UZR and Jason Bay
Today the Boston Herald and Deadspin homed in on Jason Bay’s 13-run change in his 2009 UZR. John Tomase of the Boston Herald writes that “UZR owes Jason Bay an apology” and Deadspin’s Barry Petchesky writes that “it’s foolish to jump aboard the Sabermetrics bandwagon.”
John Tomase goes on to say:
… Ultimate Zone Rating, which was treated as Gospel this winter during all the discussions of defense around these parts.
Let’s just say that it’s not UZR’s fault that it’s been treated like gospel, and anyone who actually reads FanGraphs should know that we definitely don’t treat single-season UZR like gospel either. Dave Cameron addressed “not liking or agreeing” with UZR last year in his Bay vs Cameron article and Jack Moore consulted multiple defensive metrics in his contract analysis of the Jason Bay signing.
Even with the UZR improvements, Bay is still listed as -35 runs below average since 2007. Some will point to his poor 2007 as the result of still recovering from knee surgery, but his problems in the outfield continued to plague him in 2008 between playing in PNC and Fenway.
All the other defensive metrics seem to agree that he has been pretty sub-par since 2007 as well. John Dewan’s +/- has him at -18 runs. Sean Smith’s Total Zone has him at -47 runs and Brian Cartwright’s WOWY based system likes him the best at -8.2 (having him at +9 in 2009).
Now there’s no doubt that, in 2009, it seems most systems have him somewhere around average, which is a fairly large departure from his 2007 and 2008 metrics, but let’s not forget that the fans themselves seemed to think Bay was below average in 2009. Tom Tango runs the Fans Scouting Report each year and they rated Bay a 2.69 on a scale of 1-5, which placed him 48th of the 71 rated left fielders.
Do we need to average out every single defensive system for each player to get a decent picture of his defensive abilities? Probably not. But when in doubt (like some of you were with Jason Bay), you can always get a second opinion, and a third opinion. Right here on FanGraphs we now carry two separate defensive systems (UZR and John Dewan’s Defensive Runs Saved) for all players.
Baseball statistics can be a valuable tool in your toolbox when analyzing a player. Often times they can be one of the most useful tools in your toolbox, but there’s really no reason to throw that tool away completely, just as you wouldn’t throw away scouting information. And as I’ve said before, a lot of times the metrics you use actually do contain “scouting” information.
New information and better models are coming out all the time and this is not unique to baseball. Analysts use the information that’s available to them and in light of new information, there’s a chance that the previous analysis might be wrong.
But the alternative as Barry Petchesky wrote is that “it’s foolish to jump aboard the Sabermetrics bandwagon.” Suppose we go back to the days of batting average and fielding percentage; to how many players would batting average or fielding percentage “owe an apology”?
David Appelman is the creator of FanGraphs.
nit pick: it’s “homed in” not “honed in”
http://dictionary.oed.com/cgi/entry/00333343
Most people don’t have access to the OED–and certainly not at home.
Honed in is perfectly fine.
Not to me!
Nor to Appelman, it now appears.
Just cuz George Plimpton’s editor was too chicken to point out an error doesn’t mean we all have to accept it.
M-W.com says…
Main Entry: hone in
Function: intransitive verb
Etymology: alteration of home in
Date: 1965
: to move toward or focus attention on an objective
usage: The few commentators who have noticed hone in consider it to be a mistake for home in. It may have arisen from home in by the weakening of the \m\ sound to \n\ or may perhaps simply be due to the influence of hone. Though it seems to have established itself in American English (and mention in a British usage book suggests it is used in British English too), your use of it especially in writing is likely to be called a mistake. Home in or in figurative use zero in does nicely.
This super-minor semantic argument is of vast importance to the main thrust of the article. Let’s comment on it some more.
“Honed in” is incorrect. It’s “homed in” (think of a homing device). It’s a common error, but that doesn’t make “honed in” correct. Why is everyone voting down bowie when he admits it he is nitpicking yet is also absolutely correct?
UZR has always been treated in 3 season sample terms, why make a deal about 1 year?
Except for when discussing value. Bay just gained 2 WAR last year. He was worth two more wins.
Why single season UZR was ever used for value purposes is beyond me. Ben Zobrist was the most valuable player in baseball last year.
It looks like Zobrist got knocked down to 8.3 for 2009 against Pujols’ 8.6, so Pujols was infact tops per Fangraphs WAR. I guess this had something to do with the UZR update too, because I also recall Zobrist topping the leaderboard.
If we’re trying to figure out who had the most valuable year, I think it’s just as fair to include people with potentially transient defense as it is to include pitchers that gave up low single digits of homers despite a normal FB% and an unsustainably low HR/FB% (even in stark contrast with their own career average). Or batters with insanely high BABIPs. The outcomes actually occurred, despite what many of us here think “should” have occurred.
Zobrist was a top ten hitter and played plus defense at a lot of positions, mostly 2B. If it were Utley putting up around 40 batting runs and around 20 defensive runs saved – as he has before – I don’t think many of us would bat an eye at it. Zobrist had a hot year. It happens!
Or maybe, that difference is irrelevant because WAR isn’t precise beyond +/- 0.5 either way of its listed value, and so Pujols and Zobrist should have been considered roughly equal before and after, rather than splitting hairs over who is a couple decimal points higher.
Astro – the difference between UZR and ERA with a low HR/FB rate is that the former has measurement error while the latter does not. You can argue whether or not a pitchers HR/FB rate should be considered or just his FB rate, but you can’t argue that those home runs didn’t actually happen. Just like you can’t argue that Chris Carpenter didn’t give up 2.26 runs per 9 last year (or whatever). You CAN argue that Ben Zobrist was not worth +27 runs with the glove last year because we know that UZR has measurement error in describing past events. Therefore, single season UZR should not be mixed in with other stable stats, at least without a standard error around it.
LOL, Zobrist was among the leaders in uzr at 2 different positions last year. He was a hitting stud and he was better than Utley in the field and he was a studly glove in the outfield all in one player. The adjustment makes him seem more like a human being. A very talented one still.
You CAN argue that Carpenter didn’t give up his runs, because of the defensive component. ERA is still measuring exactly what it sets to measure, though. On Zobrist, +/- had him at +30 (!) runs between 2B and RF alone. Looks like he had great RZR/OOZ numbers as well. You can of course argue that he wasn’t that good, but what evidence would you use? A Cardinal-fan belief that he wasn’t nearly as good as Pujols? Even if it were rounded with a “standard error” – which I presume would be guesswork – Zobrist would still rate to have high defense for 2009 via UZR. Combined with more transparently excellent bat work, that’s going to put him very near the top of the pack overall.
Yes, that’s why I am arguing that Zobrist was not as good as UZR and +/- have him. Because I am a Cardinals fan homer, not because it’s unrealistic to expect a player, who had put up poor UZR scores in the previous years and had to be moved off of shortstop because he wasn’t good enough, to be worth +30 runs at two different positions.
