Nate Silver is an author of Baseball Prospectus.
Nate Silver: Hi everyone, let's get chatting.
Adam J. Morris (Houston, Texas): Texas Rangers -- any chance they hang around in the A.L. West race?
And why is there no Kenny Rogers PECOTA card?
Nate Silver: I like the Rangers quite a bit. The lineup can mash from top to bottom, and if a few of the young arms come through, it's a team that could win 88 or 90 games, which just might be enough in the AL West this year. I think it's easy to forget that the Rangers are a team for which just about everything that could have gone wrong had gone wrong for much of the last several seasons; they're an 80-win team that might be able to play at a 90-win level.
Pierre (Torino): Howdy Nate, how you doin' today ?
Nate Silver: Sorry about that, gang. My laptop is in the shop, so I'm on my old Pentium I today. Just like Bud Selig's hairpiece, this thing was state-of-the-art no fewer than eight years ago.
DK (NYC): Bruce Chen just got traded to the O's for a PBTNL. The O's are the 10th organization he's played for. PECOTA still thinks he has some value, despite his prodigious ability to serve up taters. Is he AAA filler at this point, or is there still a chance he'll succeed?
Nate Silver: Home run rate is huge when evaluating young pitchers. Lots of guys can put together reasonably good K/BB numbers in the minor leagues simply by challenging hitters; it seems that HR rate is a truer test of how likely that strategy is to hold up in the big leagues. I'd be surprised if Chen puts together much of a career.
men w/o hats (safety city): On a scale of 1-10 -- '10' being a Brady Anderson-esque fluke and '1' being as expected as Bonds walking at least 100 times -- where would you place Craig Wilson's season so far?
What odds do you give him of finishing around .300/.370/.500?
Nate Silver: Somewhere between a 3 and 3.5, like maybe 3.14159 or something.
Wilson's always been a good hitter, and it's easier to be a good hitter when you're playing every day. Something like 300/370/500 is certainly at the upside of his forecast range, but it's well within it.
Randy Brown (Ann Arbor, MI): How many PA's does Wily Mo Pena have to keep posting his current .296/.387/.519 before we all get in line to apologize to the all-knowing PECOTA?
Nate Silver: Go, Wily, go.
I had the chance to see Pena play when the Reds came through Chicago not long ago. The guy *looks* like a hitter, which is to say, he's big, and strong, and the ball goes a long way when he makes contact with it. His plate discipline looks to be a bit better too this year, and he's a talented enough hitter that even marginal gains in that department could bring him to the fringe of usefulness. And when a 21-year-old is useful in the big leagues ... there's a good chance that he's going to have a career.
Colon City Bomber (Colon City, PA): Who is the geometric mean between Ben Oglivie and Barry Bonds? Or instead, Ben Oglivie was the geometric mean between Barry Bonds and whom?
Nate Silver: Greg Luzinski maybe? Cecil Cooper?
Alex Ciepley (Brooklyn, NY): My fantasy teams... so many Expos hitters... so slowly dying...
Nate Silver: Jonah Keri has been no fun this year. I warned in my last chat that the Expo hitters might be a little bit overrated by PECOTA because of the ballpark effect that we'd applied, but their collective slump looks far more serious than that. I don't know anything about the inner workings of that organization apart from what has been reported publicly, but I'd feel better about forecasting a rebound if I were confident that the team was providing their players with the best in training, instruction, coaching etc. As it stands now, the Expos are being run as a third world franchise, and I think we're seeing some of the downside of that.
Matt (Boston): How should Theo prioritize the following five free agents?: Pedro, Nomar, Ortiz, Varitek, Lowe.
Nate Silver: Nomar, Pedro, Varitek, Lowe, Ortiz.
