Master Index of Archived Threads


Value of Defense at Catcher

Rotblatt Apr 05 2006 11:38 AM

JB asked me to start a new "Catcher Defense" thread from the "I woke up this mawnin'" thread (which, by the way, should REALLY have been conducted solely in blues verse). Generally, stats-guys have had a hard time quantifying the value a catcher brings to the table defensively outside of his ability to throw out basestealers. Ability to control passed balls & wild pitches, call the game, field bunts, and handle plays at the plate are all incredibly difficult to quantify. Throwing out wannabe base stealers, however, is pretty easy, although there has been some research done on the role a pitcher plays on that front. I suppose this could be a catch-all thread (get it?) for discussing whatever articles we turn up on the subject as well as our personal thoughts. Incidentally, I also caught an article I had missed before about Piazza's CS% in [url=http://www.baseballprospectus.com/article.php?articleid=2888]2003 & 2004[/url] being surprisingly close to average. Anyway, here are the articles I posted in the other thread, along with a few other ones I ran across: Here's data for you, Sal. You may have seen most or all of this in the past--I'm not sure what I've seen on CPF or what I've found elsewhere. Basestealing Piazza sucks. He cost us [url=http://espn.go.com/mlb/columns/bp/1202793.html]18 runs[/url] between 1998 & 2001. I'm sure he's declined since then. Calling games [url=http://www.baseballprospectus.com/article.php?articleid=432]This study[/url] is from 2000. Basically, it finds that "catchers do not have significant differences among their game-calling abilities." Now, lest someone call that a cop-out, Piazza led the league in CERA (the metric they designed to measure the ability of play-calling) in 2006, and did NOT crack their bottom 10 Career Catchers--although he also didn't crack their top 10 Career Catchers. Basically, if there's an ability to call games, Piazza's probably in the middle third according to this study. Blocking balls According to [url=http://www.hardballtimes.com/main/article/quantifying-catcher-defense-and-other-stuff-like-that]this study[/url] from HT, he was neither in the top 10 nor the bottom 10 in 2005. Again, call it the middle third. If anyone's got any other articles to point to, I'd be interested. I should reference some of the other BP findings, which basically call the defensive impact of a catcher negligible: ]However, if we believe the results from this study, namely that catchers do not have significant differences among their game-calling abilities, the implications are staggering. First of all, the much-maligned stats we’ve been using for years to evaluate catchers--runners thrown out and passed balls, might actually quantify their defensive value. Furthermore, the relative unimportance of the running game could prompt teams to shift better offensive players to catcher without hurting the team’s defense. You open up another position on the field besides first base for prospects who don’t have the reflexes to play the infield, nor the speed or instincts to play the outfield. The positional is still physically demanding to play, but you could potential keep two dynamite offensive players in the lineup--say Mike Piazza and Frank Thomas, but swapping them between C and 1B so neither gets overworked behind the plate. Far from being the position with the lowest expected offense, it could flip to the other side of the defensive spectrum entirely, and become a place to hide a slow-footed slugger. I personally think that the value of defense at catcher is greater than negligible, but statistically speaking, they haven't found a way to measure it yet. I'm very insterested in what [url=http://www.hardballtimes.com/main/printarticle/the-fielding-bible/]The Fielding Bible[/url] has to say on the matter. Here's [url=http://www.hardballtimes.com/main/article/the-crucible-of-competition]another article[/url] from HT referencing the Fielding Bible. Diamond Mind [url=http://www.diamond-mind.com/articles/hldthr97.htm]Catcher Throwing and Pitcher Hold Ratings[/url] Baseball Prospectus [url=http://www.baseballprospectus.com/article.php?articleid=436]Catching Up With the General: A Postscript[/url] [url=http://www.baseballprospectus.com/article.php?articleid=1489]Aim For the Head[/url] [url=http://www.baseballprospectus.com/article.php?articleid=1343]Controlling the Running Game: Is it the Catchers, or the Pitchers?[/url] [url=http://www.baseballprospectus.com/article.php?articleid=2561]The Man with the Golden Gun: 2003's Most Valuable Catcher Arms[/url] Hardball Times [url=http://www.hardballtimes.com/main/statpages/catchers_midseason/]Midseason Catcher Stats[/url]

