Part of

« Morning thread: Phillies debunk Japanese signings | Main | Beerleaguer for breakfast: Wang to Washington? »

Tuesday, February 09, 2010


I agree. I mean sure, I'd like another back end starter due to our concerns about Moyer and Kendrick but if they pitch anything like they are capable of, we'll be fine with the 5th spot. Joe Biemel would be alright, but I'd rather we turned Bastardo loose.

Phillies Red: The key to me in that study is that it takes twice as many games to get a read on defense using UZR as it does to get a read on offense using wOBA. But in any event, I'm talking about UZR and OPS, not wOBA.

Maybe I'm not reading the stats right, but please glance at 10 players randomly and tell me UZR is as consistent as OPS. I just picked two Phillies.

Even then, just a cursory glance shows big jumps. Take Pedro Feliz: His UZR the past 4 seasons is 14.7, 20.8, 7.2 and 5.3. What am I missing here? Is that a pretty big variation? His OPS is .709, .708, .705, .694. Fairly stready, no? Are you saying Feliz is an exception?

How about Chase Utley? UZR: 9.3, 15.7, 20.2, 10.8. OPS: .906, .976, .915, .905.

Clearly I'm seeing this wrong somehow and I'd love for someone to set me straight.

Phillies Red, here's more: Victorino UZR (just 3 years because 4 years ago he was almost exclusively in RF).

UZR: 10.7, 6.5, -4.1. That's not a big variation?

Rollins UZR: 6.6, 12.8, 2.7.

Ibanez UZR: -5.8, -20.5, -12.1, 8.

Please tell me how that is not a big fluctuation. Also tell me how Ibanez went from a terrible UZR to a very good UZR if it didn't have anything to do with the stadium and who was playing alongside of him.

Obviously I am not seeing what you and Jack are seeing and I'd love to be educated on this.

I tend to agree with you, JW. In addition, even with the emphasis on getting back into the WS this year, I think it is important to find out during the early days of the season what Bastardo, Kendrick (and even Mathieson)can do in order to prepare for budgetary decisions in 2011. I suspect that, as you say, many of the options available now will be available later, with some notable additions. No objection, on the other hand, to sign one or two of them to a minor league contract- if they will bite.

Not to worry. Kyle and Cole will become beasts with the help of their pal Roy. The "Beasts of the East" triumvirate will dominate the NL. The fearsome threesome are probably working out together as we speak.

I may be a broken record but besides pitching depth the phillies should keep their eyes open for a RH INF. If bruntlett wasn't good enough to remain on the team exactly how is castro sufficient? hiding behind "he won't play that much" is rediculous. The RH INF position could easily supply 8-10 ABs a week through PH and resting either utley or rollins.

I know there is little chance of such a move being made but i still see it as a glaring weakness in the roster

Cliff Corcoran has a writeup on how many wins he thinkd Halladay and Lee will ad to their teams:

As clout would say: Interesting.

That Corcoran article is interesting indeed. I am surprised that the Phils getting Halladay is a bigger upgrade than the Mariners getting Lee, but the reasoning makes sense to me. I wonder how many wins he would have rated Polanco over Feliz?

The Corcoran piece, combined with JW's commentary, makes me wonder:

Does Halladay add 4 wins to a 93 win team?

Will the bullpen, which (with the exception of Lidge who was awful) was fairly mediocre, be a liability in 2010 or an asset that will hold its' own and propel the team to the magic 100 win mark?

That means the 'Bastardos' need to perform when called upon, and career average performances from the likes of Baez, Madson and Durbin should suffice.

With Halladay and an improved Hamels, a league average bullpen may just be enough to get them another 7 wins in 2010.

This could be the year for another Rollins prediction.

krukker, I was looking for the Polanco/Feliz comaprison also. I wonder if he'll do a second article and cover that.

I suspect that Polanco is a slight downgrade defensively and a decent upgrade offensively in 2010.

So, without any statistical support, let's speculate and say that Polanco's defense costs the Phillies one win more than Feliz's would have (-1), but that his offense at the top of the order (good contact hitter) and the movement of Victorino to the 7 spot in the lineup gives the Phils 3 more wins (+3).

That translates into about 2 more wins with Polanco in the lineup.

That, plus Halladay's 4 should bump the Phils into the 98-100 territory.