It’s very easy to argue that someone has not played like a 30 run defender because we know there is a lot of measurement error and we know that a player is more likely to be average than +50.
Total Zone rating has him at +12 last year.
All of his pre-2009 years do not amount to a single year’s worth of defensive play, and I don’t think it would make a lot of sense to use UZR in order to discount UZR. I don’t know the reasons why he was moved from SS to RF/2B, but I would expect a SS to field better at 2B. I also don’t agree that “we” know there is “a lot” of measurement error, in terms that non-specific.
As for TotalZone, I’m extremely glad it exists for historical estimations but I’m still going to heavily lean toward PBP-based defensive stats than those using Retrosheets for current analysis. Even Sean Smith, TZ’s creator, has stated that UZR is superior. On the specific subject of 2Bs and TZ, I’m reminded of Hornsby with his skillet-glove reputation and Morgan with his five gold gloves. Personality and popularity could no doubt be a factor, but enough to call Hornsby +53 and Morgan -41 for their respective careers? On that case, I guess we’ll always be left to wonder.
I could certainly see Zobrist’s 2009 as being a defensive aberration due to sample size and/or positioning – a mostly unaccounted variable which has an understandably large impact on range statistics – but I don’t see a reason to conclude that UZR is problematic to the point that it should be excluded from value considerations, either from this concern or the more clear-cut corner case of Jason Bay & Fenway.
So you are saying that Zobrist was worth +30 runs in the field last year…. gotcha
Very good response to the Deadspin article David. Those who follow sabermetrics closely are well aware that defensive stats are still somewhat of a work in progress and need to be interpreted with caution. I don’t think FanGraphs or most of the other big saber sites have suggested otherwise. There will always be those who take new tools and use them improperly but it’s hard to stop them.
Additionally, I see the willingness to tweak new stats to address flaws as a positive. Some will see that as opportunity to criticize sabermetrics but ultimately it’s more important to get it right.
Lee
“There will always be those who take new tools and use them improperly but it’s hard to stop them”
That was funny. Thanks.
Everyone on this site, most especially its authors, use yearly UZR data. It’s built into WAR, ie, value, and every discussion regarding particular players and their impact. And the thing is, as Bay minus Fenway testifies, as others before had as well, it’s dead wrong. UZR has never figured Fenway right, and I suspect it never will. Same goes to SLG at the new toilet and other venues that deviate from the cookie cutter standard stadium.
Ignore it.
It’s especially infuriating for Petchesky to write that, as Deadspin’s supposed to be “new media” and should be over this dead debate at this point. I actually like most of Deadspin’s other writers, but Petchesky is insanely glib and uninformed on a pretty consistent basis, which undermines Deadspin as a whole a little bit, or at least their consistent (and mostly valid) critiques of other sports reporting.
I never ever want to hear about Tango’s “Fan Scouting Report” ever again, ever. The only signal in that noise is selection bias, i.e. 95% of the fans that participate are saber-oriented types who look at UZR DRS +/- data on an almost daily basis. I’m as big a fan of advanced metrics as you’ll find, but I cannot stand that particular argument, when it comes to their accuracy.
That is not true. A very small percentage of the voters come from fangraphs and other sabermetric sites.
I can’t speak for all the teams but I have a pretty good idea who most of the Tigers voters are . Most of them are not sabers.
Lee
The very fact that they are reading Tom Tango’s Blog guarantees that a high percentage of them are vastly more aware of advanced fielding metrics than the general public. But perhaps this selection bias is acceptable, since I think I would trust a baseball internet nerd’s perception of defense more than a randomly sampled casual fan.
But misc, most of them aren’t reading Tango’s blog. Most of them are solicited from, frankly, FanPosts on SBNation websites, posts on forums, etc. There may be a self-selection bias in who decides to respond, but I don’t think it’s as larger or obvious as you suggest.
Well sure, most people don’t find out about the project directly by reading tango’s blog, they find out about it from various other blogs whose writers or fan-posters are typically saber-oriented as evidenced by the fact that they are linking everybody over to tango’s blog.
Or heck, maybe the first year they participate they are pristine un-sampling-biased data points. The fact that that person put time and effort in over at tango’s blog means they will probably look at the results and hang around and maybe read some articles about how the fan’s scouting report matches UZR, and then they start learning about UZR, and then the next year they are converted into a biased data point.
I myself specifically refrain from participating because I do not have time to actually sit down and watch very many games, so my perceptions of players is mostly informed by reading the interblogs and looking at advanced stats and box scores, etc. I know for a fact that my input would be almost pure bias. So perhaps my viewpoint is skewed because I assume that most of the participants are like me, when in fact they are mostly impartial observers with thousands of hours of free time devoted to amateur defensive scouting.
FWIW, from http://www.tangotiger.net/scouting/
“And, most importantly, do not, absolutely do not, look at any numbers. Don’t look at his fielding percentage, range factor, zone rating, UZR, or anything else that someone else is telling you. I just want you to rely on your eyes. You are the scout. I need you to rely completely on your own observations.”
Telling people not to look at the numbers after they spent the previous 9 months looking at the numbers is not going to magically make them unbiased. Sure, maybe it helps a tiny bit.
I guess my point is that the survey is fine for what it is, but you can’t use opinion to substantiate a metric, especially when it is the opinion of people who are likely to be aware of the metric and have some level of confidence in said metric. It is a self-fulfilling prophecy.
The Win Values metric considers only UZR (not other fielding systems) and only a player’s one year totals. It’s not as nuanced as the discussion in this blog post.
http://www.fangraphs.com/blogs/index.php/win-values-explained-part-two
Its funny that we can all know that a player had an unlucky year at the plate due to BABIP despite solid peripherals all around. But when a player has a bad year defensively when they are perceived as better than that season, its either a sample size issue or defensive stats sucking. What if a guy is just lucky or unlucky in terms of how the ball happens to be hit toward him in that single season? We see it happen for guys at the plate, it surely happens to guys in the field too. For those reasons its why we always need to look further into the stats, at larger samples and yes, even scouting reports.
A players win value is what did happen on the field. It isn’t what is likely to happen moving forward, or even what should have happened. One year of UZR data tells us what happened and what the player did. It does not however tell us how that player will perform in the future though or how good their ‘true skill level’ is, just what actually happened on the field.
Can someone tell me how the sabre community can delve so deeply into the numbers yet apportion huge swaths of events to “luck”? Seems contradictory; perhaps that is why the unfavorable perception that many of the metrics are useless prevails CONSISTENTLY.
If your numbers aren’t building solid truths, what “service” are you doing for the fan base, in general? To hear you guys talk, you’re curing AIDS.
I wouldn’t push this issue if I didn’t always see this question asked and see the asker get shouted down for “ignorance” around here. Doesn’t make you cats sympathetic, TO SAY THE LEAST.
Huge “swaths” of events? I’m not sure you’re using that word quite right. I’ve certainly never seen it used a context like that before. What do you think the word means in the context you’re using it?
