Revere trade revisited

[cmathewson]

UZR from year to year is a ping pong ball. Yet... UZR is the most quoted defensive stat. You don't see people using 3 year samples. They use less than a year of data and try to pass it off as gospel... I don't belong to that church.

 

UZR stats for 2014 are telling me that Trevor Plouffe (2.1) has more range than Brett Lawrie (-0.1)

 

Carlos Gomez went from 4.8 to 24.4 to -0.4... From 2012 to 2014.

 

Defensive Metrics will mean a lot more to me when they stabilize. They don't mean much until they do.

 

It's not a perfect stat, but it's the best evidence we have in the absence of watching him play everyday. Do you watch all the Phillies games? I didn't think so.

 

UZR fluctuates as does average. It depends on luck. A guy could get a lot of hard hit balls hit in his zone. But, like batting average, large samples tend to reveal true ability. Revere has put in a lot of games in the Phillies outfield. And his numbers have not fluctuated. It's possible he's been extremely unlucky. But with each passing game, that become less likely.

[diehardtwinsfan]

It's not a perfect stat, but it's the best evidence we have in the absence of watching him play everyday. Do you watch all the Phillies games? I didn't think so.

 

UZR fluctuates as does average. It depends on luck. A guy could get a lot of hard hit balls hit in his zone. But, like batting average, large samples tend to reveal true ability. Revere has put in a lot of games in the Phillies outfield. And his numbers have not fluctuated. It's possible he's been extremely unlucky. But with each passing game, that become less likely.

 

Cmath, I highly doubt you watch all of the Phillies games either, so I'm not sure how this makes you more qualified than RiverBrian to talk about the pro/con of UZR.

 

I don't think that watching every game is a prequisite to being able to distrust or trust UZR and other defensive metrics, especially given the fact that there's quite a bit of disagreement in the SABER world on the value of these stats as well. It's not a whole lot different than you and I saying we should not place value on stats like Wins, Losses, and RBIs.

 

I'd caution you a bit here. RB is using the eye-test from his own observations, and you are not. Baseball uses both approaches, and the general consensus of Twins fans watching Revere during his time here was that he was a pretty good defender. I have no problem with you choosing to disagree, but there's no reason to be condescending about it.

 

Thanks,

[cmathewson]

Cmath, I highly doubt you watch all of the Phillies games either, so I'm not sure how this makes you more qualified than RiverBrian to talk about the pro/con of UZR.

 

I don't think that watching every game is a prequisite to being able to distrust or trust UZR and other defensive metrics, especially given the fact that there's quite a bit of disagreement in the SABER world on the value of these stats as well. It's not a whole lot different than you and I saying we should not place value on stats like Wins, Losses, and RBIs.

 

I'd caution you a bit here. RB is using the eye-test from his own observations, and you are not. Baseball uses both approaches, and the general consensus of Twins fans watching Revere during his time here was that he was a pretty good defender. I have no problem with you choosing to disagree, but there's no reason to be condescending about it.

 

Thanks,

 

Fine, but you missed the point. I presumed that none of us watch the Phillies games. So in the absence of that evidence, what do we have to go on? UZR data. If that is off the table, we can't really have a discussion based on any evidence at all. "I think Revere is a great fielder, but I refuse to use data to draw that conclusion, and I haven't seen him play in two years" is not something I can either agree with or disagree with.

[Brock Beauchamp]

Fine, but you missed the point. I presumed that none of us watch the Phillies games. So in the absence of that evidence, what do we have to go on? UZR data. If that is off the table, we can't really have a discussion based on any evidence at all. "I think Revere is a great fielder, but I refuse to use data to draw that conclusion, and I haven't seen him play in two years" is not something I can either agree with or disagree with.

 

In all fairness, we've all seen Revere play the outfield many, many times. Yes, it's possible that he has lost a step and his defense has suffered with the Phillies.

 

It's also possible that UZR is very wrong in what is less than a season's worth of data (and UZR fluctuates wildly from year to year and should not be the sole metric used to qualify defensive acumen over a single season of play).

 

Revere has played less than 130 games for the Phillies.

[diehardtwinsfan]

I personally don't understand how UZR can fluxuate that much over an entire season. I would think given the number of balls hit to a players zone over the course of the season would even things out... unless the metric doesn't account for things like grounders through the infield, or liners off the wall...