As I posted above, much of the season will depend on the effectiveness of the bullpen.

If the bullpen is as good as 2008 this team has a chance to win 100-1005 games. If it's as mediocre as 2009 - with a closer who is ineffective 2/3 of the time, because of the psychological effect that can have on a team (losing late) IMHO they'll struggle to add 4 in the win column.

Still, unless they are hit with a rash of injuries, they should still win the division.

If the offense continues to be one of the league's elite, then yes Halladay probably does add 4 wins. If everything falls into place, we could easily win 100 games.

I think we take our offense for granted sometimes, but a regression and/or major injury or two on the offensive side will very quickly wipe out the upgrade on the pitching side.

Of course any team has this issue, but the Phils always seem to have overachievers on offense. If, say Werth regresses to .240/.340/.450 with 20-25 HR, that will probably hurt more than having Halladay helps.

OOOPS. that should say "100-105" wins.

damn fingers!

awh, I agree with your assessment of Polanco and Feliz. The only thing I'm worried about is Polanco's bat. He doesn't walk a ton or have a lot of power, so his value is very closely tied to his BA. As such, his BA needs to be very high (.300 or better) for him to be an offensive positive. Hopefully he gives the Phils 3 good years, but I am only cautiously optimistic.

I should point out that even if Polanco hits .280, with career walk and ISO rates, he still constitutes an upgrade over Feliz, albeit a much smaller one than if he were to hit for .300.

krukker, do you really see Werth - in a contract year - regressing THAT much?

Krukker: I'm curious as to why you think Werth could regress to a line like that? There is the possibility for injury, but nothing about Werth's 2009 stats lead me to believe that last year was a fluke...

awh- 1050 or 1005-- As long as the Phils make the post season and their last game is a win, I'll be happy.

awh - I was kind of hoping for 1000+ win season.

No way, it was just an example that regression due to age or significant injury on the offensive side could be potentially devastating. So much so that it would wipe out our pitching upgrades.

If I had to guess Werth's 2010, I say .275/.380/.510 32 HR and 26 SB.

No wonder Beimel's still available...would you sign a guy who looks worse than Nick Nolte's drunken mug shot?

Clout, I'm certainly not trying to be a jerk here, but looking at Phillies we all know is certainly not random selection.

In any event, the analysis that I linked to in the last thread suggests that you are a bit off in declaring UZR to be wildly more variable than offensive metrics. If you look at that fangraphs article, one of the crazy occurrences is that UZR actually stabilized faster between 2008-2009 than offense. Tom Tango pointed out in the comments that this is certainly weird (perhaps a random occurrence for 2008-2009) and that his research shows it usually takes 100 games of defense to get URZ to correlate at r=.5, and 50 games of offense. So generally speaking, you are correct that OPS (more accurately wOBA) correlates at lower samples.

But that doesn't change the fact that the difference is much less than you seem to be making it out to be. That's what at least three different analyzes pointed to in that thread. And again, this is not a claim that UZR is highly stable or predictive, but that OPS and wOBA may be less stable than we generally give it credit for, Pedro Felix notwithstanding.

In other news, the idiots in Queenshave finally figured out they need to lower the wall at Bailout Park:

awh: I wonder if they'll decide to increase the height of the wall in the right field corner. I think we started calling it Chase's corner, right?

I'm not a stat guy, so I only know what I'm told. Doesn't UZR factor in BIS video scouting? Bill James' Fielding Bible Plus/Minus?

Rollins rated as the highest shortstop last season by that measure, earning a +23. This season: -3, which ranks 23rd among his peers.

Jack Wilson, Brendan Ryan, Yunel Escobar, Marco Scutaro, Elvis Andrus, Cesar Izturis, Paul Janish, Troy Tulowitzki, Steven Drew and Jason Bartlett were the +/- leaders at SS.

krukker, maybe they should concentrate on increasing the height of their third baseman? He might hit more HR at home then, right?

Murph just updated "High Cheese". I'm going over there to see what's up.

back soon.

awh LOL!

Murph has a rundown of the ST invitees, and breaks it down into who will and might be breaking camp with the team, and who's soen the road.

Phillies Red: The Phillies aren't representative? Why? Should I list the remaining 4 Phillies starters' UZR? I picked 4 guys at random. If you want, pick another team and I'll look at that team.