Anyway, when we “apportion huge swaths of events to ‘luck'” (LOL, way to try and make yourself sound smart), what we’re actually doing is trying to explain the variance present in smaller sample sizes. Variance is something that’s present in any form of statistical analysis. Sure, once a sample size is large enough, a number will regress to the mean, but in smaller sample sizes (i.e. 1 year of UZR) variance can have a large effect on the numbers we’re seeing. It doesn’t mean the model is wrong, just that the model needs more data until its starts becoming more accurate. In the case of baseball, we attribute this variance to luck, because that’s the easiest thing to call it. Say I get a bunch of bloop singles or seeing-eye singles. Would you rather say my high batting average is due to variance or luck? Does it really matter in the end?
“One year of UZR data tells us what happened and what the player did.”
Except when it doesn’t. Like in this case.
No one should take UZR all too seriously. That’s what leads to a somewhat justified backlash like what we see here (although he goes too far in lumping in UZR with statistics that are actually good). Even the creators of UZR caution us to take their statistic with a gigantic grain of salt. Unfortunately, Fangraphs DOES take it as gospel by using it in their much-publicized WAR statistic.
When did we declare WAR as gospel?
I don’t think FanGraphs takes it as gospel by putting it in WAR. Who ever said WAR should be taken as gospel?
Even if it really is the strongest bit of evidence on a players’ performance — which is debatable — that’s still not gospel. There is no gospel, and no one at this website has ever told me there is.
Whenever you use it, and the accompanying ‘player value’ in your analysis, you’re stating it as statistical fact. It doesn’t matter what your intentions are, that’s how it’s presented.
Nowhere clearly visible on this site is there a disclaimer saying that statistics like WAR are extremely unreliable and therefore no conclusions should be drawn from it. You state it in the same way you state wOBA, which is a much more reliable and static statistic.
Look, I get why you don’t want to put big disclaimers all across your site. Reduces your appeal. But at the very least own up to the fact that you are contributing to the confusion by treating it like any other statistic, when in fact it should only be used as a kind of fun, not-to-be-used-in-any-real-analysis novelty.
David can speak for himself, but I think the error is with the reader who decides that’s statistical fact. I never read it as such. I don’t expect them to put a “statistics can and do miss out on parts of the picture” disclaimer in every single article.
“when in fact it should only be used as a kind of fun, not-to-be-used-in-any-real-analysis novelty”
Well, I can’t really agree with that either.
Which is why they should only be the start of a debate, not the end.
You rank players, you publish rankings, you sell rankings @ $9.99 a pop, people take those rankings and use them to bet money in fantasy leagues, or simply browbeat their “intelligence” into the “ignorant”.
I do?
You know there are other player ranking systems (many free, some not) that weight defensive contributions differently, or not at all, correct? Some for fantasy purposes, some not. No one is forcing anyone to pay for these rankings, or casting categorical dispersions on other ranking or value systems. To each their own. Isn’t America grand?
And I think it was Petchesky who was doing the browbeating here – categorically dismissing an entire school of thought because there was one instance of one measure that he didn’t much care for.
Dave rightly pointed out that UZR is simply one analytical tool that can be used to measure defensive contributions, and that it works better over larger samples and can fluctuate year-to-year, and should be used in conjunction with, not to the exclusion of, other defensive measures. To the extent anyone is using UZR (or WAR) as definitive proof of anything (in and of itself), they’re misusing it. They can help build a case for or against something, but cannot decisively prove anything in isolation.
Plus, I think a lot of times that misuse is on the reader’s end, not the writer’s. The writer (at least the ones here) are using UZR/WAR knowing that there is a fair amount of noise in a single year of data, but putting that caveat in with each and every use of the measure is just exhausting. You can only say it so much, and then expect your readership to know that the caveat is a given.
Well UZR had a good run.
Time to go back to my old standby, fielding percentage.
The only problem, David, is that you write that UZR is not very reliable for only a year’s worth of data, but you do not present the degree of uncertainty in your statistics, you present only the means (or average values). I think we can all benefit if standard errors or confidence intervals accompanied your presentation of statistics. If you presented that information, people will be able to have a better understanding of the context.
You provide a lot of free information for which the whole community should be grateful to you, but you can substantially improve the quality of existing information by adding the extent of uncertainty in the metrics you provide.
I’m not going to disagree with you, but I’ll defer to MGL if that’s something he thinks is going to be helpful/feasible. I think we’ve talked about getting rid of the decimal place in the past at the very least.
Even the batting metrics are not that reliable for one year worth of data, but everyone seems pretty comfortable with them. UZR has been publicly available only a little over a year and in general I’ve been surprised at the level of attention WAR and UZR has received in the mainstream.
I think there’s always going to be some growing pains when something goes from a niche stat that the few thousand people who visited FanGraphs probably had a decent understanding of, to a stat that is slowly becoming ubiquitous with defense where you don’t get all the UZR analysis straight from FanGraphs anymore.
I think on the site we try and use the statistics as responsibly as possible and we can only urge others (which we have oh so many times) that you can’t draw conclusions from just one year of UZR data.
But can you also urge others to not draw conclusions from single-season WAR data, which includes in no small way the vast amount of uncertainty from UZR data?
I don’t think he should urge them not to draw conclusions. They just need to realize their conclusions are not definitive and should be continually questioned for as long as they have time and interest.
Pitching statistics like ERA are also unreliable, maybe more so than UZR.
I think confidence intervals are nice for the saber community but I think they just add a layer of confusion for a more mainstream audience. An integer value is probably best if expanding the audience is a goal.
Even the batting stats in 1 season of WAR data are subject to all sorts of fluctuations. Half of the stuff we write about on the site is performance validation, so I think we’re used to having checks and balances on batting/pitching stats (BABIP, LOB%, HR/FB), but this is not so much the case with fielding stats.
I think we’ve done a whole lot of cautioning on evaluating players on 1 year of data for anything, more so for UZR, but still….
I think the reason UZR and WAR have received attention in the mainstream is because most people think it allows you to put a dollar value on performance. That is one thing most people understand, and can readily check against what players are making. But they don’t pay much attention to how it is derived.
It adds up things on the same scale (i.e., runs scored and runs saved and wins), but runs scored and runs saved are very different random variables because of their underlying variabilities. Which implies that the total runs contributed above replacement has pretty big standard errors surrounding them, and therefore, so do the WARs computed from such component stats. Therefore, the dollar values assigned to such WARs should also have pretty large standard errors assigned to them.
When you say the following: ” Suppose we go back to the days of batting average and fielding percentage; to how many players would batting average or fielding percentage “owe an apology”?” one comeback is reliance on BA and fielding percentage did not lead to heroic claims about how much a player is worth in dollar terms. But the presentation of WAR and value does precisely that, often without any care for standard errors accompanying them. Most readers are not statistically sophisticated enough to figure it out on their own without this being explicitly declared. That is why I think for the sake of integrity of these stats, the standard errors need to be presented.