old original jb Apr 05 2006 12:30 PM

I agree with the ESPN article that the devaluation of the running game is an oversimplification. If the catcher against whom you are running is poor enough, it pays to run, and catchers who can't control the running game will surely shift the odds in favor of base stealing enough to cost their teams runs. It should be possible to plot the CS% against run impact/chances to find a breakpoint at which if a catcher's CS% is below it, it will pay for the average guy to run on him. I would guess this would occur somewhere aroung a CS% of 30%. The ESPN chart does not normalize for total attempts because they are interested in actual runs given up, not potential to give up runs; actually there is an added complication in that the impact of a poor CS% is increased when teams get wind of it and run more while the impact of a good CS% may actually be reduced in calculated runs prevented but increased in reality when tteams choose to run less against good catchers. Whether to build your team around base-stealing is a different matter, especially if you are giving up other tools in favor of speed. On average, it is not a good idea. But it might make sense for a team in a division (and therefore having a schedule) with a lot of poor catchers to calculate their own schedule adjusted opponents CS% to get an idea of whether they would benefit markedly from making the running game a somewhat bigger part of their plans.

MFS62 Apr 05 2006 01:51 PM

Great post. Lots of info there. I just had time to scan the first linked article. I'll replace this post with some comments after I've had some time to read it more carefully. But I just wanted to give you some kudos. Later

Edgy DC Apr 05 2006 02:09 PM

My general problem with devaluation of the running game is that baserunner runs are more crucial than your typical run. You can say that Rickey Henderson only accounted for 12 runs stealing (or Bob Boone saved 12 runs throwing), plug those runs into a pythagorean equation, and say, "Eh, doesn't amount to much in terms of wins." But stolen base runs greatly tend to come in game situations where one team is within a run or two of the other, and those runs are far more likely to be the difference than a run (a theoretical run as measured in the Runs Created statistic) contributed by hitting. Hits come when they come. Some come in close games, others in blowouts Not so with steals.

Frayed Knot Apr 05 2006 02:28 PM

Can we all agree that Piazza was "Death on Pop-ups"?

Rotblatt Apr 05 2006 03:15 PM Edited 1 time(s), most recently on Apr 05 2006 03:28 PM

Edgy DC wrote: My general problem with devaluation of the running game is that baserunner runs are more crucial than your typical run. You can say that Rickey Henderson only accounted for 12 runs stealing (or Bob Boone saved 12 runs throwing), plug those runs into a pythagorean equation, and say, "Eh, doesn't amount to much in terms of wins." But stolen base runs greatly tend to come in game situations where one team is within a run or two of the other, and those runs are far more likely to be the difference than a run (a theoretical run as measured in the Runs Created statistic) contributed by hitting. Hits come when they come. Some come in close games, others in blowouts Not so with steals. Good (and very cutting edge) point. That's where the concept of win probability comes in. The good people at [url=http://www.beyondtheboxscore.com/story/2005/11/8/225921/879]Beyond the Box Score[/url], as it turns out, have addressed the question of stealing in late and close games using win probability. ]I'm going to use this little toy to analyze a something that has bothered me for a long time: the runner on first in the bottom of the ninth. Nothing is more aggravating than watching the tying or winning run sitting on first base while the closer K's the next two batters and gets the last one on a popout. Well, actually, it is more aggravating when the runner is nailed trying to steal second base. . . . . If you're down by two runs, you better be damn sure that the runner gets a good jump, since you need at an 87% success rate just to break even, even with nobody out. If there are two outs, the breakeven stolen base success rate is an astounding 98%! At the other extreme, the breakeven point is only 60% when there is a tie game and nobody out. This is well below the gross-sabermetric-average breakeven rate of 70%-80%. The results are surprising for the "Down by 1 run" data set; the marginal rise in the breakeven point is actually quite low across the number of outs. One can explain the situation for one out by considering the possibility of a double play ending the game. In a tie game, the double play is not as bad as you might think, since the home team still has the opportunity to bat again in extra innings. The trend for two outs is harder to explain. Honestly, I expected the two-out situations to converge to a very high breakeven point as it did with the tie game and down by two scenarios. My explanation is that moving the runner to second is far more important when down by one run because a single - the most common type of hit - will often score the runner. When down by two runs, plating one run is not as important and protecting a tie game is important because the home team always has the opportunity to bat again. To extrapolate: when it's close and late, scoring a run is more valuable. Stealing 2nd (or 3rd) in those situations, which increases the probability of scoring a run, is more valuable. Therefore, preventing a runner from advancing is ALSO more valuable. How much so? Well, if we're away, and we're up by 1 run in the bottom of the ninth, with THEIR runner on first, our win expectency is 68.5%. If they successfully steal, it's 58.8%. If we gun them down, it's 89.6%. That's a swing of 30.8%--pretty darn big. What this credits is the idea of late-inning replacements in key spots. Substituting a Piazza with a Molina in the ninth makes a boatload of sense if the opposition has got a couple Pierre's & Reyes's coming up.