I'm not talking about predictive value. I'm talking about fluctuation. The UZR numbers look to me like they fluctuate quite a bit. If I'm looking at them wrong, please enlighten me. Maybe a swing from -16 to +8 isn't a big fluctuation, but just looks that way. I honestly don't know. But on the surface UZR seems to fluctuate quite a bit and that makes me wonder how valid it can be in terms of truly telling you a guy is a good or bad fielder. I do agree that themore years you have the better picture you can get.

In other news, Larry Bowa is quoted as saying that the dodgers have to keep a close eye on Manny Ramirez because he wore down last season:

Hey Larry, did it ever occur to you that the reason he wore down and was unproductive is that he got caught using a recovery agent for steroids, and that thereafter he no longer could artificialy enhance his performance?

BTW, is there any place one can go to compare the list of PED users to Scott Boras clients?

With both A-Roid and Manny being poster boys, I wonder how much 'overlap' there is between the two groups, and whether any other agents have had a higher percentage of their clients busted?

JW: According to UZR, Feliz, Rollins, Vic, Howard, Werth and Utley were all worse defensively than they were in 2008, some by a considerably margin. Ibanez was hugely better than the previous year.

Does all of that confirm what your eyes told you last season?

From what I understand about UZR, any single season can be completely out of whack with the general numbers, and therefore it's more useful taken in approximately 3-year averages to get an accurate picture. Which, to me, makes it a lot less indicative/predictive than people like to make it out to be.

Chris in VT: That's sort of where I am on it. There has got to bhe a lot of noise in the thing or it wouldn't fluctuate wildly like that. Did Rollins really go from one of the top fielding shortstops in baseball in 2008 to one of the worst in 2009?

If Manny wore down playing only about 100 games last year, are they going to have to make him a platoon player to "keep him fresh?"

He's just not the same player since he stopped trying to get pregnant...

I didn't say they weren't representative, but that it certainly isn't random, which is the word you used at the start of the thread. What are the odds that you would pick 5 random players and they would just happen to all start for the Phillies? The Philles' players may or may not conform to the norm in this case, but looking at such a small sample isn't likely to get us very far either way.

And you're right, we're not talking about predictive power - though this is really the underlying issue. But we are talking about correlation, and therefore variance, or fluctuation, year to year. I don't know about the UZR scale either, or how exactly it compares to the OPS or wOBA. But I have cited a reasonably reliable source that, using statistics, suggests that UZR and offensive measures correlate to themselves at similar rates. Thinking a bit more about this, just because they correlate at similar rates doesn't necessarily mean they don't fluctuate differently, that is, their variance might be different. But it does seem that, on average, these two stats correlate with similar sample sizes. That's what the data says, regardless of how it looks.

I have been taking a break for the past few days to concentrate on the super bowl and such, but it looks like I haven't missed much since it has been pretty slow around here.

Just a few notes:

- Regarding that VORP article form last thread. Do posters really need that stat to make you relaize that Rube upgraded EVERY position that we needed to fill?

I don't need a computer to tell me that Hallday (best pitcher in the game) is better than Lee (Top 10 pitcher in the game). Same goes for Polanco over Feliz and even Schneider vs. Stairs.

- Also, it takes people until February 8th to realize that this team can/will win 100 games? Really? I guess the past 3 months of complaining over losing Lee and wondering about what prospect pitcher will pitch in 2013 made people forget about the point that this team actually improved.

- The best was during the Super Bowl clout grumbling how he lost 500 bucks on the puppy bowl? Again, only clout would actually bet on "the puppy bowl".

And...for my line of day by our very own AWH:

"Does Halladay add 4 wins to a 93 win team?"

That has to be the biggest no-brainer/understatement in the history of BLer right there.

Chris in VT: "From what I understand about UZR, any single season can be completely out of whack with the general numbers, and therefore it's more useful taken in approximately 3-year averages to get an accurate picture. Which, to me, makes it a lot less indicative/predictive than people like to make it out to be."

What I'm suggesting is that this is also an applicable statement for offensive numbers taken in isolation.

Clout: No question. It's shocking, especially with Rollins, who I thought had a career year defensively. But since it's a +/- based on what is deemed to be league average, I wonder if the league got that much better. There has been a noticeable shift among teams toward defense.

