Baseball Prospectus - Chat with Nate Silver

Chat: Nate Silver

Welcome to Baseball Prospectus' Friday February 10, 2006 1:00 PM ET chat session with Nate Silver.

Nate Silver is an author of Baseball Prospectus.

Nate Silver: Hi, gang. I'm glad that people are enjoying the new PECOTA cards - we're very proud of them. Lots of questions in the queue. I'm going to try and bang out answers to some of the more technical kinds of questions, but hopefully we'll mix in some good baseball talk too.

jak (Santa Cruz): Aren't the 5 year predictions fluff?

Nate Silver: Starting out with a softball question.

The 5-year projections are determined by the exact same method that the 2006 weighted mean projection are, and fundamentally they should be just as reliable.

For *most* players, you will see the range on their EqA/ERA percentiles expand as we get deeper into the future, which is only natural. But I do think people probably overestimate the degree of uncertainty involved in the exercise: "anything can happen in five years!". Well, technically that's true, but for most players there are only a certiain number of career paths that have any sort of finite probability attached to them.

birkem3 (Dayton, OH): Hey Nate - What's up with Adam Dunn being projected defensively as a LF?

Nate Silver: We designate a player's primary position as the one that he's played most frequently over the past couple of years (with 2005 weighted *much* more heavily than 2003 or 2004; most position changes stick). We do consider the secondary positions he played too. Dunn's VORP and so forth, for example, has him as a hybrid 75% LF / 25% 1B, since that's been about his mix in recent seasons.

What we haven't done historically is change a player's projection based on anticipated position changes for next year.

We actually changed this policy for pitchers this year - if a pitcher goes from being a starter to a reliever or vice versa, we'll now run a new projection for him. It's pretty important that we do this for pitchers, since his role (starting/relief) not only affects his innings pitched, but also has a quite profound effect on his ERA and peripheral stats.

Introducing an ability to do the same thing for hitters is on our radar screen, but most of the time it isn't going to make more than a 5 run difference in the player's VORP, etc.

molnar (St. Cloud, MN): taking jak's question a step further, Wily Mo Pena has a 50%+ chance of a breakout season every year for the next five years. These can't be independent, can they? Can someone realistically break out three years in a row?

Nate Silver: Molnar,

That's correct. A 2008 breakout means a breakout versus where WMP was at during 2003-2005, not versus where we think he'll be in 2007.

johnklein (improvement rates): Explain how a pitcher (Ryan Dempster in this case, but the question is more general) can have an improvement line of 38% - 62% - 18% but be forecast with a stat line significantly below his last year's performance. Also (if I can sneak somewhat of a correlary in here) what gives with Brett Myers? His strikeouts seemingly came from nowhere, but his 2006 projections look very strong. Thanks, jk

Nate Silver: The breakout metrics evaluate changes in performance versus a player's last three years (weighted on roughly a 50%-25%-25% basis), not just 2005. So even though we're not expecting fantastic things from Dempster next year, he's still making up a lot of ground versus where he was in 2003 and 2004.

Ian Magnusson (Calgary, AB): Why no PECOTA card for Julio Franco?

Nate Silver: Tell me which comparables I should use for him and we'll run one.

We're able to run a simple projection for Franco (which is what appears in the weighted means spreadsheet) simply by running a baseline forecast and extrapolating out a generic aging curve. But all of the nifty charts and detail that you see on the PECOTA cards depend on having a set of comparables, and Julio don't got none.

molnar (St. Cloud, MN): Defense: the projected defensive runs above or below average seem very muted. For example, Edmonds, last two years, +13 and +21 runs. This year, +1. I'm down with regression to the mean, but this looks like it's saying we just can't project fielding at all. Is this just an unusual side effect of how the projected WARPs are backed out to form the complete stat lines?

Nate Silver: The defensive projections are regressed to the mean a lot. But Edmonds projection is particularly harsh. It's a consequence of his age, as well as the fact that he no longer runs very well. Speed score turns out to have a quite important predictive impact on FRAA, especially for a center fielder.

George (DC): Nate - why does PECOTA hate the White Sox pitchers again? Vazquez is the only guy below a 4 ERA? He's the 5th starter on this team! Last year it got them very wrong as well. What's the bias there? Does it think that US Cellular is just too tough to have any kind of sustained success. Even Buehrle looks average in PECOTA.