Elster88 Apr 05 2006 03:19 PM

Frayed Knot wrote: Can we all agree that Piazza was "Death on Pop-ups"? What does that mean?

ScarletKnight41 Apr 05 2006 03:21 PM

I think it means he was aggressive about going after pop-ups and he was generally successful getting them. Which seems to be a fair statement.

Johnny Dickshot Apr 05 2006 04:50 PM

I'm waiting for a park-adjusted for fair-territory, first and third basemen-range-influnced analysis of a catcher success rate in Popup Situations before we say that.

Bret Sabermetric Apr 05 2006 04:54 PM

A shocker: I saw him drop or misjudge a popup or two.

Frayed Knot Apr 05 2006 08:40 PM

Elster88 wrote: ="Frayed Knot"]Can we all agree that Piazza was "Death on Pop-ups"? What does that mean? It's a joke (far too inside really) referring to a certain long-ago poster who would defend Mikey at all costs. S/He decided he was unusually good at tracking pop-ups and so used that particular phrase almost as if a nickname.

KC Apr 05 2006 09:00 PM

JD: >>>I'm waiting for a park-adjusted for fair-territory, first and third basemen-range-influnced analysis of a catcher success rate in Popup Situations<<< Gotta break that out into day/night too, sun/clouds, what color the home crowd is wearing to distract popup situations.

Bret Sabermetric Apr 06 2006 04:14 AM

Frayed Knot wrote: ="Elster88"] Frayed Knot wrote: Can we all agree that Piazza was "Death on Pop-ups"? What does that mean? It's a joke (far too inside really) referring to a certain long-ago poster who would defend Mikey at all costs. S/He decided he was unusually good at tracking pop-ups and so used that particular phrase almost as if a nickname. Which, if you think about it, is grossly insulting to Piazza. I mean, to single out a fairly easy play that few catchers fuck up more than once a season --well, that really doesn't say much for his fielding, does it? "Death on Bunts" I could see, because a catcher who can cover a wider range, maybe cover an extra foot or two of fair territory and still get off a throw --well, that's a valuble skill. But tracking popups? 90 % of the time you simply throw your mask off and wait for it to come down. The rare plays, the ones where the catcher goes sprawling horizontal while running at full speed towards the stands, or into the stands--those would be good plays, though Piazza never did one of them (while I was watching) in his 7 1/2 years here.

Johnny Dickshot Apr 06 2006 05:07 AM

There was game in LA where where he made a parellel bar dismount and wound up in the first row.

Bret Sabermetric Apr 06 2006 05:41 AM

No, that was Jeter.

Johnny Dickshot Apr 06 2006 05:45 AM

No, really. Piazza was fine on popups. How do you figure James arrived at the C- by the way? I figure relative to the league, Piazza's throwing arm gets a D or worse, which would indicate the rest of his game scored enough Cs to pass him.

Johnny Dickshot Apr 06 2006 06:03 AM

Note how Piazza allows this guy a clear path to home plate. He's practically in the dugout! Observe the wussy sweep tag he always uses He does anything to avoid contact: What a pussy! Could he be any more far away? Always afarid, I tell you what:

Bret Sabermetric Apr 06 2006 06:04 AM Edited 2 time(s), most recently on Apr 06 2006 06:33 AM