[tobi0040]

I personally don't understand how UZR can fluxuate that much over an entire season. I would think given the number of balls hit to a players zone over the course of the season would even things out... unless the metric doesn't account for things like grounders through the infield, or liners off the wall...

 

Another metric we should look at is RAONA (Runners advanced on noodle arm). I am guessing Revere is leading the league.

[Brock Beauchamp]

I personally don't understand how UZR can fluxuate that much over an entire season. I would think given the number of balls hit to a players zone over the course of the season would even things out... unless the metric doesn't account for things like grounders through the infield, or liners off the wall...

 

I believe it has to do with the nature of defense. If you're capable of playing in an MLB outfield, all players are going to get to the same balls 95% of the time. It's the outlier hits/outs that define a player's defensive ability and those are relatively rare. It takes some time for a player to accumulate enough of them for the metric to define his ability as a defender.

 

In 1+ seasons with the Phillies, Revere has just over 300 chances on defense. Let's say 90% of those balls were gimmes that pretty much any defender would catch. That leaves only 31 balls for a metric to decide if he's a quality defender or not... Obviously, 31 chances does not provide enough data points to be anything close to accurate.

[TheDean]

46% of Revere's innings in CF have been played in Philly. We've discussed how UZR takes a long time to regularize. I do not think that his cumulative UZR in Philly is accurate, but I think his UZR in MN was probably too high. The answer lies somewhere in the middle...around league average, which is exactly what it says (UZR/150 at 1.0 for his career in CF).

 

If we're looking at just his stats in Philly or just in MN, the sample size is too small. Once we combine them, it matches the eye test. It says that he has a noodle arm, good range, and is somewhat error-prone (scores of -4.5, 8.8, -2.6, respectively). This is in agreement with what the fan bases in both cities have seen. There could be an argument for a reduction in range over time since no player can stave off aging, keeping in mind that we remember Revere in his prime. However, too SSS to say that definitively with UZR, which is where a combined Twins/Phillies fan would help (if that exists).

 

I just have a feeling that MN fans are maybe biased by nostalgia, and Philly fans are biased by their saltiness over the trade (which includes their nostalgia for Vanimal ).

[Brock Beauchamp]

If we're looking at just his stats in Philly or just in MN, the sample size is too small. Once we combine them, it matches the eye test. It says that he has a noodle arm, good range, and is somewhat error-prone (scores of -4.5, 8.8, -2.6, respectively). This is in agreement with what the fan bases in both cities have seen. There could be an argument for a reduction in range over time since no player can stave off aging, keeping in mind that we remember Revere in his prime. However, too SSS to say that definitively with UZR, which is where a combined Twins/Phillies fan would help (if that exists).

 

Can't argue with any of this though I will provide one small counterpoint:

 

It's very possible that Revere played a much more valuable centerfield in the cavernous Target Field than he does in thimble-sized Citizen's Bank Park.

[TheDean]

Can't argue with any of this though I will provide one small counterpoint:

 

It's very possible that Revere played a much more valuable centerfield in the cavernous Target Field than he does in thimble-sized Citizen's Bank Park.

 

Good point. So in theory, Revere has fewer opportunities to "flex" his range in Philadelphia compared to playing at TF - therefore his ratings suffer because he's making fewer low-probability plays compared to league average?

I can definitely buy that.

[Brock Beauchamp]

Good point. So in theory, Revere has fewer opportunities to "flex" his range in Philadelphia compared to playing at TF - therefore his ratings suffer because he's making fewer low-probability plays compared to league average?

I can definitely buy that.

 

Yep. It's pure speculation on my part but I think it's a possibility.

[Riverbrian]

I was gonna let this drop and move on with life but OK... One more.

 

If you believe that UZR is quotable... Then OK!!! My wife likes watching reality programming and I don't belong to that church either.

 

However... I stated that I would take Ben back and I meant it... I would... You disagree and so does Brock and I assume others and that's OK as well.

 

However... You responded to my declaration... with UZR metrics and when fed small sample sizes... It means very little to me.

 

In your last post addressed to me...

 

#61. You defend your view by stating "large samples tend to reveal true ability". I agree that there is some truth in that... However...

 

In post #40... You quote UZR/150 for 2014. That's a direct contradiction to what you posted in #61.

