Clay Davenport is an author of Baseball Prospectus.
Clay Davenport: Hi, I'm here for an hour. Let's hit it.
Will Carroll (Indy): A chat during Survivor? Are you mad? ;)
Clay Davenport: Welcome to our pre-Halloween chat here on Baseball Prospectus. Tonight I'm coming out of hiding to take a turn in the chair, and I'll answer whatever questions you can come up with and I can get to. Keep in mind that I'm not as
plugged into front offices as some of my colleagues here (he says with extreme
understatement); I'm a proud outsider, so I won't have a lot to say about
Texas' winter strategy.
John Pontoon (Chicago): Well, Dick Matsui didn't translate as accurately as my roto team would've liked. Care to excusify and/or put in an early whammy on Kaz Matsui?
Clay Davenport: A: No, he certainly did not do as well as I expected (dragging one of my roto
teams down the tubes as well), and it sent me back to the drawing board with respect to Japanese translations, with Extra Special Attention paid to power. The result is a revised system specifically meant to deal with Japan, and not treating it like every other league in the States. If I had had these
revisions in the spring, my forecast would have been more like .290/.375/.479
(22 HR) instead of the .290/.421/.567 (32 HR) we actually forecast - since his
actual line was .287/.353/.435 (16 HR), that cuts more than half the error
away. I get comparable improvements on most players. Kazuo's translation in
the old system looked like Derek Jeter; in the new it's more like Orlando
I'd talk about it some more, but it would spoil a perfectly good 2004 book
Paul Covert (Lynnwood, WA): So, when will the 2003 Prospectus Fielding Runs be available (or have I missed them somewhere)?
Clay Davenport: I'm actually running figures for the chapter authors today, and I'll
re-run all the Player Cards once I get book-related stats squared away.
Definitely before Thanksgiving; I actually wanted to have them out already, but that was yet another wish that didn't come true. Like my Powerball numbers.
Pearse McDade (Washington, DC): Clay,
You're the statistics maven for BP and father of the Davenport translations. What books/articles would you recommend to the statistics illiterate among us to more fully understand your approach and performance analysis in general???
Clay Davenport: I'm not sure I have a good answer for you, Pearse. The BP group has talked uite a bit about creating a 'Stats 101' course for us, and I was planning on
cranking out a series of articles related to the player cards when I get those out on the site. I'm thinking I can literally write a book's worth of articles just on that.
Dr. C (Mobile): How accurate have you found your translations to be and what other factors keep someone from living up to the level you expect? I'd expect things like pressure, stress, and job security all factor in, but would be tough to deal with on a pure-stat basis.
Clay Davenport: I actually ran a test of this recently, when I first got the 2003 data, and
its a good thing you sent in this question early, because it took me a half hour to find the right notebook. Roughly 40% of players with significant
major and minor league playing time had a major league eqa within 10 points of
what their minor league numbers indicated, 70% were with 20 points, and 90% were within 30. 'Normal' season-to-season variability is 20-25 points.
Paul Covert (Lynnwood, WA): ...And so I take it, then, that the FR calculations are labor-intensive enough that they're not going to become one of BP's Daily Statistical Updates in the foreseeable future?
Clay Davenport: They are labor-intensive, primarily because they have to integrate everything - so I need all the stats. It is doable, though, and I did experiment with an automated routine to handle it this year, so I expect we will have it next season.
BMM (Midwest): What's the best use of the BP Player Cards to use in Scoresheet's winter game that uses past seasons? WARP1 for both hitters and pitchers? BRARP or BRAR + FRAR for hitters and PRAR for pitchers? Thanks.
Clay Davenport: I don't know Scoresheet well enough to answer that. For scoresheet, roto,
whatever, the best answer is always going to be to use whatever model the game was designed around. There are certain assumptions in the Player Cards that probably don't hold up under Scoresheet play. A problem with WARP1 (for you) is that the fielding rating plays a bigger part in the overall rating as you go farther back in time; a problem with WARP3 is that league quality issues are brought in. All the PRAR values have feedbacks from the estimated fielding ability of
the team, and I don't know if Scoresheet cares about that.