Clout: Rollins OPS+ by season:

2007: 119
2008: 103
2009: 86

Did Rollins really go from one of the best offensive SS in the league to one of the worst?

JW: You thought Rollins had a career year last year defensively? I thought that, while he obviously didn't make many errors, his range looked a little bit decreased. But I don't know, it's tough to remember now. This is what makes judging defense by eye so hard.

There's also another theory, which I believe I read a quote from Theo Epstein about, in why defense fluctuates more. Which is that, in addition to smaller sample size (again, a guy gets a lot less fielding chances a season than PAs), defense is also much more reliant on pure athleticism, which declines much faster than hitting skills (hand-eye coordination, patience, etc). Things like injury and age can affect a guy in the field a lot more than at the plate.

Howard's OPS:

2006: 1.084
2007: .976
2008: .881
2009: .931

That's a pretty good amount of fluctuation, no? Looking over multiple years, we can certainly estimate about what Howard is (.270/.360/.570), but in any given year in that range, his OBP has bounced between .425 and .339, and his SLG between .659 and .543. That's a pretty big range of numbers.

Halladay will add 5 wins to his buds Cole and Kyle.

I will say this: Defensive decline could be one of the more underrated issues headed into the season. This is a group all into their 30s.

Phillies Red: I should've said random Phillies. I picked Phillies because we're all familiar with them. You seem to be saying that even though the numbers wildly fluctuate we should ignore that because the study says they really don't. (Maybe the problem is the study has nothing to do with fluctuation.)

Tommy: You think Halladay's performance is 4 wins better than Cliff Lee's?

I know you don't really do "statistics", but that's quite a bold statement. 4 wins was the difference last year between Jayson Werth and Ben Francisco. Or, in other words, between an All-Star and a bench player.

JW: That's a good point. Also have to wonder how Polanco will make the transition from 2B to 3B at age 34. Ibanez seemed to play defense very well in CBP after Safeco. Can he keep that up? Can Howard keep his defensive improvement up?

I'd count on top-notch defense still from Rollins, Utley, Werth, and Ruiz, as well as solid play from Victorino. If those guys decline though, could definitely be an issue.

Jack: "Can Howard keep his defensive improvement up?"

UZR says Howard was worse defensively last year.

Jack: Which is greater, Rollins OPS+ fluctuation from 86 to 119 or his UZR fluctuation from 0.8 to 12.8?

Clout: True, and +/- says he was just about the same as well. But he was worse by both measures in 2006 and 2007, so you can say 2008 and 2009 have been improvements. Can he get even better? Go from league-average to above-average?

Jack: "If those guys decline though, could definitely be an issue."

According to UZR, every one of those guys declined last year (except Ruiz who doesn't have a UZR).

Clout: My point is that fluctuation occurs a lot in offensive stats as well. But you can get a good picture of a guy's offensive profile by looking at a lot of different statistics over many years, and combining that with what you see subjectively.

You can do the same with defense.

Jack: I agree. My only point is that UZR fluctuates wildly and, the way I read numbers, it fluctuates year to year far more than OPS. That makes it a far less reliable stat for a single year performance. My opinion on why is that it's because there are far more external factors in UZR than in OPS.

Clout: And our Team ERA declined from 3.89 to 4.16, despite adding Cliff Lee for two and a half months.

I have no idea how much of that can be attributed to defense (as opposed to something like Jamie Moyer pitching his age), but it wouldn't surprise me at all if worse defense in 2009 led to more runs being allowed than in 2008.

"Which is greater, Rollins OPS+ fluctuation from 86 to 119 or his UZR fluctuation from 0.8 to 12.8?"

My guess: I might not understand the stats correctly, but I'm fairly sure that the OPS+ fluctuation is greater (or, at the minimum, more important).

In UZR, he went from being average to above average when compared to other shortstops.

In OPS+, he went from being well below average to above average, when compared to other hitters.

Clout, there may be differences in the exact variance, but let's look at the article:

"I am constantly hearing how UZR is an unreliable way to measure fielders because there are “wild fluctuations” from year to year. I decided to put that to the test doing an apples to apples comparison of UZR from 2008 to the nearly completed 2009 season. I only compared players that played the same position and had the qualified number of UZR Chances. Here are the results:"

So this article is addressing your exact concern. Let's skip the numbers and cut to the chase:

"So the lesson is, when there’s not a lot of UZR data on a player, there will be a lot of noise, but as the sample size increases, the data (at least from 2008 to 2009) actually becomes almost as highly correlated year to year as the stats that are considered to be the most reliable."