I think confidence intervals are nice for the saber community but I think they just add a layer of confusion for a more mainstream audience. An integer value is probably best if expanding the audience is a goal.
I don’t think it is difficult to present at all. It is extremely important to hammer the point that without a standard error, a statistic is worth nothing. This site keeps repeating the sample size mantra all the time, writers crying themselves hoarse at times. You know what is truly affected by sample size? Statistical inference or the degree of uncertainty surrounding the statistic. In other words, standard error. The estimator is not really biased in a small sample, it is unreliable. If this can be spelled out to mainstream audience, why not the quantification of it?
Lee – it depends on what you mean by “unreliable”. ERA (or RA) is a perfect descriptor of past events. Batting average is a perfect descriptor of past events. UZR is not. tRA is not. Anything that deals with batted ball data is going to have retrospective uncertainty and should be displayed as such.
No one would argue that Zobrist was better than Pujols last year, because nobody owuld argue that Zobrist really played like a +27 run defender.
This seems to highlight a growing trend I’ve observed: people have a tendency to conflate a record of performance with an assessment of ability. Just like on offense, despite the axiom to the contrary, good defenders can have bad years.
Going from +5 to -5 does mean my true talent necessarily changed. After all, a given sample of performance (like a year) is just that, a sample from the population of possible defensive performances given your true defensive ability.
Too often, journalists and lazy analysts substitute the two, claiming that the metrics are bad because they suggest they player’s skill suddenly changed. But while UZR may be a good estimator of skill, it is still functionally a counting stat. Good players have bad years. Bad players have good years.
re, Barry Petchesky wrote is that “it’s foolish to jump aboard the Sabermetrics bandwagon.”
There’s a new version of cell technology that is rolling out that allows for faster data access than the one we’ve all been using. I guess, following Barry’s twisted logic, that it would be foolish to jump aboard the “interwebs over cell technology thing”.
We should all toss their smart phones in the trash.
Forget Jason Bay, FanGraphs owes me an apology. I defended UZR to many people last year, and was rewarded with my defenses being flat out wrong. The handling of this upgrade announcment has been poor, and it is extremely difficult to check where things have changed, as the old numbers are no longer available. From Tom Tango, we got a nice little piece on how the improvement hadn’t really affected that many guys, but that was flat out untrue. And Detroit Michael is right, if UZR is so unreliable and requires nuance, then don’t base such absolute definitions of value off of it. If FanGraphs expects us to trust UZR again, they need to A) admit they screwed up before this most recent change, B) explain where there continue to be issues so that we know how many grains of salt to apply where, and C) hire an outside statistics analyst to verity their explanation.
And after they do all that, have an intelligent BASEBALL person review the outcome for sanity. “Sanity checks” are done all the time in the professional world, they are vital when you dive so deeply into numbers that the grand scheme is lost to the grains.
That’s ridiculous. All Statistics are very useful and fun part of the game for me. Sam has it right, that we simply need to know more about what the issues are, so that we can draw our own conclusions, much like we do with OBP vs AVG etc.
“BASEBALL person” = anonymous scout?
Yeah, Fangraphs, run all of your ranking by LibertyBoy before publishing.
Dude, its a free website. They don’t owe you a damn thing. If you don’t like what they’re doing, stop visiting the site. Simple enough for you?
I’ve basically stopped using UZR, not because I think it’s wrong, but because I don’t know where it’s got issues. Since the site is Add-supported, I expect they’d like a lot of traffic, and I’m merely suggesting a way that they could get added traffic (assuming confidence in their product would lead to that). Also, since Tango, Marc, Carson and others here are respected in the sabermetrics comunity, I would think they’d want to do in this situation what they tend to do so well elsewhere, which is create great statistics with transparent methods that instills the confidence of peer-reviewed work.
You’re merely suggesting a way that they could get additional traffic? Is that what you meant when you claimed they owed you an apology? Come on now, no one is actually going to buy that’s what you were trying to do.
Like I predicted, shouted down.
He asked a free website to apologize to him because their numbers weren’t accurate enough in his mind. He deserves to be shouted down for that.
FanGraphs sells their rankings, or that was that big $9.99 ad on the front page my eyes lying to me?
Did you buy that package?
No you idiot, that would be Bloomberg Sports. Their only affiliation with this site is as an advertiser. Nice try though.
I now see you were talking about 2nd Opinion. I don’t see how that makes the site anything other than free, considering all the numbers that are being talked about here are available for free. Their fantasy companion has nothing to do with this.
Alex, Their apology won’t bring additional traffic. There admission of what is wrong with the stat will. Just like we all know what is wrong with stats like AVG, we’d like to have a clearer example of UZR. I don’t need an actual apology, but I would like the first couple of things I originally listed to be implemented, and I think it would help the site.
Just an FYI, Deadspin actually said:
“We are saying that it’s foolish to jump aboard the Sabermetrics bandwagon while they’re still working out all the kinks. ”
Not an unfair statement, really.
Then it’s foolish to jump on the bandwagon for any for of baseball research or analysis because they all have kinks.
I’m not sure how well that attitude works in other fields, but that author sure would have no place in the sciences…
Well, there’s a lot of silly bandwagon-jumping in the sciences too. Doesn’t mean it’s right.
His statement is perfectly valid. Nobody should be using UZR to make definitive statements about a player’s value, like so many of the wannabe-sabermetricians who read this site foolishly do. I don’t think even David would disagree with that.
Is it possible that Fenway, especially LF is unique in some way that makes UZR unreliable there? One angle I thought of is that the LF has to play shallower there so it’s more difficult to get an angle on balls hit down the line or in the gap.
Yes, and that is what was adjusted for. Bay and Ellsbury were the two biggest changes in UZR after the update. Micheal Cuddyer also got a boost.
For people who want to rail UZR for not being precise enough, what’s the better alternative? How confident are you that the alternative is more accurate the UZR? Just because its not perfect and has to be tweaked from time to time doesn’t mean it should be ignored.
The alternative is my eyes, and they’re pretty unreliable. I’d like to use UZR, but want to know where it will be deficient. Is that too much to ask?
How many games do your eyes watch? How many players do they see play the same position? If the answer isn’t somewhere in the thousands for games watched (in person too, because watching on TV doesn’t give you anywhere near a large enough picture of defense) and at least 15 players per position for large samples, then your eyes and brain simply don’t know enough to base ones talent level at a defensive position compared to the peers for that season. I am guessing short of MLB advanced scouts, no one watches enough baseball in person to make such statements with validity based on ‘in person defensive evaluations’. Sure, watching a guy like Adam Dunn play left field is easy to say ‘he sucks’. But what about most other guys? You simply can’t see enough of everyone else in person to make statements on 80% of the league based on what you can actually watch.
Wow, I didn’t know Deadspin writers promoted ignorance like that. Pretty surprising.
Good article.