If you'd asked me at the start of the season, I would have placed Pedro first, ignoring the usual risks inherent in picking an elite pitcher before an elite hitter. But, while the guy is clearly a wonderful pitcher who is capable of making adjustments even while at less than 100 percent, I'm not sure that we're ever going to see the 2001-2002 version of Pedro again. Considering that he's likely to demand his market rate and then some, there's a universe in which it might be in the Sox' best interest to part ways with him.
DK (NYC): How much weight do you give to PECOTA when you're doing your own evaluation of prospects? There were a bunch of selections on the BP top 50 prospect list who PECOTAed terribly (like Dustin McGowan), and guys who PECOTAed very will (like John Maine) were left off the list. How much say did you have in that list?
Nate Silver: The top 50 list was truly a consensus effort, and PECOTA was an important factor in the placement of folks like Jeremy Reed (who is tearing it up at Charlotte) and Justin Morneau. But it wasn't the only piece of information that we looked at, nor should it be. If you look at a guy like Joe Mauer ... PECOTA is of limited use there, simply because he's such a unique player that there's not much of anybody to compare him to.
Terry (San Francisco): Do you see any chance that Bonds will finish the year with an obp of .600 or above? His PECOTA card 90th percentile BB was 161, with an EqOBP at .549. At the current rate, though, he'd go off that chart.
I've seen a handful of games so far this season, and it's a shame we don't get to watch that swing more often. Does it make any sense for the opposition to keep up this walk rate rather than pitching to him?
Nate Silver: I was truly amazed by some of the numbers in Joe Sheehan's article a couple of weeks back, that Bonds had more home runs than swinging strikes for example. Barry wasn't all that far from a .600 OBP in 2002, and considering how little incentive there is to pitch to him at the moment, it wouldn't shock me one bit if the guy maintained that level.
This doesn't get said very often, but I really thik that the Giants need to consider acquiring a big hitter at some point this summer to place behind Bonds in the lineup. Not only do they need the offense, but it would make Bonds much more effective. If somebody like Magglio Ordonez became available, he'd be perfect.
Southcoast (Texas): Is Brandon Larson going to continue hitting successfully? He's looking pretty comfortable at the plate. If you had to pick between 2 of the following 3 players who would you keep in a NL only 4x4 keeper league, all players with approx the same salary and all available as keepers for next year. [B.Larson, Holliday, Jason Bay]. I can only keep 2 of the 3 players listed above.
Nate Silver: Larson's had about as rotten a run of luck as you could imagine, and it's good to see that he's hitting again. I think he'll be one of the better third basemen in the league for a couple of years. There's no question that I'd protect Larson and Bay ahead of Holliday. That's Holliday, not Halladay, right?
DK (NYC): Thanks for PECOTA, I've found it to be a very useful tool.
One question: Why is there no PECOTA card for Ponson? Is he too big to fit into your projection machine?
Nate Silver: Our apologies on some of the PECOTAs not being updated. They'll be fixed as soon as the PECOTAputer is back from the shop.
Yoopi Doopi (Itinerant entertainment worker): Nate, love your work. A couple questions...
1) A fundamental clarification - I know that PECOTA uses just the past 3 years in calculating a player's "baseline", but how far back (in the minors and/or majors) does it look in comparing the player to potential historical matches?
2) Did your 2004 PECOTA forecasts for Expos players use last year's park effects for Hiram Bithorn, or some kind of theoretically guestimated version of them that accounts for the profound changes the stadium underwent this offseason (fences moved back 10-30 feet, but fast new turf)?
Nate Silver: PECOTA looks at all major league players since 1946, and all minor league hitters since 1998. We'd like to expand the minor league database in the future; it would be useful to see how well B.J. Upton compares to Derek Jeter, for example.
The Expos projections assume the same Olympic/Bithorn hybrid as the team played in last year. They do not assume that the team would start to hit like the 1912 St. Louis Browns. Or the 1892 Cleveland Spiders.
