X runs better than the Indians in 2014", that would go a long way toward making the number more believable.

[TheLeviathan]

*EDIT: and maybe even a hot dog. But no GD sushi. That doesn't belong at a ballpark...can we at least agree on that?

Split the difference and go with Walleye on a stick.

[Brock Beauchamp]

 

In the case of Hunter, it's not hard to believe when a defender gets to his age (and with his type of build combined with that age), his defensive ability starts going down at a pretty good rate.  I'm not sure how he'd be immune to it when others haven't.

I agree, but Hunter's 2012-2013 is the third time he lost 20+ runs from his DRS in a single season. The two previous times, he regained most of the DRS the following season.

 

It's really hard to trust a metric that has that kind of swing from season to season, particularly when people insist on quoting it in single season usage.

[Willihammer]

 

I'm not accusing you of doing any such things, for the record... But that kind of statement is incredibly common, even by some veteran writers around baseball.

 

I'm not sure exactly what I want from UZR. Probably some way for the metric to better use contextual runs saved. If a metric is going to use runs as a measurement, it should find a way to apply those runs into game situations. That should help remove some of the insane outliers and stabilize the metric a bit.

 

For example, if Gutierrez saved a bases loaded double that 90% of center fielders can't reach, he gets 3 runs x .9 (not exactly perfect math, as he also added an out and removed a person from second base, but you get the point). Those were runs he actually saved in a game situation.

Isn't that a subjective judgment too? What if 90% of fielders don't catch the ball, but 50% of fielders would at least stop the ball rolling all the way to the wall? What about the speed of the baserunners, the decision of the 3rd base coach to send/hold the last runner?

[jimmer]

 

I'm not accusing you of doing any such things, for the record... But that kind of statement is incredibly common, even by some veteran writers around baseball.

 

For example, if Gutierrez saved a bases loaded double that 90% of center fielders can't reach, he gets 3 runs x .9 (not exactly perfect math, as he also added an out and removed a person from second base, but you get the point). Those were runs he actually saved in a game situation.

 

If people are going to say things like "the Royals defense was X runs better than the Indians in 2014", that would go a long way toward making the number more believable.

I think you're talking about WPA.  Spycake wrote earlier on this thread:

 

That is an argument for defensive WPA Win Probability Added. A diving catch with the bases loaded and two outs is more valuable than an identical catch with no one on. Interesting, but not really a good way to estimate fielding performance or ability.

 

[Brock Beauchamp]

 

Isn't that a subjective judgment too? What if 90% of fielders don't catch the ball, but 50% of fielders would at least stop the ball rolling all the way to the wall? What about the speed of the baserunners, the decision of the 3rd base coach to send/hold the last runner?

It's less subjective if Field f/X is generating the data.

 

Nothing is going to be perfect but the more human elements that can be removed from a metric, the more accurate the metric should be in the long run (and it'll take years to accumulate enough data to really make it accurate).

[Willihammer]

 

It's less subjective if Field f/X is generating the data.

 

Nothing is going to be perfect but the more human elements that can be removed from a metric, the more accurate the metric should be in the long run (and it'll take years to accumulate enough data to really make it accurate).

Fieldf/x will make the catchability % more precise but it what you are talking about with taking actual game situations and assigning run values based on that seems just as error prone, perhaps moreso, than using weighted values.

[Otto von Ballpark]

However... I don't want it thrown away... I want it... it's info... I just want it tempered... and I don't like players being hung by it year to year knowing the sample size isn't sufficient.

 

Perhaps the biggest problem with UZR is this: The data that makes up the stat isn't transparent for the average fan to recreate and this leads to massive misunderstanding and misuse..

Definitely agreed it should be tempered. Personally, if I want to estimate a player's defense, I look at multiple metrics, try to discount outliers a bit, apply a + or - 10 runs, check it against scouting reports, etc.

 

To your latter point, I find it an understandable annoyance than a problem that a detailed accounting of every player's UZR isn't published. To be useful, it would have to show the low level data collected, which isn't free and thus can't be shared freely. MLB was awesome to release PitchFX data, but I have a feeling they may not be so generous with the fielding equivalent. So the problems of transparency might get worse before they get better.

[USAFChief]

How is that? I guess I'm reading it wrong.  I looked at Inside Edge Fielding at Fangraphs (since they've had it there since last year) and for qualified RF he was 16th out of 16.  Then I extended it to OF who had 700 or more innings and he was 21st out of 21. Then I extended it to 600 innings and he was 26th out of 26. To 500, he was 31st out of 31.

Perhaps I'm mistaken...got a link?

[TheLeviathan]

If people are going to say things like "the Royals defense was X runs better than the Indians in 2014", that would go a long way toward making the number more believable.

Personally, I don't want anything from UZR. I think what it tries to do with the limitations we have measuring defense is fine, IF (big if) it is used properly. Fangraphs states pretty clearly that using it over larger samples greatly increases it's usefulness, yet stating KC saved X runs over Cleveland in one season betrays their very recommendation. Ditto using WAR alone to say Trout > Cabrera for MVP.

 

I like looking at UZR and WAR as ranking tools over 3+ seasons, which is sort of what we're told they are for and best at. So why does the saber community so frequently not follow their own advice on defensive metrics?

[jimmer]

 

Perhaps I'm mistaken...got a link?

http://www.fangraphs.com/leaders.aspx?pos=rf&stats=fld&lg=all&qual=y&type=3&season=2014&month=0&season1=2014&ind=0&team=0&rost=0&age=0&filter=&players=0

 

The link is already set to Inside Edge Fielding for RF.

[Otto von Ballpark]

I'm not sure exactly what I want from UZR. Probably some way for the metric to better use contextual runs saved. If a metric is going to use runs as a measurement, it should find a way to apply those runs into game situations. That should help remove some of the insane outliers and stabilize the metric a bit.

 

For example, if Gutierrez saved a bases loaded double that 90% of center fielders can't reach, he gets 3 runs x .9 (not exactly perfect math, as he also added an out and removed a person from second base, but you get the point). Those were runs he actually saved in a game situation.

 

If people are going to say things like "the Royals defense was X runs better than the Indians in 2014", that would go a long way toward making the number more believable.

That is Win Probability Added, which already exists for hitters and pitchers but is very hard to calculate for defenders.

 

And while that might help balance a ledger, it is worse at evaluating skill and performance (see the numerous criticisms of WPA for evaluating hitters/pitchers already).