 

Without a significant sample size... I'll use my eyes and I'm comfortable with that. I think everyone should.

 

Even with larger sample sizes... I'm not totally comfortable with the stat.

 

My opinion... I don't think zone ratings handle positioning correctly. Luck is a part of it but I think positioning is the biggest flaw. Unless it has been refined... And I guess that's possible... But I'm not aware of refinement. It doesn't account for the shifting that takes place... It also can punish players for catching balls out of the zone because balls caught out of the zone as a result of positioning can also result in hits falling inside the zone based on positioning.

 

Also... The dude (for lack of a better word) tracking the metric has to determine the degree of line drivedness and fly ballness of each ball put in play. You will find those opinions vary from dude to dude.

 

Also... Consider that a MLB CF averages 2 to 2.5 PO's a game with the overwhelming majority of those catches routine and you can see its gonna take a long time to acquire enough data to produce a measurable variance that's worth talking about. Plus factor in 64 different zones that are used to compile the data and how many balls have to fall into each one of those zones to produce a variance and positioning becomes huge because one ball falls in and the whole model is screwed until more data is pumped in... The result is piling small sample size on top of small sample size because each individual zone further narrows the amount of data for an individual player because they use 64 zones. Therefore... Jamming routine data into a formula that is overwhelmed with routine result tends to produce swings in small sample sizes because if you stand in the wrong spot for one play... It nails you and the swing is going to be enormous.

 

Hopefully that helps explain why I don't belong to the church.

 

Now if you want to talk about OPS and OBP in regards to Ben Revere... It's a much more convincing argument and I'm on much more shaky ground. That's where I tend to plug my ears and say... I like him anyway. You just won't convince me that he is defensively flawed.

 

From a moderation standpoint... Don't ask a question of me and then answer it. That's disrespectful. Of course I don't watch every Phillies games and you know that and I won't pretend to... but I watch a boatload of MLB baseball and I've seen the Phillies quite a bit. My opinion on all MLB players is earned. I've put the time in.

 

I don't watch any minor league baseball and I think its clear that I go out of my way to not give my opinions on the talent of Sano or Buxton or whoever because I haven't seen them. I don't offer opinions on who the Twins should draft because I haven't seen them.

 

I've seen Revere play... I have seen Bonofacio play... I have seen Seth Smith play... I've seen Hank Conger play more than once and I have an opinion and have the right to express it and I've put the time in.

[Riverbrian]

I personally don't understand how UZR can fluxuate that much over an entire season. I would think given the number of balls hit to a players zone over the course of the season would even things out... unless the metric doesn't account for things like grounders through the infield, or liners off the wall...

 

It's not one zone... It's 64 Zones and each zone has percentages of balls caught based on league wide percentages of balls caught inside of each zone.

 

If you boil it down to just one of the 64 zones for example purposes. Let's call it "Zone Brian"

 

If across the league... 90 Percent of balls hit into that "Zone Brian" are caught. And Ben Revere or any CF doesn't make a catch in that zone where the ball is normally caught 90% of the time. The Weighting is going to kill him. And he may not make that catch because he was positioned strangely because of a defensive shift.

 

When you consider that a CF makes 2 to 2.5 catchers per game on average and then you reduce it to how many of those 2 to 2.5 catches actually land in that particular "Zone Brian" and those small sample sizes are spread across 64 zones. One ball not caught in a zone is going to take a lot more balls caught inside that zone to stabilize it and it can take years.

 

Because the Average CF catches 2 to 2.5 balls per game and the majority of them are going to be standard routine catches. That easily fit into that 90% league wide average of "Zone Brian". Or in Reverse... If Ben Revere makes a catch in "Zone Diehard" which the league wide average is 20%. Ben can be awarded because of positioning and again... it's going to take a lot of future data to stabilize it.

 

If Ben or any CF is out of position because of a shift. He can be penalized or awarded severely for something that is beyond is control.

 

And what happens when Chase Headley is playing a deep second on Jason Kubel. Is he now responsible for that zone around 3B. To my knowledge... Yes he is.

 

Now if someone tells me that they figured out how to halo players based on where they are standing and have figured out how to assess speed, angle distance and height of a batted baseball. Sign me up!