But I also think there are a lot of people who do treat some of these metrics as gospel, and who do need to be more open minded to the possibility that there is other useful data and information out there, including the opinions of professional scouts, which are worthy of consideration.
And the biggest mistake I see being made by many who consider themselves fans of this site, is relying on any of these measures in small (single season) samples. Yes, that is partly user error. But, I think partly it is also confusion over WAR, and the way it is presented, and what it is really intended to be.
I see some defending WAR as being what actually happened, not meant to be a projection. But, for offense, something like WPA is actually a better measure of what actually happened in a given year. By basing WAR on stats like wOBA and FIP, it certainly seems designed to get more at underlying talent, even though it is being presented as though it were about single season contributions.
I think WAR might be more useful if you were to present a single projection line at the end. I realize CHONE is available there now (if you click to unhide it), but that only gets updated long after seasons end. If you want to avoid creating yet another projection, you could maybe just present a 162 game career average instead, which would be pretty close in most cases to a reasonable projection. Another possibilty might be a 3 or 5 year moving average. And include it on the last line in a way that stands out (not optional). At least it would be something to encourage those who are only going to look at one number, to look at something broader, that captures a bit more of the bigger picture.
While I do think some were off on Bay because of this, it wasn’t by much. Still, it shows in your fan projections, where Bay was projected as only worth 3.0 WAR, when a more disciplined look at the data might suggest more like 3.5 (or as CHONE projects, 3.7).
As a Met fan though, I saw this mistake being made even more this offseason with Jeff Francoeur. Looking at UZR, DRS, and TZ (Rtot/yr), they all have Francoeur as an above average defender, saving anywhere from 5.6 (TZ) to 8.7 (DRS) runs per year. Using UZR (+7.7), he has roughly league average overall defensive value (after the -7.5 positional adjustment). Even the Marcel projection had him only about a -5 run batter, while those which appropriately included aging had him near league average there (CHONE had a .327 wOBA). Almost any experienced sabermetric analyst should have projected Francoeur, still only 26 this year, as around a 1.5-2.0 WAR player. And yet there were many seemingly smart people this off season arguing against the Mets tendering him a one year contract at $5M – primarily based on not understanding how to properly use these tools.
So, you can blame the users, but I think that it’s a lot of these loyal users so often getting so many things wrong, often with an overconfidence that borders on arrogance, that is causing some folks like Petchesky to question the value of some of these stats.
It’s absolutely a good thing if people are having their confidence shaken in a metric that the creator himself doesn’t think should be given anything close to 100% confidence.
What’s not a good thing is people deciding less than 100% accuracy means it should be given 0 attention, or half-brained rants about how people shouldn’t be “jumping on the sabermetrics bandwagon.”
Agree. The Boston Herald article is actually very reasonable, and the Dead Spin one starts off reasonable, but the last paragraph, especially the swipe at Bill Simmons, is really uncalled for.
It’s one thing to caution against giving too much value to one stat, which is new and maybe is still getting the kinks worked out, but to dump on another writer for actually trying to understand these things and explain them is just dumb. Especially when the article lined to says this on UZR:
“The flaw: A tendency for wild fluctations in the year-to-year data. Like Carl Crawford’s UZRs since 2003: 14.7, 23.3, 15.5, 9.4, minus-1.2, 19.1, 17.6. So, um … what happened in 2007? Was he playing with a fake leg? Was he doing heroin? You’re better off taking three-year samples with UZR over grabbing one season.”
Maybe if Barry had actually read Bill Simmons, he wouldn’t have been so surprised by this adjustment.
When I asked about Fenway’s LF issues on both this website and The Book Blog last season, the response from both Dave Cameron and either Tango or MGL (can’t remember which one replied) was that there weren’t any problems with UZR in Fenway.
I pointed out that Bay hadn’t made an error all season, and that he led the AL in outfield assists at the time with 15. My argument was that he simply wasn’t an atrocious fielder – he was slightly below average, at worst – and that UZR’s Fenway left field issues made him look far worse than he was. I was basically laughed out of the comments section.
Now, shockingly, Bay goes from being one of the worst left fielders in the game last year to average, performance-wise, in the ’09 season.
So, it’s being less than honest for the rebuttal of both this website, and The Book, in the wake of the radical UZR re-adjustments for Bay and Ellsbury to be, “Hey, we always said you need three years of defensive data, and that you need multiple defensive metrics, and that none of this should be taken as gospel.”
Obviously! The intelligent commenters here already knew that. What WAS presented as gospel, to me, directly, in replies on both this website and The Book was: “UZR does not have issues in LF at Fenway.”
What other way can I possibly interpret that statement?
So, in review: I understant that we need to look at multiple years of defensive data. I get that we should look at multiple sources. I get that UZR isn’t perfect, and the website never claimed that it was.
Nonetheless, it would behoove fangraphs/The Book Blog to openly, transparently say, “Hey, these are the figures/ballparks that we have less confidence in. We are continually tweaking the figures in ballpark X, so take the figures there with a grain of salt.”
How hard is that? Is it arrogance that prevents the proprietor of UZR from doing that, or from answering my initial question of last year truthfully?
I agree Jacob. I know there were a lot of us that pretty much always assumed that there had to be something off in UZR caused by LF in Fenway. However, we were almost always told that this effect was either not there or insignificant. This is where some of us feel they failed. I do not care about people using Bay’s “value” incorrectly. I do care that they are using the problems with the latter to justify the former. Such a cop out argument is beneath fangraphs.
A much more interesting piece would be it turns out we now have reason to believe we were not adjusting for certain things enough. Here are some of the players and places that it had an interesting effect on (check out coco 2007 now etc). Here is what these changes mean for these players and UZR in general (ie proactively explain why people are going to overreact to these changes and Bay’s change instead of having to react to the situation). This is why some people, IMO justifiably, are questioning credibility. They likely would not be if fangraphs etc had owned the situation instead of the way they handled it.
walkoffblast: I ran the correlation between the previous UZR and the new UZR, I personally didn’t think it was going to be an issue because the two were so close. We had a post a week ago explaining the new UZR changes and MGL answered a bunch of questions in that post. Perhaps delving into the specifics of a few players is something we should have done. The Hardball Times independently ran correlation analysis of the new vs old uzr and came to similar conclusions. http://www.hardballtimes.com/main/blog_article/out-with-the-old-uzr-in-with-the-new/
UZR has always been a hot button stat and I think no matter what we did with the rollout there would have been a lot of articles on the subject about Jason Bay because it was a fairly controversial topic to begin with.
As for what you were told about Fenway, I can’t really answer that because I’m unaware of the details and when it comes to what was and wasn’t being accounted in UZR and its previous incarnation, I am not the proper spokesperson.
Thanks for the response. This whole thing is a relatively minor issue compared to what is the majority of awesomeness that is fangraphs.
“Perhaps delving into the specifics of a few players is something we should have done.”