What is there to not understand about this? UZR is as "correlated year to year as the stats" like OPS. That means that it doesn't fluctuate, on average, any more than offensive stats.

So, tommy, because I posed it in the form of a question, are you saying that makes me as silly as you usually are?

Are you GUARANTEEING 97 wins?

Just askin'?

Will Ohman to O's on minor league pact.

Phillies Red: And yet when you line up the individual numbers they DO appear to fluctuate more. Again, assuming the numbers have the same relationship to each other (i.e,. the difference between 1 and 2 is the same amount).

I encourage you to spend an hour looking at UZRs and then OPS for various players and then tell me there's no differnece in the size of fluctuation.

What is the unit on that UZR? What kind of percentage increase is it from .8 to 12.8?

Andy: That is exactly the issue. Posters here are trying to say that the pct. difference between 0.8 and 12.8 is less than the difference between 86 and 119. If the numbers are equal then that's absurd. On the other hand, if UZR uses some numerical scale were 1 does not really equal 1, they may be right. I don't know the answer.

Clout, dude, what are you talking about? The scales are different, yes, but that is not an argument against what a handful of researchers have shown when they actually crunch the numbers, which is that the fluctuations aren't any greater given the appropriate sample size, again, on average.

You are clearly a smart guy and know lots about baseball, so I'm not sure why you are just rejecting solid analysis because it "appear[s]" that one stat fluctuates more than another. Are you really ready to believe what a small sample of say 10-20, or even 30 players, tells you when you can look at a much larger sample? To make the point even sharper, are you really arguing that appearance is a better judge than a statistical analysis? Really?

"What is the unit on that UZR? What kind of percentage increase is it from .8 to 12.8?"

UZR (ultimate zone rating): The number of runs above or below average a fielder is in both range runs, outfield arm runs, double play runs and error runs combined.

My reading of that suggests than an UZR of 0 is average (for that position).

Jack - I think Tommy's point is that Halladay wil be 4 wins better than the slot in the rotation that Lee occupied for a 3rd of the season last year. Remember, Lee only made 12 starts for the Phils in the regular season, and went 7-4 3.39 in those 12 starts. But Brett Myers, Rodrigo Lopez, and Antonio Bastardo all filled that slot in the rotation at one point or another in the season, and they collectively combined for an ERA over 6.

It's not much of a leap to say that Halladay would represent a 4-win improvement over that group collectively (Lee/Myers/Lopez/Bastardo) last year.

So a fluctuation of 12 runs between two years...hmmm...

Well maybe that's not too much. But if just about anybody can fluctuate that much, doesn't it make it kinda meaningless? Just wondering, not taking sides.

And having the numbers state that Howard was worse in 2009 than in 2008 really makes me kinda question them. That's not what my eyes told me. And it's not what this board saw during the 2008 and 2009 seasons either.

Phillies Red: You keep ignoring my point: Why do the NUMBERS in UZR appear to fluctuate more? Please spend some time looking through UZR numbers, as I have. There may be a perfectly reasonable explanation of why they APPEAR to fluctuate. If you can't do that, then let's just move on.

Chris in VT: Fair enough.

What the heck is UZR? Can you posr a list of the non-routine abbreviations?

"So a fluctuation of 12 runs between two years...hmmm..."

Actually, the 0.8 value was from 2005 and the 12.8 value was from 2008.

Also, UZR is a counting stat. FanGraphs uses UZR / 150 as a rate stat. In 2005, Rollins had an UZR/150 of 0.7. In 2008, his UZR/150 was 15.0.

I guess the Howard thing indicates a different problem with UZR. Since it is runs abov e average for that position and (this is the problem point) for that YEAR, then you can't compare one year to another. Because what if all the great shortstops except for Jimmie get injuries - suddenly he's a better fielder? Because he will look better compared with scrubs; his number might be 25, but he's not playing any better than he was when he was at 12.

UZR = Ultimate Zone Rating...It's supposed to be a measure of a player's fielding ability/range.