 

Maybe they have... If they have... Like Emily Latella... "Nevermind"

[Willihammer]

Another respected defensive metric - Ultimate Zone Rating (UZR) - also uses BSI's data but does not give credit for runs saved by shifts.

 

http://www.baseballprospectus.com/article.php?articleid=17183

 

DRS gives credit (or blame) for plays where a shift is on. However Revere doesn't score great in that either (-4 for his career).

 

So basically if Revere is getting burned (or helped) by shifts, it won't show up in his UZR one way or the other. But, what constitutes a shift for a CFer? I can't think of a single outfield shift. Mostly all they do is move in a little, move back a little, shade to the left, or shade to the right.

[Willihammer]

I remember reading a while ago how Denard Span was happy to be able to play shallower than he played in MN. Well, he posted pretty strong UZR numbers in MN but has been up and down as a Nat. Might be that Revere is doing the same thing, and costing himself big time UZR points by missing a few gappers that he would have caught if he'd been positioned deeper, as he was in MN.

[cmathewson]

Brian,

 

The shift thing makes little sense. Guys get more credit for getting to balls outside of their designated zones. Also, positioning is part of being a good fielder. Ripken was slow and old, but he was great at positioning, which is also Hardy's secret sauce. I want to know how many balls he's getting to that he should be getting to, whether he gets to them by positioning, good jumps, or because he busted his tail adjusting to a crappy jump from a poor position. It's always been my belief that Revere makes up for his poor positioning and crappy jumps with speed and hustle. He will slow with age. But his positioning and jumps are not likely to improve.

 

I do agree that we are talking about smallish samples. He plays behind a staff that gets a lot of Ks, so he hasn't had as many chances as I would have expected. Then again, you have to get to the ball for it to be a chance. The jury is still out on his defense. His offense, on the other hand, is not defensible.

[Riverbrian]

http://www.baseballprospectus.com/article.php?articleid=17183

 

DRS gives credit (or blame) for plays where a shift is on. However Revere doesn't score great in that either (-4 for his career).

 

So basically if Revere is getting burned (or helped) by shifts, it won't show up in his UZR one way or the other. But, what constitutes a shift for a CFer? I can't think of a single outfield shift. Mostly all they do is move in a little, move back a little, shade to the left, or shade to the right.

 

Thanks Hammer... That was a good read... I always amazed by the differences between the different advanced metrics and those differences make me pause. I encourage the continued work on Defensive Metrics. I don't think they are cooked yet but I long for the day that they are. I've always been a big defense guy and I think fully cooked metrics will help prove my personal hypothesis of the importance of defense.

 

I agree with the author when he wrote: "There are, I think, two larger points at work here. One is that seemingly trivial details can have large impacts on the results of “advanced” fielding metrics".

 

As for Ben Revere and any questions about his range.

 

I Offer this... Check out the video below:

 

It's one catch... It's a small sample size... Does he get this kind of jump all the time... I don't know... Many claim he doesn't... But if he's capable of doing this once... He's capable of doing it again and (just a guess) maybe 1 or 2 OF'ers could have done this. So if he has jump issues... They are fixable. But for obvious reasons... You can't make Willingham do this ever.

 

www.youtube.com/watch?v=tAts428eqQ0

[jokin]

Thanks Hammer... That was a good read... I always amazed by the differences between the different advanced metrics and those differences make me pause. I encourage the continued work on Defensive Metrics. I don't think they are cooked yet but I long for the day that they are. I've always been a big defense guy and I think fully cooked metrics will help prove my personal hypothesis of the importance of defense.

 

I agree with the author when he wrote: "There are, I think, two larger points at work here. One is that seemingly trivial details can have large impacts on the results of “advanced” fielding metrics".

 

As for Ben Revere and any questions about his range.

 

I Offer this... Check out the video below:

 

It's one catch... It's a small sample size... Does he get this kind of jump all the time... I don't know... Many claim he doesn't... But if he's capable of doing this once... He's capable of doing it again and (just a guess) maybe 1 or 2 OF'ers could have done this. So if he has jump issues... They are fixable. But for obvious reasons... You can't make Willingham do this ever.

 

www.youtube.com/watch?v=tAts428eqQ0

 

Huh. I'm sure his RF and DRS was docked by .2 for that lackadaisical effort.