That was basically my point in a nutshell. You are right. People were going to write about Bay no matter what. But, if you addressed it (you as in fangraphs not you personally) before they did, I was just suggesting it would have made you look better and could possibly have shaped the discussion in a way that was more productive than the deadspin type pieces. MGL did touch on it in the comments of that other piece but only after someone else brought it up and thus the tone already had become defensive.
For UZR as a whole this is a non-issue but I think the public perception is damaged. Why? Probably because true or not they feel like this whole thing was at first attempted to be swept under the rug. In their defense it is a little hard to believe that the adjustments were made for very specific places and then no one bothered to look at the obvious players this could have effected and thought hmm this might be newsworthy.
Regardless, anyone who uses this as an excuse to dismiss UZR’s usefulness in evaluating players is fruitlessly engaging in a two wrongs make a right scenario.
Agree completely… things like FIP are real good (in my view) for looking at performance moving forward and understanding whether a player will regress, but these stats now seem to be used (by some) for evaluating past performance.
I know ERA is not the best stat, but I don’t see using FIP for Cy Young voting – if a pitcher gave up a run on a single instead of a HR, it’s the same thing in terms of what happened – now if you are going to look at what that might mean in terms of future performance,it’s a different story, but I don’t think it should rework past results. A pitcher who strikes out 9 batter per game might be more likely to be successful going forward than a pitcher who strikes out 6 per game, but it doesn’t mean he was more successful that past year if they each recorded a similar # of total outs and hits and runs…. I understand there is a defense and luck component involved but there’s luck involved with HR’s as well (wind blowing out, game times, temperature, etc….)
Having something be luck-independent is not the same as having it accurately describe true talent in a given sample size. It helps a lot, but the definitions don’t really match up.
Ditto TCQ
I think looking at something like FIP for Cy Young voting is just fine, as long as you don’t just look at FIP and call it a day. ESPECIALLY when there is a disagreement between expected performance and true performance, I think looking more in to the details to figure out where the vote should go is the right course of action.
This goes back to the UZR issue. The problem isn’t necessarily the stat itself, it’s anyone who uses it lazily. Quick assessments do have their place, I don’t think every blog post on the internet or article in the newspaper should have to be built on extensive research and sometimes simply reporting what the numbers say to raise an interesting point is just fine. But I don’t see any reason to be lazy in awards voting, and I don’t see any reason to stop pursuing a point if you truly want to be as accurate as possible.
How come some players have a positive UZR total but a negative UZR/150?
Hey David,
Couldn’t find an email for you, so this will have to do.
Great points, and I certainly didn’t intend to disparage UZR in itself, but rather the sabermetric newbies who want to use it as the be-all-end-all of fielding prowess.
My full quote, which I take a little issue with you shortening, is “We are saying that it’s foolish to jump aboard the Sabermetrics bandwagon while they’re still working out all the kinks.” That last part is kind of important, because it’s still an emerging science, and may never stop being fine-tuned, and it’s too simplistic to point to UZR, or WAR, and say “this is the absolute measure of this player’s worth.”
I don’t think you, or most of your readers do this. But the more recent converts do, and my post was aimed more at them.
Thanks,
Barry
So what did you actually mean by the sentence in question then? It comes across as “you shouldn’t trust the numbers until they’re perfect,” which is a completely unrealistic expectation with any statistic. Honestly, if anything, the comment about working the kinks out almost makes it worse in my mind, as it implies there is nothing wrong with jumping on the bandwagon once the kinks are worked out. At least with the initial quote, I could take it as meaning don’t ever place too much trust in sabremetric stats because all stats have their faults.
Barry, thanks for stopping by. I agree with you on those points, just didn’t quite read your original article that way. But sounds like we’re really on the same page.
I think WAR and UZR are a great starting points for discussion, but right, other factors can and should be taken into consideration.
It’s very easy to look up a player’s UZR or WAR and use it in ammo for whatever point you’re trying to make on the internets, without any additional consideration.
Barry –
Maybe you shouldn’t rely on one year’s worth of data? Haven’t any of us stat nerds who live in our mother’s basement ever screamed “sample size” at you before? If not…
“Sample size!Sample size!Sample size!”
He isn’t. He’s disparaging those that are, and thereby misusing a stat that should in no way be used for single-season valuation of a player (but, unfortunately, is – by this very site).
Taking a 3 year average is meant to assess a players fielding ability. 1 year fielding is equivalent to perhaps 1/3 of a season hitting. A hitter can get off to a poor start, and his numbers would be lower than his true talent level, yet finish the season about where you would expect. Same is true with fielding.
However, this does not alter the fact that the hitter had a pretty poor stretch for 54 games. The offensive numbers duly record what happened since counting stats have zero uncertainty (and rate states based on counting stats).
UZR OTOH had Jason Bay as one of the worst OF’er in 2009, at -12.8 runs below average. That is what they said HAPPENED. As a Red Sox fan, I had argued he did not seem that bad, and it had to be how UZR handled Fenway. DRS, not available on FG at the time had Bay rated significantly better.
Now UZR says Bay was slightly above league average. That’s is a big spread for a stat that is supposed to say what happened. And if teams really are using stats like UZR and WAR to make judgements that affect players careers, and if anyone promotes such stats as valid for this purpose, you you better try and estimate the uncertainty in your numbers like those in the real world do.
I suspect MGL went overboard on his Fenway adjustments. The new and improved UZR has the Red Sox as one of the best defensive teams in the AL.
It ain’t so. Only those who have not watched any games would say otherwise.
DRS agrees with my eyes.
You need to use an average of DRS and UZR in your WAR calculations to smooth things out.
Also, when you introduced DRS, you studied DRS and UZR old and said there was good correlation. How about UZR improved. Where are the old UZR archived, it would have been nice to get an announcement before it was eliminated.
Furthermore, MGL was seeking opinions on how UZR could be improved only 1 month before such changes were actually made, that was on Tangos site. Perhaps the changes were not well thought out, or rushed to market due to DRS coming out on FG which allowed direct comparisons for the first time.
Well, if you want to blame me that the old UZR’s weren’t archived, then that’s fine. I didn’t see any reason to since the two were so similar and I still see no reason to do so.
The proposed changes MGL was asking about on insidethebook.com were not any of the changes that made it unto UZR this year, so that had nothing to do with anything. You realize there were 13 players for 2009 who had their UZR change more than 4 runs. 13! This was not a massive change by any means. The correlation between the two was .95. If UZR changed completely changed its methodology about players then it would probably be called UZR2 or something.
I’m sure there are teams that use similar methods for analyzing players. FanGraphs does not consult with any teams, so what teams wish to do with our data is purely at their own discretion.
How sure are you? Let me guess, not quite 100%.
This is becoming a time share pitch. I’m just glad I have a final closure over the host of internet smartasses who have been beating the UZR drone for a good couple of years. You guys sell these rankings, by the way, that makes you liable for the fallout. Reap it!
Sell them? We definitely do not sell the UZR or WAR rankings. We sold a book for $7.95 which didn’t even have UZR rankings in it, so I’m not sure what you’re talking about.