Thomas Prowell (Seattle): How come Davey Johnson's name hasn't come up at all for the vacant managerial position in Boston? You'd think a bright guy like Theo Epstein would recognize Johnson's value to a team.
Clay Davenport: The problem is that, while Davey Johnson is a smart guy, he knows it. This creates a problem with working with other people, especially those who are not as smart as he. He's good, but everyone figures that there are people as good or nearly so who are a lot easier to deal with.
But hey, he's a champion right now (filled in as manager for the Netherlands in their European Cup title run this summer).
Manny Ramirez (?): Will I stay or will I go?
Clay Davenport: I have no idea; whatever forecast I make will end up being the wrong one, count on it (Schrodinger's cat has nothing on me; as a meteorologist, I make a profession of publicly blowing forecasts). No player of Manny's caliber has ever been placed on waivers like this, to the best of my knowledge, so there's no precedent to follow. My gut says he stays.
jschmeagol (carlisle, pa): In which ways does you creation, RARP, differ from VORP (Keith Woolner's?). Which one is a more accurate measure of a player's offensive value?
Clay Davenport: Actually, I'd say that neither one does, because I don't really think offensive value measurements should take position into account at all;
positional differences should be accounted for from the defensive statistics, because nobody is playing any position while they're up.
That's a rather extreme position, I know. Positional adjustments based on
average hitting at a position is a convenient shorthand (and in a managerially efficient world, you would expect that the average hitting difference between positions would, in fact, equal the average fielding difference) and is a lot easier to measure. VORP uses different replacement levels for each position, while RARP uses a fixed percentage of league average at the position; VORP essentially uses runs created, RARP uses equivalent runs.
Benjamin (University of Maryland, Baltimore County): Bill James claims that Drysdale isn't HOF-worthy, as he didn't pitch well in must-win games. Should a similar judgement be levied upon teams that often fail in playoff elimination games, or other "big" games? Thanks.
Clay Davenport: Personally, I think that's hooey. Certainly, a player/team that fails in a key situation should be marked down somewhat, but to take off so much that a Drysdale-level performance isn't Hall-worthy? Please. You can't fail in a big game unless you were good enough to get there in the first place, and I'm one who puts more weight on the 35 games he worked to get there than the couple of post-season games he supposedly choked.
Dixieflatline (Chicago): Clay,
Thoughts on the Brooks Kieschnick experiement in Milwaukee? Do you think other teams might try this approach? Also, would it not be easier to evaluate his impact if he was listed under DH/other as that was the role he was trying to fill?
Clay Davenport: Yes, I think so. There were a couple of players who were trying this in the minors this year - Ryan Minor springs to mind - and there's plenty of guys who are drafted with either/or potential (like VanBenschoten and Markakis) who can test that water if their regular career starts to sour.
Yes, he probably should be listed as DH/other. Keep in mind, though, that it is an automated routine. It labels a player wherever he appeared the most, and with 69 games total, 42 at pitcher, he's going to be listed as a pitcher. However, his batting stats are weighted by position, so his baseline is against 42 games worth of pitchers and 17 games worth of DH/other, which is going to be substantially higher than what pure pitchers are being compared against.
CaptanRon (Washington DC): Hi Clay,
Haven't seen you around Fanatics lately but I'm sure you're still dominating as HRBAKER on the NTN board. Please settle a bet, now that Gary Carter has made the HOF, who is the single most deserving player eligible (who retired in the past 40 years) but not yet in the HOF (not counting players eligible for the first time this year)? We've decided you shall be the final arbiter on an intense debate (and I'm betting you'll say a 3rd baseman who's been in the news lately, not that I want to sway you're decision or anything). Thanks.
Clay Davenport: Got married and moved to the suburbs of Maryland, which puts the downtown bars a lot farther away. And less frequent. Anyway, I would say the top player as you've defined it is Ron Santo, although it is not a slam dunk; I think Bert Blyleven is a pretty close, with Bobby Grich and Dick Allen back a couple of steps.