There is obviously quite a bit of debate about it's actual usefulness.

And the easiest way to find these things is to (google) them.

Clout: Yes, there are fluctuations. But if you sort through the numbers, you can find a pretty reasonable evaluation of someone. Take Shane Victorino, for example.

If you look at Victorino's page on Fangraphs, you'll see a ton of UZR numbers, all of which seem to deviate wildly. But that's because he's played a bunch of different OF positions over the years, and many of those samples are tiny. If you just look at the CF numbers from the past two years, when he was a full-time CF, you'll see UZR of 6.5, and -4.1.

So he was above-average in 2008, and slightly below-average in 2009, and not by a huge deviation (that's within 10 runs we're talking about). Those seasons contained just over 300 putouts each, so combine them, and you have, over two years, about the same number of data inputs as one season's worth of PAs, and a picture of a CF who is probably about league-average defensively. Add in this season's upcoming totals, and we should have a pretty good picture of what Shane Victorino is defensively.

Now, saying Victorino is a league-average defender in CF, maybe slightly above-average, probably matches up with your subjective opinion, right? Additionally, we see that he was a VERY GOOD defensive RF when he played a full season there, posting a UZR of 13.4, which makes sense, as we would expect RF to be an easier defensive position than CF (thus, a league-average CF would be a very good RF).

Jack: There were only 6 CFers with a worse UZR than Victorino last year. I wouldn't call that "slightly below average."

"a league-average CF would be a very good RF"

Garry Maddox would have been a lousy RF.

I'm using the numbers. -4.1 is slightly below-average. 0 UZR does NOT equal the median (or the mean) CF in any given year, because I believe "average" is calculated over a larger sample than just the CFs from that given year. UZR in a given year is not scaled to itself. It has calculated an "average" defensive player based on years' worth of data on batted balls turned into outs in any given position.

I could be wrong, but I believe that is the case. Phillies Red, do you know anything about how the "average" is calculated?

Jack: So, theoretically, the CFer who finished 30th in a given year in UZR might actually be pretty good? If that's true I really don't get this stat at all.

Jack: My point is EXACTLY what Chris from VT described.

AWH: "Are you GUARANTEEING 97 wins?"

Barring long term (more than 20 games) injuries to Howard and Utley grouped with another horrific year from Lidge, yes. I am pretty confident with all the upgrades along with the best pitcher in baseball we can manage to win 97 games.

** Keep in mind the irreplaceable Clifton Lee only went 7-4 with a 3.39 ERA here (in the regular season). Also, we only managed 1 more win in 2009 than in 2008.

The last 3 seasons average in the AL (since we use multiple season stats to prove points here) Halladay went 18-9 with a 3.09 ERA. Who would you rather have???**

Buster Olney ranks the Phillies as the sixth best rotation in the majors behind the Red Sox, Yankees, White Sox, Angels, and Cardinals. Well actually he has them as 5a (and of course mentions that they'd be #1 if Lee was still on the roster).

He also has the Braves (would be #3 if Vazquez was still on board), Giants, Rays (depends on Garza and Price), and Diamondbacks (depending on Webb) in the conversation.

Clout: Yeah, I mean. If the Phillies and all the other teams promoted guys like Freddy Galvis to play SS, then you could have 30 SS with above-average numbers. Their offense would suck, but sure. What's so hard to understand about that?

The other thing is you can use your common sense. If I told you a guy had a 115 OPS+, you'd say he's above-average. But if you learned he was a 1B, you'd think he was below-average. Right?

I know you can handle that. It's not hard to use UZR as a helpful measure in evaluating defense. Even you understand that. So we're just arguing for the sake of it, right?

Jack: Here's another problem: Ibanez UZR of 8.0 was 4th best in baseball last year.

Is Ibanez a great fielder? Does he have great speed and range? If so, why was his UZR in the 3 previous years -5.8, -20.5 and -12.1??? Is it possible his ballpark change might be a factor? How many other outside factors impact UZR? As in, who is playing next to Ibanez?

What I'd love to see is the defensive equivalent of FIP. Something that removes all outside factors that have nothing to do with the player's skill.

Clout: No question the move from Safeco to CBP impacted Ibanez's defensive stats. And you should obviously take any LF numbers from Fenway, for example, with a huge grain of salt.