Easton (Clayton CA): Hey Nate
I think you'll agree with me that Hank Blalock of the Rangers has a tremendous future, including being the all-time best of the left-handed throwing 3rd basemen. See pages 1 and 517 of the "Prospectus".
Nate Silver: Blalock is a wonderful player, but I'm hesistant to classify him as a potential inner-circle HOF guy before he learns how to hit left-handed pitching. Even with his hot start this year, he's managing just a .366 SLG against them. Can he be the next Eric Chavez, with maybe a little bit more home run power and a little less defense? Sure, but I don't know that he's the next Mike Schmidt.
ncooke (Seattle): What happened to Edgardo Alfonzo? Peaking at 27 is fine, but then he fell off a cliff at 29. Shouldn't he be better than this, and did PECOTA see this coming?
Nate Silver: The thing I try an emphasize is that every player ages differently; the typical aging curve that you see, which presumes that a player peaks at 27 and then slowly declines thereafter, is really just a cheap shorthand that's made up of a lot of divergent experiences of a lot of different types of hitters (see my BP Basics article for more on this). It would surprise me, frankly, if Alfonzo ever performs at an elite level again.
Mike (Alexandria, VA): Are the Reds shopping Sean Casey and/or Ken Griffey Jr? What could the Reds reasonably expect in return for either of those guys?
Nate Silver: I don't know whether they're shopping those guys because, for the time being at least, the Reds look like they might have a real team on their hands. Their case is somewhat similar to the Rangers, I think, in that they had a ton of bad breaks last season that probably diminshed expectations to an unrealistically low level.
Besides that, Casey has an albatross of a contract, and I can't see anyone taking him on. Griffey ... maybe he could end up in the Bronx, but I don't know.
leicester (Burlington, VT): What is the underlying distribution used by the PECOTA system to produce a player's distribution? Is it simply the empirical distribution based on the comparables, or something like a Poisson distribution for counting stats with the parameter estimated by the comparables?
Nate Silver: It's empirical. A player's PECOTA forecast range is determined by the performance of his comparables, nothing more and nothing less.
Rich M. (Chicago): Nate, Enjoyed the CIRBI article. However, Is the data available to break down the 3rd base situations into 2 outs (where only a hit will do) and 0 or 1 outs (where "productive outs" will do). Just wondering if among your leaderboard of players efficient in driving in the runner from 3rd there are those who had an abnormally large amount of those chances with less than 2 outs.
Nate Silver: I didn't check that, but you make a valid point -- a hitter who comes up a lot with 2 outs is going to be at a disadvantage when it comes to driving in runners from third.
cloakscott (Champaign, IL): Where'd the adjusted standings go? And all your contact information on the site is busted.
Nate Silver: Will Carroll has already offered to put Dave Pease up in a day spa for all the hard work he's done getting the new site up and running. We took the Liz Taylor route by deciding to upgrade our front-end and our back-end at once, and the long term result is going to be improved performance for our customers. The feedback we've gotten so far on the new site design has been universally positive. But there's no doubt that there are going to be a couple of bumps along the way, and we appreciate your patience.
prophet (Mountain View, CA): How did PECOTA do with last year's playing time estimates? I saw that the rate state predictions went pretty well, but how well does it predict playing time from past results?
Nate Silver: Last question for today. Thanks for hanging out, guys.
It's hard to evalute how well PECOTA did in evaluting playing time, because all the other forecasting systems predicted playing time exogenously, which is to say, they filled in the playing time numbers by hand. Doing it by hand is probably going to produce a more accurate result than doing it by computer, simply because you're able to take into account so many more factors that way. PECOTA didn't know that Dusty Baker hated Mark Bellhorn, for example; there's no variable for that. That said, PECOTA is pretty good at picking up on things like the propensity of injured playres to stay injured, and the propensity of older players to become injured even if they've been healthy in the past, and I think there's some value in considering the playing time numbers that it spits out, along with a host of other considerations.
Nate Silver: Thanks for all the great questions.