[Riverbrian]

Brian,

 

The shift thing makes little sense. Guys get more credit for getting to balls outside of their designated zones. Also, positioning is part of being a good fielder. Ripken was slow and old, but he was great at positioning, which is also Hardy's secret sauce. I want to know how many balls he's getting to that he should be getting to, whether he gets to them by positioning, good jumps, or because he busted his tail adjusting to a crappy jump from a poor position. It's always been my belief that Revere makes up for his poor positioning and crappy jumps with speed and hustle. He will slow with age. But his positioning and jumps are not likely to improve.

 

I do agree that we are talking about smallish samples. He plays behind a staff that gets a lot of Ks, so he hasn't had as many chances as I would have expected. Then again, you have to get to the ball for it to be a chance. The jury is still out on his defense. His offense, on the other hand, is not defensible.

 

The highlighted part... We are in complete agreement. I also want to know this. I just don't believe that the current defensive metrics can tell me that for a wide variety of reasons.

 

The important thing to me is the damage that can be done by one ball in a zone that a large percentage get to but he doesn't get to because a coach waved him over to Dominic Brown's back pocket or he had a cramped foot or slowed up thinking it was Marlon Byrd's ball. And once you have that screwy piece of data in the mix. It can take a real long time to stabilize it because of the weighing involved and the small number of opportunities in that particular zone.

 

In the end... It's Ok... If you believe in what UZR or other metrics are telling you... I just don't belong to that church.

 

I love stats... I think Bill James is the man... I think he revolutionized baseball for the good and smart people are continuing to advance things but my personal belief is this: Some stats can tell you pretty solidly what they have done in the past but the eye can tell you what they can do in the future. Once a player fixes something... the data changes. It doesn't matter if a hitter is hitting .184 when he steps to that plate. If he gets a hit he's batting 1,000 into the future. It's why I don't want to throw Hicks away. He can fix things and he just might. I'd love to have Ben Revere on my team but I'm guessing that Hicks, Arcia and Buxton will all be better choices someday. That's just my personal opinion.

 

And I fully admit that I overvalue Speed... Probably to a fault but I can't see myself changing. Too much baseball under this bridge.

[jokin]

In one of the defensive metrics, I can't remember which, there's the final 10% area of batted balls measured that are rated- "impossible" for anyone to catch. Revere must be a leader in that category.

[Brock Beauchamp]

Brian, I look at that play and say "Boy, that Ben Revere sure is fast... But what the &^%$ was he doing in left field against a guy whose spray chart looks like":

 

 

It's possible that's coaching, but...

[Riverbrian]

Brian, I look at that play and say "Boy, that Ben Revere sure is fast... But what the &^%$ was he doing in left field against a guy whose spray chart looks like":

 

[ATTACH=CONFIG]7524[/ATTACH]

 

It's possible that's coaching, but...

 

LOL... To that I don't know. It's quite possible that right before Scooter made contact... Ryne Sandberg looked up from reaching into the sunflower seed bag and said. "What the Hell is Revere doing over there".

 

If Revere is positioning himself and having a consistent problem with it. The coaches need to put the sunflower seeds down and lean over the railing and wave their hat and position him for each batter.

[diehardtwinsfan]

of course if they were pitching him inside, perhaps they were expecting him to pull. I learned a ton about UZR today, so I really appreciate this conversation. I cannot help but wonder if it would just be easier to track the number of catches an OF makes relative to their peers at the same position. You can normalize it by the number of innings played in the position. I'd think the better defenders would typically lead those categories. If they are out of position, it shows. If they are in position, it shows. A lot more simple, and then you can start calculating how many outs/game a good defender can make.

[TheLeviathan]

of course if they were pitching him inside, perhaps they were expecting him to pull. I learned a ton about UZR today, so I really appreciate this conversation. I cannot help but wonder if it would just be easier to track the number of catches an OF makes relative to their peers at the same position. You can normalize it by the number of innings played in the position. I'd think the better defenders would typically lead those categories. If they are out of position, it shows. If they are in position, it shows. A lot more simple, and then you can start calculating how many outs/game a good defender can make.

 

It's a valid thought. I have a hard time swallowing most defensive metrics because they pretend to have a great deal of accuracy and objectivity but they are still, fundamentally, based on a significant amount of subjectivity.

 

They're nice to look at, they're interesting, but I haven't seen much reason to trust them.