It was a C+. I'l look up his methodology. Check out the UMDB memories [url]http://ultimatemets.com/profile.php?PlayerCode=0120&tabno=7[/url] of Grote's defense for the kinds of things i found lacking (or at least no evidence for) in Piazza. The things that people say are "positive" for Piazza don't impress me much: when a pitcher says, for example , that Piazza was easy to throw to, I think "Well, no one would say that throwing to Grote was easy. After a game of throwing to him, you felt worn out, hassled, challenged, but I don't want my pitcher thinking 'That was easy'." Piazza was fine on pop-ups. That's my point. Is there a catcher who isn't 'fine' on popups? I want a catcher who's hell on popups, who catches everything possible and a few popups that aren't possible. I want a catcher whose uniform front is brown from all the popups he goes sprawling into the dirt after. I want a catcher who goes into the third row twice a game, I want a catcher who's off and running full-tilt for popups 20 feet off the first base bag. One of the Grote memories that provided a reminder of Piazza's lack of hustle behind the dish had to do with backing up first base on a grounder to ss or 3b. Grote was full-tilt running on every grounder (with no other runners on) to back up first. With Piazza, if a throw got away from the 1bman, the ball would bounce around foul territory while the runner would breeze into second every time.

Bret Sabermetric Apr 06 2006 06:11 AM Edited 1 time(s), most recently on Apr 06 2006 06:41 AM

BTW, I wasn't trying to avoid your five part post in the other thread before it turned into a Rusty-fest. I answered your points one-by-one but it got swallowed by my computer at work yesterday and when I got back to it, the thread had devolved. one of my points, before even seeing your fine collection of Piazza photos, was that your memories and mine are necessarily anecdotal. You may have seen more games of Piazza than I have (though I'm not convinced of that, I'll concede the point) but I've seen enough (and have taped enough) examples of him setting up before the outfielder has even released his throw with both feet on the grass part of the infield and then doing a swipe tag when the throw came in where he was standing to start me thinking this way in the first place. Watching such a play, you think "Hey, Piazza HAD to go get a wild throw and then return too late to nab the runner" but usually it was his choice that made that play work out that way. Obviously I'm not saying that he literally never got into contact at the plate, just that he'd do things to prevent contact from happening.

Bret Sabermetric Apr 06 2006 06:30 AM

James' 4 point method for catcher's defense (in ascending order of importance) caught stealing %, fielding percentage, passed balls and successful bunt hits.

KC Apr 06 2006 07:01 AM

Fine pictures, JD - you can lead a mule to water but you can't make him drink.

MFS62 Apr 06 2006 07:05 AM

Some comments: I was surprised by the CERA rankings. They show that Charlie O'Brien wasn't as statistically good a defensive catcher as the reputayion for being one that kept him in the majors for so long. He uses certain batter stats to calculate the effectiveness of a catcher, but I didn't notice ERA. IIRC when Piazza came to the Mets it was reported that the ERA of his staffs in LA had about a haf a run lower ERA when pitching to Mike than it was when the same pitchers threw to his replacement receivers. I would have liked to have seen a similar comparison with the catchers who were rated highly using CERA. Two of the lower rated catchers, Parrish and Stanley were moved to other positions (1B, DH). I guess their managers didn't need the CERA to know that they weren't that good defensively. The list of guys who started out as catchers and were moved before they could amass the number of games to qualify for his evaluation includes Rico Carty, Joe Torre, Dale Murphy (what is it with those fucking Braves?) and Craig Biggio. When considering popups, are catchers rated on "range factor" too? It is important. While, as was noted, few catchers miss many popups. But there is a disparity in distances they have to run in vatious ballparks. it used to be over 50 feet to the backstop in Forbes Field (Pittsburgh) and the foul territory in Oakland is huge. Getting to a popup is therefor more important than being able to catch it. Range factor would be informative here. As for being able to throw out runners, some catchers had the misfortune to be the receiver for staffs that were terrible at holding runners close or were breaking ball pitchers. If you have three starters who allow the runner an additional first step or two, or the ball doesn't get to them quickly, they probably won't have a high success rate, no matter how good their arm. Finally, wrt the comment above about backing up first on infield grounders - maybe Mike was running down the line just as hard as Grote. It (to paraphrase a Tom Seaver line) just took him longer to get there. Later

Bret Sabermetric Apr 06 2006 07:57 AM

KC wrote: you can lead a mule to water but you can't make him drink. I saw what I saw, KC, and I taped what I taped. When we were at that Met game a year ago, in deep right field, do you remember I asked you to look where the guy set up on plays at home plate? It's not like I'm making this stuff up long after no one could verify it, with my mind made up, I was curious and I was observing. I'll drink--I'm just not drinking Koolaid. I'm not sure if CERA is credited when replacement catchers played as rarely as Mike's replacements did--small sample size plus staff turnover from year to year and all that. What I'd very interested in would be the w/l pt in games he started vs. those he didn't. Youi'd think the Mets would get killed with their best hitter out of the lineup, but I wonder what the degree of decline was in the 100s of games that, for one reason or another, he didnt start in. This is actually a very popular and useful stat in hockey (+/- when a player is on or off the ice) and I'd live to see it in wider use in baseball.

