Examining UZR

[757690]

Since this is being discussed in numerous threads, I figured this would be a good time to start an overall UZR thread.

Clearly there is a strong disagreement over UZR and how it should be used, especially it’s place in fWAR. I think it’s a huge improvement over previous attempts evaluate defense production and skill level, but it still has major problems. Here are what I think are the major problems with UZR. These are just opinions and I welcome anyone to show me where I am wrong:

1. This has been discussed before, but UZR doesn’t take into account positioning. As we have all seen, fielders positioned properly in regards to where the hitter is likely to hit the ball, save plenty of hits and runs. But UZR ignores these.

A player who starts the play out of his normal position, and easily makes the play because of it, saves a hit, but UZR counts it a ball in his zone, since they count where the fielder starts in determining what a normal zone is for him, not where he would normally play. If he had started where he normally plays, the ball is either a hit, or he catches it out of his zone, and he get credit for an out of the zone play. Either way, he’s penalized for positioning himself properly.

2. Arm rating. This is only used in the outfield, and it penalizes veteran fielders with strong arms and over benefits most new fielders with even decent arms. Veterans with strong arms, almost never get run on. Which means they don’t get many assists. UZR doesn’t give a point to OF’s who prevented runners from taking extra bases with the fear of his arm. Yet that happens far more often than an OF who nails a runner taking an extra base.

With most new OF’s , runners are more likely to run on them, as no one knows exactly how strong their arms are. Therefore, new OF’s get many more assists in their first season than they do afterwards, and therefore, in their first season, they get a much higher arm rating than they deserve.

3. This one I’m not so sure about but I think I’m right about it. UZR gives double points to fielders in valuable positions who have above average range. For instance, SS already get around 6 extra runs added to their WAR, simply for playing SS. I get why, since their UZR is a number that compares them to other SS, so the extra value of being a SS needs to be added.

But the reason why SS are more valuable than say, a 3B, is that SS get more balls hit in and around their zone. Which means that a SS with above average range, will get more chances to make plays out of his zone than a 3B with the same range. A SS gets around 32 more balls hit out of his zone than a 3B, according to my attempt at math and Fangraphs. Which makes sense, since his zone is bigger, a 3B can only go out his zone to his left, back and in, while a SS can go a full circle out of his zone. If both fielders get to 25% more balls out of their zone, the SS will get 8 more balls out of his zone, with the same skill level. That’s almost a run different. So SS with above average range, will have a higher UZR and WAR than a similarly skilled 3B.

[Brutus]

So you're saying it suffers from UZR error?

[jojo]

3. This one I’m not so sure about but I think I’m right about it. UZR gives double points to fielders in valuable positions who have above average range. For instance, SS already get around 6 extra runs added to their WAR, simply for playing SS. I get why, since their UZR is a number that compares them to other SS, so the extra value of being a SS needs to be added.

Position adjustments are not part of the UZR "score". UZR is a rate stat calculated relative to a player's positional peers. UZR/150 is UZR normalized for playing time, i.e. "a full season".

[757690]

Position adjustments are not part of the UZR "score". UZR is a rate stat calculated relative to a player's positional peers. UZR/150 is UZR normalized for playing time, i.e. "a full season".

Thanks for the correction. The position adjustment is only for fWAR. So the UZR rating doesn't have the issue I mentioned, but FWAR does.

[jojo]

Thanks for the correction. The position adjustment is only for fWAR. So the UZR rating doesn't have the issue I mentioned, but FWAR does.

It's not an "issue" with WAR-it's a necessary adjustment.

[757690]

It's not an "issue" with WAR-it's a necessary adjustment.

The adjustment is necessary, however, it seems like fielders playing some valuable positions, who have above average range, are getting rewarded doubly for their range, unfairly adding their fWAR value.

[jojo]

The adjustment is necessary, however, it seems like fielders playing some valuable positions, who have above average range, are getting rewarded doubly for their range, unfairly adding their fWAR value.

How so? The position adjustment is independent of their raw UZR and all players at the position get the same adjustment.

[PuffyPig]

1. This has been discussed before, but UZR doesn’t take into account positioning. As we have all seen, fielders positioned properly in regards to where the hitter is likely to hit the ball, save plenty of hits and runs. But UZR ignores these.

A player who starts the play out of his normal position, and easily makes the play because of it, saves a hit, but UZR counts it a ball in his zone, since they count where the fielder starts in determining what a normal zone is for him, not where he would normally play. If he had started where he normally plays, the ball is either a hit, or he catches it out of his zone, and he get credit for an out of the zone play. Either way, he’s penalized for positioning himself properly.

I would suggest that a player gets positioned where the coaches tell him to get positioned. I"m not sure why a player should get credit for being piositioned correctly when he likely has nothing to do with it. But I may be wrong with respect to who's in charge of positioning.

[757690]

How so? The position adjustment is independent of their raw UZR and all players at the position get the same adjustment.

