Master Index of Archived Threads
It may not exist, but they get an award for it anyway
| RealityChuck Nov 01 2006 09:12 AM |
|
|
[url=http://mlb.mlb.com/NASApp/mlb/news/article.jsp?ymd=20061031&content_id=1728319&vkey=news_mlb&fext=.jsp&c_id=mlb]Pepsi All-Clutch Team[/url]
|
| metirish Nov 01 2006 09:14 AM |
|
|
Shit,those are some nice numbers...is he available?
|
| Edgy DC Nov 01 2006 09:15 AM |
|
It's fun how they come up with a different stat to underscore the clutchness of each player.
|
| metirish Nov 01 2006 09:18 AM |
|
Yeah, and does clutch only exist in late innings?
|
| Frayed Knot Nov 01 2006 09:38 AM |
|
|
If only someone here had led a campaign to get him earlier!! A show (read: sponsor tie-in) on this "clutch" team was run on ESPN last night - I flicked by a couple of times. Like most discussions on clutch they admitted they couldn't nail down any way of defining it. They gave Bill James a sound bite or two, talked about "close & late", and "BA w/RiSP" numbers, then in the end pretty much wound up pulling a kind of Potter Stewart on the whole deal: I can't define it but I know it when I see it'. Problem was (at least as far as I could tell) they used some of those supposed clutch stats to make their choices w/o any consideration as to what the players' normal stats were. IOW, it turns out that the guys with the real good "clutch" stats were ... the ones with real good stats to start with! Wow, whoda thunk that?!? And how'd those clutch abilities help young Mr. Wright out last week?
|
| RealityChuck Nov 01 2006 09:38 AM |
|
|
Certainly not*. But there is no way to define "clutch" that will include all instances that you consider a clutch performance and will also exclude all instances you don't consider clutch. We need to parallel Damon Knight's definition of science fiction and say, "A clutch performance is what I am pointing to when I say, 'That is a clutch performance.'" That would drive the statheads nuts. *What were probably the most clutch baskets in New York Knick history? Willis Reed's two baskets in the 1970 finals. They made the score an insurmountable 4-0, Knicks.
|
| MFS62 Nov 01 2006 09:46 AM |
|
I think Durham is a free agent.
|
| Edgy DC Nov 01 2006 10:03 AM |
||
|
How'd they help Beltran with the whole damn gonfalon depending on one swing?
I don't know if I'm a stathead, but that would drive me nuts.
|
| Vic Sage Nov 01 2006 10:40 AM |
|
come on Edgy, you know who a stathead is.
|
| RealityChuck Nov 01 2006 10:46 AM |
|
|
It's also people who refuse to believe that baseball players are human beings and that their statistics reflect personality, preferences, changes in they play, and mistakes.
|
| Edgy DC Nov 01 2006 10:50 AM |
|
If you don't like a criteron I choose, let me know you don't like it and why.
|
| Yancy Street Gang Nov 01 2006 10:55 AM |
|
Hey, when do they announce the Silver Slugger awards? The Mets may have a winner or two this year.
|
| Rockin' Doc Nov 02 2006 07:35 AM |
|
MFS62 -"I think Durham is a free agent."
|
| patona314 Nov 02 2006 08:44 AM |
|
|
Honestly, I don't think the Mets should pursue Durham. He's 35 and could be hit or miss in terms of productivity. I'd like the Mets to concentrate on defense at this position aka Anderson Hernandez and spend money on pitching.
|
| RealityChuck Nov 02 2006 09:32 AM |
||
|
Let me give a clear real-life example: Jose Reyes. A year ago, Baseball Prospectus said Reyes was a bad leadoff hitter because his OBA was too low and castigated Willie for batting him first. They said that, at his age, he was too old to learn to be a patient hitter (with the statistics to prove it), and that Willie's patience was a sign that he wasn't up to the mangerial job. Sports Illustrated even brought up the idea that maybe the Mets need to turn to Anderson Hernandez as their leadoff man. Remember? Then someone came along and said, "If you look at runs scored per times at bat, Reyes's numbers are pretty good." So here's someone picking a new criterion to prove his point. Then Reyes started becoming more selective at the plate and raised his OBA to levels that made statheads happy. But I never saw anyone who claimed Reyes couldn't do it say "I was wrong." Because, after all, to admit their mistake would be to admit that i numbers don't tell the story. It's the same thing with a Baseball Prospectus report on Dan Uggla, who said, essentially, "he's doing much better than his PECOTA (aka "Pulling Every Conclusion Out The Ass") would indicate." That meant the PECOTA was wrong. But you'd never see anyone at Baseball Prospectus ever speak such heresy. Now, let me make this clear: I grew up a baseball statistics freak. (Now I've outgrown the stage, but never mind.) And the various ways of analysis can be very useful in analyzing past performance. If you wonder why the Mets lost to the Cards, their lack of two-out hits with runners in scoring position is a key factor, for instance. You can discover all sorts of useful information about players through it. However, Sabermetrics are no good at explaining anything. Nothing in it explains why Guillermo Mota pitched poorly for the Indians, but pretty well for the Mets. All the numbers could conclude is that he did. But why? Was his bad numbers with Cleveland due to his being over the hill? I'm sure Baseball Prospectus concluded they were. Were his good numbers for the Mets due to a regression to the mean? If they were, why didn't anyone predict it? Or were his bad numbers with Cleveland due to his tipping his pitches and his good numbers with NY due to steroid use? Statistics can't say, or even come to a conclusion about it. Finally, Sabermetrics is piss poor at predicting. PECOTA doesn't even claim to be all that good -- the numbers given are supposed to be in a range. Given a big enough range, one can predict anything accurately (I predict next year Jose Reyes OBP will be between .000 and .500), but why not just give a range? Say that PECOTA predicts a BA of .257-.267? No, they give a single number that is only a range if you read the fine print. That's intellectually dishonest. A coach might predict a player hits around .250, and that's no different from a PECOTA score, but because of the extra significant digit, PECOTA sounds more authoritative. Also, do they ever compare the PECOTA predictions with actual on a systematic basis? I've seen it on an individual basis, but does anyone go through all predictions and find how well their doing? And how do these match with a baseline? Take this one: I predict all players will put up the same numbers in 2007 that they did in 2006. Is PECOTA any better than that, given the same measuring criteria (i.e., same range they use for each criterion)? Is Baseball Prospectus doing this sort of comparision?