Benjamin Grimm Apr 06 2006 08:02 AM

Interesting idea for the UMDB. I can easily add a stat for record in games started by position player. (Games they didn't start would be more tricky, since I don't track when they're on the roster.) I'll put that on the to-do list.

Bret Sabermetric Apr 06 2006 08:09 AM

Wow, excellent. Instead of correcting my typo above, I;'ll simply note that this is a stat that apparently I WILL live to see. Of course you could do "games not played" by subtracting "games played" from 162 (most years) and figuring out that w/l pct.

Johnny Dickshot Apr 06 2006 08:14 AM

Interestingly, "range factor" doesn't attempt to measure "range" as in distance, but as in "success over range of opportunity." I was only busting your balls with the photos and obviously being sarcastic. Like I said, it'd be interesting to know if certain catchers set up differently, and if so whether they were trained to do it in such a way (since other defensive positioning plays tend to be consistent with the team); or if -- my perception -- a catcher sets up more to the situation than a set plan, and that's all before you weigh the merits and risks (outs gained or lost, injuries caused and prevented) of specifically choosing one style over another -- and whether THAT even matters over the course of the year (I would guess: probably not). So, before coming to some conclusion that can be boiled down to "Piazza is a wuss whose defense beyond throwing-out basestealers is costly" I think you're going to have to show it. And like I said, an Overall C+, considering a 'D' arm, indicates to me that James doesn't find Piazza's fielding pct, PBs and bunt coverage to be grossly below average. That he doesn't even address "home plate coverage" as a metric worth measuring ought to say something too.

Benjamin Grimm Apr 06 2006 08:22 AM

Bret Sabermetric wrote: Wow, excellent. Instead of correcting my typo above, I;'ll simply note that this is a stat that apparently I WILL live to see. Of course you could do "games not played" by subtracting "games played" from 162 (most years) and figuring out that w/l pct. True, but I'd leave that to the visitor to do mentally. In Piazza's case, for example, it could be confusing. In 1998, do his games not started include the games the Mets played before the trade? Since some might assume yes and others no, I'd rather just go with the more clear numbers.

KC Apr 06 2006 08:22 AM

I was just joking too BS, you really need to take it down about 6,000.

Frayed Knot Apr 06 2006 10:08 AM

Except that, unlike the plus/minus ratings in hockey which is derived from in-game stats, attempting to judge a player's worth on a team's won/loss over the course of the year involves so many other factors that I don't think it'll tell you a whole lot. Hot & cold streaks involving other players, key injuries, mid-season trades, etc. have to be factored in or you're not measuring what you're setting out to do.

Bret Sabermetric Apr 06 2006 10:29 AM

KC, I recognize your light tone, but you were drawing some humorous analogy between me and a stubborn, obstinate, unreasonable, mean-spirited cantankerous animal, weren't you? Just for the record. Dickshot (and anyone else), before Yancy gets his new stat together, would you care to take a guess on the w/l pct dropoff between Mets game with Piazza in the lineup 1998-2005? I'd say it would be fair game to include the pre-Piazza 1998 Mets, since the team was pretty much the same, other than him, but we can set it up with any parameters you choose. This is the bottom line issue here, as far as I'm concerned, because I think that there was a dropoff but far less of one than Piazza's fans think there was. When I was advocating trading him for younger players after 2001, that was the nub of my argument: if the Mets's dropoff is only .xxx percent when Piazza's on the DL or otherwise not playing, wouldn't it be even less if we dropped him and gained the players coming back in a deal? How good would those players have to be to get us showing a positive percentage for such a deal? I posit that it's a smaller percentage than most Mets fans think. If the Mets played (WAG) .475 ball over those 8 seasons, maybe they played .472 ball when Piazza didn't start. My reasoning is that say Piazza's offense is worth, say 7-3 in a good season. If his d is as mediocre as I think, say that's another 2-4, making him 9-6 overall. Still a good overall rate, still helpful, but not the spectacular Schmidt-like 12-3 that you think you're getting. (Schmitty was a hell of a defensive player, as good or better than Piazza with the bat, and a good baserunner, though slow). If you're think his d is more like 3-3 or even 4-2 then you're evaluating him right up with Schmidt and Babe Ruth, and I think deceiving yourselves as to his true value. That's the crude form of my argument at any rate.