I tried to explain it here:

But the reason why SS are more valuable than say, a 3B, is that SS get more balls hit in and around their zone. Which means that a SS with above average range, will get more chances to make plays out of his zone than a 3B with the same range. A SS gets around 32 more balls hit out of his zone than a 3B, according to my attempt at math and Fangraphs. Which makes sense, since his zone is bigger, a 3B can only go out his zone to his left, back and in, while a SS can go a full circle out of his zone. If both fielders get to 25% more balls out of their zone, the SS will get 8 more balls out of his zone, with the same skill level. That’s almost a run different. So SS with above average range, will have a higher UZR and WAR than a similarly skilled 3B

[CySeymour]

Good discussion. I'll admit to not being the most well read on the defensive stats, so I appreciate these insights.


Please, carry on.

[jojo]

I tried to explain it here:

The position adjustment wouldn't impact or be impacted by that scenario though.

[757690]

I would suggest that a player gets positioned where the coaches tell him to get positioned. I"m not sure why a player should get credit for being piositioned correctly when he likely has nothing to do with it. But I may be wrong with respect to who's in charge of positioning.

That's an excellent point, and probably why UZR doesn't consider positioning.

However, it does give an advantage to those players who are positioned poorly, since they will have more balls hit out of their zone, and thus more chances to make an out of the zone play. And it penalizes players for being positioned correctly for each hitter, since they will see less balls hit out of their zone.

So basically, if a fielder wants to get a better UZR, he should play out of position.

I'm not sure how to fix it, but it does seem to be a problem.

[jojo]

That's an excellent point, and probably why UZR doesn't consider positioning.

However, it does give an advantage to those players who are positioned poorly, since they will have more balls hit out of their zone, and thus more chances to make an out of the zone play. And it penalizes players for being positioned correctly for each hitter, since they will see less balls hit out of their zone.

So basically, if a fielder wants to get a better UZR, he should play out of position.

I'm not sure how to fix it, but it does seem to be a problem.

If a player wants a better UZR, he should have better range. Being out of position doesn't help him if he can't get balls he should get in his zone as a result.

[jojo]

Since this is being discussed in numerous threads, I figured this would be a good time to start an overall UZR thread.

Clearly there is a strong disagreement over UZR and how it should be used, especially it’s place in fWAR. I think it’s a huge improvement over previous attempts evaluate defense production and skill level, but it still has major problems. Here are what I think are the major problems with UZR. These are just opinions and I welcome anyone to show me where I am wrong:

1. This has been discussed before, but UZR doesn’t take into account positioning. As we have all seen, fielders positioned properly in regards to where the hitter is likely to hit the ball, save plenty of hits and runs. But UZR ignores these.

A player who starts the play out of his normal position, and easily makes the play because of it, saves a hit, but UZR counts it a ball in his zone, since they count where the fielder starts in determining what a normal zone is for him, not where he would normally play. If he had started where he normally plays, the ball is either a hit, or he catches it out of his zone, and he get credit for an out of the zone play. Either way, he’s penalized for positioning himself properly.

2. Arm rating. This is only used in the outfield, and it penalizes veteran fielders with strong arms and over benefits most new fielders with even decent arms. Veterans with strong arms, almost never get run on. Which means they don’t get many assists. UZR doesn’t give a point to OF’s who prevented runners from taking extra bases with the fear of his arm. Yet that happens far more often than an OF who nails a runner taking an extra base.

With most new OF’s , runners are more likely to run on them, as no one knows exactly how strong their arms are. Therefore, new OF’s get many more assists in their first season than they do afterwards, and therefore, in their first season, they get a much higher arm rating than they deserve.

3. This one I’m not so sure about but I think I’m right about it. UZR gives double points to fielders in valuable positions who have above average range. For instance, SS already get around 6 extra runs added to their WAR, simply for playing SS. I get why, since their UZR is a number that compares them to other SS, so the extra value of being a SS needs to be added.

But the reason why SS are more valuable than say, a 3B, is that SS get more balls hit in and around their zone. Which means that a SS with above average range, will get more chances to make plays out of his zone than a 3B with the same range. A SS gets around 32 more balls hit out of his zone than a 3B, according to my attempt at math and Fangraphs. Which makes sense, since his zone is bigger, a 3B can only go out his zone to his left, back and in, while a SS can go a full circle out of his zone. If both fielders get to 25% more balls out of their zone, the SS will get 8 more balls out of his zone, with the same skill level. That’s almost a run different. So SS with above average range, will have a higher UZR and WAR than a similarly skilled 3B.

Have you any data to suggest these assumptions about positioning/arm are actually valid? Also, how much are you assuming each assumption impacts a player's UZR score?

[757690]

If a player wants a better UZR, he should have better range. Being out of position doesn't help him if he can't get balls he should get in his zone as a result.

True. But still, UZR gives players with above average range, who play out of position, a higher rating than players with similar range, who play on position. That seems to be an issue.