John M. Perkins (Macon, GA): EQA's for Southeastern League (thank you) but not for the low rookie Arizona and Gulf Coast Leagues? Is there a stat based reason?
Clay Davenport: No, there's no reason, other than that I dismissed complex baseball when I started doing DTs - keep in mind I had to hand-enter all the stats back then, and those seemed like a pretty good place to draw the line of what I was willing to do - and haven't had a really good reason to revisit that decision. The adjustments are awfully low - but that's true of the lowest independent leagues as well. They're a short season - but not much shorter than the Pioneer or Appy. Or the AFL, which I've done stats for. I'll work on getting them added next summer.
Petrus (Torino, Italy): Who's the winner in a Beltran for Soriano trade?
Clay Davenport: They both could be. It isn't obvious that one is a significantly better player than the other, even if Beltran is more sabermetrically correct in his skill set (and din't just spend three weeks looking like an overmatched rookie). Advantage goes to whoever pays the least, and that will probably be Soriano.
Greg Wilson (Boulder, CO): I've read that the average replacement player fields at the MLB average level. If that's true, why do you have Fielding Runs Above Replacement and Fielding Runs Above Average? Thanks!
Clay Davenport: Is that true? A replacement level player is not a fixed entity; it is in many ways a theoretical construct, and the theorist can base it on whatever assumptions he wants, although Keith W.'s done some really good work at injecting some reality into those theoretical constructs. A replacement level player fields at the major league average only because somebody defined it that way.
To answer your question, I start with a different assumption - an assumption that the ultimate replacement level team was the Cleveland Spiders of 1899, a combination of craptastic hitting, pitching, and fielding. That puts my "replacement level" player at a .130 win pct., significantly below the "freely available" threshold (which typically involes a .300-.350 win pct), but still above the "no contribution whatsoever" of Win Shares. Echoing what I said earlier about making positional adjustments with defense, I use the same replacement level baseline (a .230 EQA) for all hitters - yes, even pitchers. Positional adjustments come from the difference between an average and replacement level fielder - about 30 runs for catchers and shortstops, about 12 for first basemen. Ideally, the worst regular fielder in the league should be around 0 in FRAR and facing a position switch.
Jeff Biddle (Chicago): Do you foresee more indy league players making it into the high minors and majors? Who are some players out there who you know deserve a good shot?
Clay Davenport: Absolutely. The Indy leagues are sort of restoring the minor leagues to the size they were prior to the 70s, and there is no doubt in my mind that the draft is inefficient enough to miss players - it would have missed Mike Piazza if he hadn't been a friend of the family. That's especially true for players who have the drive and determination not to quit when they go undrafted, and who then set out to prove to the majors what a mistake they made in passing him over.
Now, one problem I have with evaluating players from the indy leagues is that I have a harder time getting bio data, so I don't always know how old they are. I don't have a top indy player list ready to go, but I'll see if I can get something for you after the chat.
Greg Wilson (Boulder, CO): Will my Baseball Prospectus '96, with its "detachable Cardinals chapter for easy traveling!", one day become a collector's item on Antique Roadshow?
Clay Davenport: I hope so. I've got two of them, myself; the one I've had on the shelf, and the one that got lost in the bottom of a box of notebooks that I didn't find until I moved.
R.C. Cook (Dallas TX): So, what will it take to get a Dallas Pizza Feed?
Clay Davenport: I might prefer some kind of BP Chili Cookoff feed if I found myself in Texas, but let's see. The essential thing would be BP people in Dallas, and I'm scratching my head trying to think of anybody on BP that's anywhere near Texas. It seems everybody who used to live in the middle of the country has moved to one of the edges. Maybe we should try to set up a traveling caravan next year, touring ballparks like in the Ma$ter Card commercial...
Clay Davenport: That was a fast hour. Thanks for all the questions, including the ones I didn't get to answer, and I guess I have to go to the videotape to find out who's still on the island.