He apparently thinks you guys are somehow affiliated with Bloomberg sports since their advertisement shows up on your site. You shouldn’t pay him any mind.
I didn’t buy it. I just copy the B-R stat pages into excel.
But you do sell rankings, and people put stock in those rankings based on all the high-minded talk regarding new metrics like UZR, and the arrogance and disdain that you allow, on your site, to be flung at detractors of the new metric. A previous commenter said it in more polite terms than I can muster.
UZR has nothing to do with a Fantasy baseball guide they are selling.
Worst thing about the change are the detractors that’ll never shut up about it.
Thanks for the reply and confirmation.
I couldn’t realize anything without the archived data to compare it to.
Also, there were only 200 regular position players (exclude catchers and pitchers) who might have played enough where 4 runs is considered insignificant. That’s about 6.5% and if that’s every year, you could be looking at over 10% of regular position players over a 3 year period who had such an adjustment for one year.
PFT, That’s a great point. The refusal to admit that this update is a serious change is something that I think is going to give FanGraph and UZR a tough time. After touting UZR all of last year over at SoxProspects.com, I and everyone else have basically sworn off of it because we can’t be sure what data is good, and what is not. Simply writing a piece explaining that you understand that a good chunk of players (6.5% of regulars if PFT’s point is correct) were affected, and here are the known issues with the data. Presenting known issues is something that all beta and many final cut, and if your suggestion is that UZR is still in Beta, then it shouldn’t be too big a deal to offer hints at where it might require a grain of salt. To not do so David is arrogant and dangerous for your site’s reputation.
I don’t understand why at SoxProspects you couldn’t just relay the information as strongly suggesting something, but not definitively so.
I have the old UZR downloaded for 2002-2008 and I suspect that many other analysts do as well. If you would like to see the numbers email me at nicks0817@aim.com
Tomase is generally not that bad, but he’s clearly in over his head in this case.
And by the way, the quote you take from his article here, it’s 1000% wrong on Tomase’s part. When he says “these parts”, he means Boston, specifically the local media.
He’s conflating “constantly being brought up” with “taken as gospel.” The exact opposite is true. Outside of Alex Speier and maybe a few Globe pieces, anytime UZR was brought up it was either to disregard it or flat-out mock it.
Actually, come to think of it, Tomase referenced WAR in an article last year defending the JD Drew contract, so this is probably just Tomase feeling like he’s made himself look foolish by touting faulty metrics. Like I said, he’s in over his head.
“When I asked about Fenway’s LF issues on both this website and The Book Blog last season, the response from both Dave Cameron and either Tango or MGL (can’t remember which one replied) was that there weren’t any problems with UZR in Fenway. ”
Dave Cameron is a good baseball analyst who also happens to be basically a douchebag (with a funny voice), the sooner you accept this you will save yourself some aggravation.
Can we please just start banning people from commenting who are clearly adding nothing to the debate? LibertyBoy is not contributing remotely and is being actively hostile – is there even a reason he should be here? It’s one thing to disagree but actively engage your opposition, and it’s another thing to trash fangraphs on the hilariously inaccurate notion that this UZR update has something to do with paying which (obviously) undermines this entire site, best freely available statistics and all.
I also want to say something that is somewhat unrelated to the current debate: Why is David Appelman left alone to defend something that he did not design himself, but simply provided as data on his website? Where are all the analysts who used UZR frequently to ascribe value to players? It should have been one of them writing this one, not David Appelman. Because he has a sense of integrity, now he has tackled this issue head-on, and is the only one facing commentary on this topic that has more variability than Jason Bay’s UZR swing. Where is the sense of integrity on the part of the other writers?
My only problem here is the “what? we never said anything like that!” approach David is taking.
“When did we declare WAR as gospel?”
Insinuated it plenty of times. Ask Devil Fingers
Brett Gardner is better than Jason Bay.
So is Josh Willingham!
Bay will always be a LUZR.
I agree with many of you that far too many people use UZR as the be-all-end-all for rating a player’s defense. With defense more than any other facet of the game, I believe it’s vital to actually watch the player in action.
I’m a Mets fan, so call me biased if you must, but I’ve watched nearly every game this season, and I haven’t see one ball hit to left field that Bay didn’t reach where I thought “soandso would’ve made it to that one!” On top of that he’s made several “webgem” plays, including a diving catch just yesterday – and hasn’t made an error since 2008.
Bay turns routine flies into desparation dives (or singles and doubles). So yes, you are biased. But not as a fan, but because your data set and analysis does not compare Bay with other LFs. You are merely comparing Bay with himself, and what you perceive to be balls that he sould or shouldn’t reach.
then why was he better than most leftfielders last year? Maybe your own eyes are telling you BS?
I’ve only made one comment in this thread and it was entirely negative, which I don’t feel good about. This site has contributed a lot to my understanding and deeper enjoyment of the game. Hard to put a price on that for me, and yet, the daily trips here are free. Pretty awesome.
I also appreciate Mr. Appelman graciously replying within this thread.
Trying to be constructive, as something to think about going forward, I think a great article for MGL to write for this website would be:
“UZR hot spots” – three zones that we struggle to get accurate UZR data for:
1. Fenway LF
2. Pac Bell RF
3. Bermuda Triangle in Sun Life Stadium
(Last two are just completely random guesses – I’m trying think about parks with quirky angles, dimensions or caroms).
I think that would be a great article, with MGL then giving us some analysis like, “although I have 90% percent confident in my ratings at most major league stadiums and their particular zones, I’d place my confidence level for the data in Fenway LF at 70% and 85% for RF at Stadium X,” etc.
I don’t think there would be any backlash with future tweaking and minor corrections to UZR if the community simply knew these “hot spots” up front and that they were in the process of being dealt with.
Ive been reading these responses (something I never do) only because of my love of baseball and my interest and amazement at the neverending focus on, and use of “sabernetrics,” anf all of these other attempts at using statistics in an evr increasing level of complexity to evaluate the performance and skills of athletes.
I’m not naive–I fully understand that statistics is an intrergal part of baseball, and has always been used, to some degree, to evaluate players–ERA, OBP, BA etc . . . (as in all sports). For the avrage fan statistics are used and understood. However, over the past years the use of ever increasing convaluted mathamatical equations at many levels of baseball has become a problem. Some teams have jumped enthusiastically on he bandwagon, including the Red Sox–to my dismay, since I have followed them for years. Not to get off point, but I wanted to point out that even the team I love and follow I believe is turning good intentions into a quasi-treligious belief in any and all numbers and equations!
As I’ve mentioned stats have there role, but let’s not forget that this is a “sport” with human beings. An analogy I will use is when academics in the Humanities (like History), increasingly attempt to use stats, equations and science to analyze people. It can help to an extent–but math and science can only answer sme questions, and many times they over-analyze “human” issues and miss the point.