Rotblatt Apr 06 2006 11:17 AM

Attempting to reduce a complicated concept like measuring a player's impact in a game by looking ONLY at W-L record is frankly silly. It's like looking at Reyes' 99 runs last year and concluding that he must have reached base a lot. I'd suggest looking at Win Shares, VORP, etc.--whatever metric floats your boat.

KC Apr 06 2006 11:17 AM

Nice pics JD, you can lead a stallion to a gold platted trough of purified spring water but you can't make him partake in the pleasure.

old original jb Apr 06 2006 11:29 AM A theory about models to predict defensive impact:

Predictive models based on the sum of the predictive statistics of individual players calculated using league averages will work best to predict average performance and will [u:202a443da8]underpredict[/u:202a443da8] extremes in either direction. The estimates of how likely it is that making or not making a defensive play will prevent/lead to a run surrendered are based on the average of the situation accross all teams. But if you are on a very good defensive team, it is likely that your mistakes will have less impact and if you are on a bad defensive team they will have more; conversely, on a bad defensive team, a star defender may save more runs than on a great defensive team. The worst combination would be more than one well below average defender plus below average pitching. You would expect adding a poor defensive player on such a time to have a much worse impact, and that such teams would do much worse than predicted on the basis of the sum of the impact statistics of their players. Therefore, the balance of offense/defense value is profoundly affected by how far below or above average someone is and how far below or above average his teammates are. Adding a slightly below average player to a good defensive team would have minimal impact. Adding a well below average player to a below average team could be the straw that breaks the camel's back.

Johnny Dickshot Apr 06 2006 12:07 PM

My guess would be the record is more or less the same, a few points up or down, and agree that it wouldn't say much nor show cause and effect.

Bret Sabermetric Apr 06 2006 01:31 PM

Piazza aside, don't you agree then that the impact of individual players on a team's performance over the course of a season is overstated, usually in proportion to the evaluator's sophistication? Somewhere James makes this comment: If a team loses its big star for two weeks, that team tends to get by surprisingly well while he's out. That's one of the bases for my argument to trade stars having good years: If trading Star X actually costs you only 2 or 3 games over the course of the season (after you've gotten far behind in the pennant race anyway) and trading him lets you get out of an unfavorable LT contract, get younger, get more versatile, look at your own players in Star X's role a little bit, etc., then it may be worth doing. Fans typically respond "NOOOO!!!!! It will kill the team to deal off Star X for a lot of garbage, and we still haven't mathematically been eliminated until the fat lady has sung yyybbb." Individually, this might be true, but when a team persists in hanging onto veterans year after year to make pennant runs that either don't happen at all or fail, well, I have a problem with that, as you may have noticed.

Johnny Dickshot Apr 06 2006 01:39 PM

Bret Sabermetric wrote: Piazza aside, don't you agree then that the impact of individual players on a team's performance over the course of a season is overstated, usually in proportion to the evaluator's sophistication? Well, I prolly agree with the first part though every situation is different. I noticed your other problems.

Rotblatt Apr 06 2006 02:01 PM

="Bret Sabermetric"]Piazza aside, don't you agree then that the impact of individual players on a team's performance over the course of a season is overstated, usually in proportion to the evaluator's sophistication? Sure, but that's why we have metrics--to look objectively at the contributions of each of our players. I can look at the 2005 WARP3 scores, and verify that, in fact, Wright was our most valuable hitter, contributing 8.5 wins over a replacement 3B. ="Bret Sabermetric"]Somewhere James makes this comment: If a team loses its big star for two weeks, that team tends to get by surprisingly well while he's out. Yes, that's true. The point, however, isn't that the big star isn't valuable, but that his value is spread out across the entire season. The difference on a day-to-day basis isn't meaningful--it has to be taken in the context of the entire season. I mean, Wright created 0.053 wins per game last year. Matsui created 0.016. Wright would have to sit out for 27 games in order for us to "lose" one single game because Matsui's playing instead of him. Now, over the course of the season, that difference spreads out to 6 full wins, but on a game-by-game or even week-by-week basis, the impact is negligible.