|
| MFS62 Nov 02 2006 09:44 AM |
|
|
By a strange coincidence, the answer to one of those running clues in this morning's Daily News crossword puzzle was a quote from someone named Evan Esar. It may be what some are trying to say about "clutch".
Later
|
| Edgy DC Nov 02 2006 10:25 AM Edited 3 time(s), most recently on Nov 02 2006 11:50 AM |
||||||||||||||
|
I'm not sure what claim I'm making that you're talking about. When persons make rational scientific arguments, and they're disubuted, communitites of like self-styled rational thinkers can analyze the dispute, how well or not the initial claimant defends their claim, and consensus can be reached, ideally so broad as to reduce the individual biases of the community members to make bias progressively less and less significant. If, in the middle, someone is left saying, "I don't care about the data, it's how I feel," the rational thinkers recognize that he or she has nothing left to offer the argument. If consensus is not reached, his or her conclusions are labeled controversial, noted as such, used withe caution and rarely, but revisited on occassion when new data is culled that can be applied to them. That's science. Newton puts forth ideas that he feels are so backed up by the data that he has the audacity to call them laws --- not theories, but laws --- but his peers reviewed his work and the scientific community accepts them as such.
No. Please provide references for these articles. This is hightly controversial theorizing and not the consensus of any baseball analytical community. Vic Sage believes plate discipline is a skill like bat speed or foot speed that can't be learned, only maximized to some extent within the context of a player's talent level. I think otherwise. I look at Barry Bonds and Mookie Wilson and others and think it can be learned, though it's harder to learn at the big-league level. We both have some data to back us up, obviously neither one is compelling enough to create a consensus. Which one of us is the stathead? The one you say is? What kind of crap is that? I say Vic's a Martian.
It made me happy, but that's because I'm a Met fan.
Anybody who speaks in terms of "can't" and "won't" is not being responsible. We're working with probability here, and you should use terms like "is likely to" and "is less likely to."
Do they really claim that their projections will be exactly reflected in the upcoming season, or do they rather effectively say 'this is as accurate a projection as we can come up with, but so much is beyond our ken'? It's a projection, not a promise. And isn't saying that he's doing much better than his projection acknowledging that the projection was off? What do you do with such data? If you're scientific, you look for whatever factors the guys who most exceeded (or fell short of) their projection have in common, and work it into the formula. Did you have a more accurate projection, on average, for every player in the league? I'm rather certain, nobody nailed down the outcome to the last bunt single of every occurance in baseball, but the guys who used good science did better than those who didn't.
Great. But you've outgrown that.
No. They can also conclude the likelihood that he might continue to pitch poorly or the likelihood of a rebound. The why isn't as much what they're after, so much as the "what next?"
Please, show me where they concluded this.
They can take data and suggest a likelihood. Which is worth a lot.
Listen, I didn't invent PECOTA. I don't defend PECOTA. And PECOTA does not equal Sabrmetrics. When you're more accurate than PECOTA, you can call it piss-poor. Until then, you're being just as arrogant as any abuser of statistics. Bill James himself --- even as he looks back at some of the snotty iconoclasm he pushed in the eighties --- decries those who read his books and think they're replacing untried assumptions with truth, when they're often replacing one set of assumptions with another set of assumptions, that they accept because it was advanced by a self-styled statistician, but they lack the grounding to analyze themselves, and so rather adopt it on bias.
Feel free to. You seem to have already concluded that their success rate will be piss-poor, but I'd love to see your systemic analysis.
I'll make a reasoned prediction that PECOTA is somewhat more accurate than that. But if you make such predictions and analyze that, you might get a derogatory tag like Stathead, which is what I really object to.
|
| Vic Sage Nov 02 2006 11:41 AM |
|
Mars to Edgy... Mars to Edgy... message coming in....
|
| iramets Nov 02 2006 12:11 PM |
|
Great work and great arguing, Edgy. I really learned a lot from you. Thanks.
|