Yes, use stats and other scientific ways to evaluate players, in their skils and performance. But, baseball is a human endeavor that requires people to use hier eyes and judjement. When a complex equation tells Red Sox execs, sports writers, and fans something that goes against their experience on a daliy basis, and even is counter to the other more “mundane” stats that have been used and trusted for decades, no offense but the geeks have gone too far!
Iwatched him play all year, and I saw a slightly above average let fielder–for Fenway with only one year–the experience of watching Yaz, Rice dominate the wall, and a multitude of opponents misplay the wall and angles weekly demonstrated what the following statistics will back-up.
In Boston, especially under the new regime, winning two world series, although great, clouds many Red Sox fan’s and other Bostion’s judgment. UZR stats among others seem to blind all to reality. Yet, if stats are King, let’s take a look!
Aside from Bay’s team leading offense, talk of him being la lazy outfielder, with lacking skills seemed to abound. Here are some stats!!!
-Started 150 games and was 2nd in the league in innings played (all league rankings)
-He was 3rd in outfield chances and 4th in putouts
-He was 1st in assists with 15 (10 at 2nd, 1 at 3rd and 4 at home)
-He was tied for 2nd with 2 doble plays and had ZERO errors
FANCY NEW STATS
-Rtot 3rd in league with a 7.4
-RTZ 10th in league with a 4.9 (only 4 of above 9 were 100 game starters)
-ROF3rd “” ” with a 3 ( 1st and 2nd had only 35 starts each)
-Rtot/yr 35th ” ” with a 10.6 ( * only 3 of above 35 were 100 game starters)
-RTZ/rd 7th ” “with a 4.5 ( 2 of 6 were 100 game starters)
ALL OF THE ABOVE STATS HAD BAY ABOVE THE LEAGUE AVERAGE.
-HE WAS ALSO 2ND IN DEFENSIVE CHANCES AND 2ND IN PUTOUTS, THROWING OUT 6 RUNNERS TRYING TO ADVANCE A BASE ( LAEGUE AVE WAS 4)
– HE FIELDED 743 BALLS, 4TH IN LEAGUE ( AVE WAS 713)
***ALL THESE STATS CAME FROM THE MLB WEBSITE. I POINT THEM OUT FOR A REASON–IT IS UNREASONABLE TO ASSERT THAT BAY WAS A BELOW AVERAGE FEILDER. THE MAIN ARGUEMENT I’VE SEEN IS THAT FENWAY IN LEFT IS EASY TO PLAY–I BELEIVE THAT IS FOOLISH. THE ONLY BENEFIT FRM PLAYING IN A SLIGHTLY SMALLER OUTFEILD IS THE POSSIBILITY FOR EASIER PUTOUTS CAUSED FROM PLAYING SHALLOW. HOWEVER, ONE STILL MUST PLAY THE WALL , WITH ALL ITS ANGLES AND STRANGE BOUNCES, AS WELL AS THE CORNER AND THE STANDS THAT JUT OUT BEYOUND THIRD BASE RIGHT ON THE FOUL LINE. ALSO, ONCE YOU LEAVE THE LEFT FEILD CORNER IN BOSTON THAT 37 FOOT WALL IN LEFT CENTER IS 380 FEET, AND A BALL THAT HITS OFF ITS FACE IS ANGLES TOWRDS CENTER. THERE IS AN OLD ADAGE, WHAT THE WALL GIVES, THE WALL TAKES AWAY. i BELIEVE ITS ONE OF THE MOST DIFFICULT POSITIONS IN THE OUTFIELD TO PLAY IN ALL OF BASEBALL.
YET, BAY HAD GOOD, IF NOT EXCELLENT NUMBERS, EVEN MANY OF THE NEW ONES USED TO RANK FELDERS. HOWEVER, BEACUSE OF THE SOX DEFENSIVE ISSUES LAST YEAR, MANY LOOKED FOR SCAPEGOATS, AND THE HIG PAYED BAY LOOKING FOR A NEW CONTRACT WAS AN EASY TARGET. SO, THE SOX MANAGEMENT USING THEIR “VOODO” STATS THAT THEY CHERRY PICKED TO SINGLE OUT BAYS SUPPOSED WEAKNESSES, WAS PICKED UP BY BOSTON’S NOTORIOUSLY FICKLE MEDIA, AND BEFORE OUR EYES A MAN WHO PLAYED MOST OF THE GAMES, WAS A LEADER IN PUTOUTS, ASSISTS, HAD NO ERRORS RANKED HIGH IN MOST OF HE ABOVE MENTIONED “NEW “STATS,” WAS LABELED A POOR FEILDER. (REMEMBER MANNY!) BAY REPLACED HIM, AND HIS OFFENSE WAS CLOSE 36 HR 119 RBI, WITHOUT A 25MILLION PRICETAG. IHAVE NOTHING VESTED IN BAY, I JUST AM BOTHERED BY THE B.S.THAT OCCURS WHEN THEO AND GANG WNAT TO RID THEMSELVES OF A PLAYER FOR CONTRACT REASONS, AND THIS BOGUS CAMPAIGN IS WAGED TO MAKE HIM INTO THE WORST FIELDER IN THE LEAGUE!! BASED ON QUESTIONABLE DATA. THE STATS MENTIONED ABOVE SHOW THAT HE WAS A ABOVE AVE PROFESIONAL OUTFIELDER, IF GIVEN A CHANCE TO PLAY THE WALL FOR AS LONG AS MANNY, HE COULD HAVE BEEN GOOD. IN 2009 HE WAS THE BSET OFFENSIVE PLAYER ON HE SOX, UNQUESTIONABLY. 2ND IN HR AND RBI IN THE AMERICAN LEAGUE.
BASEBALL PLAYERS WHO HIT, FIELD, HUSTLE FDAY AFTER DAY–REAL FANS AND BASEBALL SCOUTS KNOW IT WHEN THEY SEE IT. MATHEMATICIANS AND COMPUTER GEEKS CREATING COMPLEX EQUATIONS ABOUT WAHT A BASEBALL PLAYER IS AND SHOULD BE IS JUST ANOTHER PERVERSIONOF THE GAME.
MONEY RUNS THE GAME–THE OWNER HAVE NOW, AND ALWAYS HAVE HAD THE RIGHT TO MANIPULATE THEIR TEAMS–THAT’S LIFE. THEY DO SO BECAUSE THEY ARE UNABLE TO PLAY THE GAME OR EVEN UNDERSTAND WHAT ITS LIKE TO BE AN ATHLETE. IF IT LOOKS LIKE A PLAYER, SMELLS LIKE ONE AND PLAYS LIKE ONE . . . . .ALL THE B.S “SYSTEMS” INVENTED BY THE QUASI-SCIENTIFIC HUCKSTERS AND SWALLOWED BY SOME IN BASEBALL MY HURT THE CAREER OF A JASON BAY, BUT WILL NEVER BE ABLE TO “BOTTLE” OR COPYRIGHT THE TALENTS OF A RUTH, A MANTLE, A WILLIAMS OR A JETER.