old original jb Apr 06 2006 02:25 PM

And I'm saying that the predictive value of the metrics can depend in complex ways on how far from average the rest of the team is. Especially true for defensive (see my argument above), but could also be true for offense. If I am not mistaken, win shares are based on the impact of what a player contributes calculated using data from all teams and all games. But the impact of a hit, a run, a catch, an error etc. may be very different depending upon the expected performance of the other players. For offense the relationships may be even more convoluted than on defense and may in fact have a different relationship to overall team skills-- that is weaker effects when teams are at either extreme. If an offense is very weak, I think that the loss of the star hitter will mainly result in losing by slightly bigger margins than usual, but may not be a big factor since he will have few opportunities to drive in or be driven in by his teammates. If an offense is great, than one player going down may also not have a big impact because others will pick up the slack. It will be a bigger impact than on the bad offensive team, though, because on a good offensive team the players increase one another's impact. I would predict that the biggest impact for loss of a slugger would be on teams with average to somewhat above average offense where that player provides a large chunk of the production, but the other players are on base or driving him in enough to give a multiplier effect to his individual production. If I ever find the time to find a way to prove this with math or with real raw data and statistics, I will do so. If someone can draft a grant proposal for this work, I'll send it to NIH and see if they'll fund me to do it as my day job

Frayed Knot Apr 06 2006 02:52 PM

]I can look at the 2005 WARP3 scores, and verify that, in fact, Wright was our most valuable hitter well, by one method of measurement is does. But let's not go mistaking whatever metric you want to use for the scientifically-stamped final word.

Rotblatt Apr 06 2006 03:01 PM

Frayed Knot wrote: well, by one method of measurement is does. But let's not go mistaking whatever metric you want to use for the scientifically-stamped final word. Yeah yeah yeah. I'm willing to be that most metrics (including the ubiquitous CPF SPoY) would show David to be our most valuable player last year, but your point is well taken. Metrics are to be used in moderation and taken with a grain of salt.

Bret Sabermetric Apr 06 2006 03:57 PM

Rotblatt wrote: [The point, however, isn't that the big star isn't valuable, but that his value is spread out across the entire season. The difference on a day-to-day basis isn't meaningful--it has to be taken in the context of the entire season. I mean, Wright created 0.053 wins per game last year. Matsui created 0.016. Wright would have to sit out for 27 games in order for us to "lose" one single game because Matsui's playing instead of him. Now, over the course of the season, that difference spreads out to 6 full wins, but on a game-by-game or even week-by-week basis, the impact is negligible. Say the difference between playing Matsui 162 games at 3b and playing wright 162 games at 3b is 6 games (M=5-11, W=11-5). If Wright were to cripple himself today, and Matsui were to take his place on the roster for the rest of the season, Mets fans would go absolutely nuts. They'd say that we just lost the season. If you suggested that difference was the difference between going 85-77 and going 79-83, they'd say you were nuts ,the difference is more like 90-72 and 70-92 because they tend to minimize the contributions of a Matsui and exaggerate the contributions of a Wright.(and oversstate the strength of the team generally). That makes it very hard to move a star player, because if you did move a star, and if contrary to expectations the move failed, you will get vilified. The risk of that worst-case scenario is much greater than the possible gains of moving the star, at the least from a GM's perspective. In reality, there is usually immediate value coming back to the other way, which, coupled with the surprisingly low impact of just losing the star player even without compensation, makes each such deal much safer than you'd think. For example, I don't think the Mets got anything like real value in the Cameron/Nady deal. I hated it. But I'll concede that Nady is no worse than a 5-7 player this season, and that Cameron is no better than 7-5, and I'll also concede that maybe I'm wrong in my evaluation of these players. It's a big deal in the context of the Mets making several other foolish deals, and the total is maybe 10 or 15 lost games, but the Nady/Cameron deal will probably cost the Mets under 2 net games even if the deal shakes out the way I think it will. i don;t recommend making stupid trades on General Principles, but I think each one's downside is limited.