Ibanez also added 80 points of SLG, which was obviously park-related. But you're not questioning SLG as a stat, are you?

Once again, you just have to use your common sense and apply some context to help you understand the numbers. You can do it, I promise.

Clout: Hey, I agree I'd love to see some more refined defensive stats. And I think we will, in the future. For now, I'll use what we have (UZR, +/-) along with my eyes and some common sense to try and evaluate players. Not sure what is wrong with that.

Jack: The difference between UZR and FIP or OPS+ (which corrects for park factors) is like the difference between one of those plastic magnifying glasses you give to kids and an electron microscope. Both give you useful information, but one is a sharper tool.

Jack: Ibanez' slugging was better away than home last year. Find a better example.

"0 UZR does NOT equal the median (or the mean) CF in any given year, because I believe "average" is calculated over a larger sample than just the CFs from that given year. UZR in a given year is not scaled to itself. It has calculated an "average" defensive player based on years' worth of data on batted balls turned into outs in any given position."

I don't see that anywhere. Additionally, if you take all of the numbers from the FanGraphs leaders board (being sure to select 'All Players', as opposed to 'Qualified Only') for a given position, the pieces used to calculate UZR add up to a pretty small number.

I'm assuming that there's rounding which prevents the numbers from actually adding up to 0.

i can't find the link, but I posted it about 3 weeks ago that MLB is installing 4 object recognition cameras in all ballparks that will track each players movements - which will enable data on players range and speed in the field to be analyzed. Should make UZR a forgotten tool.

The old reverse Beerleaguer curse: Shoeneweis to Brew Crew. Here we go Bastar-do, here we go! *clap-clap*

No pitchers/catchers?
This is not a major problem
Let's talk about stats!

"Yeah, I mean. If the Phillies and all the other teams promoted guys like Freddy Galvis to play SS, then you could have 30 SS with above-average numbers. Their offense would suck, but sure. What's so hard to understand about that?"

Wait a according to UZR, the worst defensive SS in the majors could still be an above-average defensive SS? Wouldn't their definition of "average" have to change to better account for the standards of the position at that point? Is this like getting graded on a "weighted curve" back in High School, where some people could get like a 150 and others a 50 on a test so that the teacher looked like less of an idiot?

Did that make any sense at all or am I talking out of my @$$ at this point?

This UZR discussion has thrown me for a loop.

No worries Chris:

"There are three kinds of lies: lies, damned lies, and statistics." - Charles Wentworth Dilke (1843-1911).

"They throw the ball, I hit it. They hit the ball, I catch it."

Willie Mays (pre UZR)

Jack: "Ibanez also added 80 points of SLG, which was obviously park-related. But you're not questioning SLG as a stat, are you?"

Ibanez SLG at Safeco 2008: .497
Ibanez SLG at CBP 2009: .511


Halladay, Hamels, Happ, Kyle and Joe have a combined .618 winning percentage. That should set the tempo to win 100+ games.

Ibanez SLG away 2008: .462
Ibanez SLG away 2009: .639

That's a lot of fluctuation. But I doubt it's park related. Unless it's far easier to hit in NL parks than AL parks.

No, it's probably just far easier to hit NL pitchers.

Average UZR and UZR/150 is 0 by definition:

I've fallen behind in this conversation, so I'd don't want to derail it yet again. But clout, first, I have to laugh, because I've already spent way too much of my life looking at UZR data. And I agree, at first it looks weird and inconsistent. But I'm not sure we need to actually explain that. I believe that the 4-5 analyses examining this issue are way better at giving us perspective on this than looking through a bunch of semi-randomly selected players.

But I guess the problem is scale. UZR likely (at least functionally) falls on a -30 to +30 scale whereas OPS+ seems to be spread over a larger scale, say 0 to 200+. But again, I wonder why we need to explain an appearance if we have reasonable data suggesting what is actually the case.

UZR is a tool...not the end all be all of defense.

Its based on a subjective scoring system and is thus just as subjective as that basis.

So it's kind of like a personality test. There's data, which can be quantified and analyzed. But at it's base you have to use a subjective tool to provide the data.

The comments to this entry are closed.

EST. 2005

Top Stories


Rotoworld News

Follow on Twitter

Follow on Facebook

Contact Weitzel