Nate Silver: George,

I buy into the projections for the White Sox pitching staff, with the possible exception of Contreras and maybe Bobby Jenks, each of whom have a lot of backstory going on that PECOTA won't be able to pick up on. There's really three different things going on that harms the White Sox pitchers:

1) Most of them performed significantly better in 2005 than in 2003 or 2004.

2) As you mention, it's an uphill battle to put up good ERAs in The Cell.

3) The White Sox got something like +50 runs from their defense last year, all of which finds its way into the ERAs of individual pitchers. We're projecting that their defense this year will be much closer to league average. Part of that is because they punted Aaron Rowand, but part of it is that a lot of guys like Konerko and Podsednik posted defensive numbers that were out of line with their past histories.

ChuckR (Addison, IL): How much of the 'new' information will be in the book as opposed to on the cards online? There are a lot of cards to go through, but are there any Stars and Scrubs charts that is especially interesting that we should be looking at?

Nate Silver: I think the comparison between Jim Thome (lots of green and red, not much yellow), and Aaron Rowand (all yellow) is pretty interesting.

steve S (Davis, CA): Nate, can you explain why for some players PECOTA has very different weighted mean projections and median projections? I noticed this in particular with the White Sox starters. In the case of Freddy Garcia (whose median ERA projection is 3.86), the weighted mean ERA projection (4.32) is even higher than his 25th percentile projection (4.26).

Nate Silver: Sticking with the White Sox theme.

Good eye, Steve. You did catch a little bit of a big with the percentile forecasts for the pitchers. Namely, the percentile forecasts had backed out park effects (you'll see the same problem for a lot of Rockies, for example). My apologies, and we'll get this fixed over the weekend.

Unfortunately, it won't help the White Sox pitchers at all. The weighted mean forecasts are "right", as are the various valuation metrics. It's the percentile numbers that are too favorable.

Aaron (Allston, MA): I love the new PECOTA cards. That said, I can not understand PECOTA's reasoning behind the Jonathan Papelbon projection. He was pretty good statistically speaking but his projection is not in the same world as a lot of his peers....particularly the 5 year forcast where his mean projection is out of baseball. At the moment I'm considering the projection a considerable outlier. Care to defend or disagree with the forcast?

Nate Silver: We'll talk more about Papelbon when we get into the pitchers in the PECOTA Does Prospects series.

He posted great ERAs last year, all the way from Portland to the Fens, but past ERA is not really a very good predictor of future ERA, and his peripheral stats aren't as good. At times I've thought about blanking out the past ERAs on the PECOTA cards. It can be a very deceptive metric without context.

Pap's peripheral numbers are decent, but not "oh, wow!" good. Strikeout rates drop off a *lot* from the upper minors to the majors, and he failed to strike out a batter an inning in a *very* favorable pitchers environment in Portland. His command is a little bit of an issue. His flyball rates are pretty high. He's be 25 next year, and most star pitches are going to have accumulated more than 30 innings of major league playing time by the time they're 25. There are a lot of small issues that add up to a forecast that's pretty out of line with people's perceptions.

mwball75 (Cincinnati): What's the ETA for the PFM?

Nate Silver: Monday. Possibly sooner.

George (DC): Nate In response...weren't the real improvements from Garland and Contreras in 2005? And given Garland's 2005 was his age-26 season, isn't his improvement more likely to stick? Also, Podsednik moved from CF to LF last year, so I'm not surprised that his numbers moved up considerably. I wouldn't count on much of a dropoff in 2006 as he continues to play a position once manned by Carlos Lee (Horsehands, as he was known to Sox fans). Konerko was only about 10 runs better than average, so his defense isn't going to make THAT much of a difference. I'm just surprised by the pessimism for 2006, when PECOTA was so far off in 2005 (Buehrle 4.47, Garcia 4.55, Garland 5.05, Contreras 4.91, El Duque 4.35).

Nate Silver: George,

I've said that I think Contreras' forecast is low.

Garland's improvement in 2005 was the result of a fairly ordinary maturation process - he got better at figuring out which pitches he should throw on which counts, and things like that. He didn't develop a new delivery or a new pitch or something like that, and he still has some trouble missing bats. I think the comparison to Jeff Weaver is instuctive (although he has an advantage over Weaver in the durability department).

HugeShoulderpad (Poughkeepsie, NY): Nate, what's with the pessimistic prediction for Casey Kotchman? Has his shine worn off or has the loss of playing time due to injury caught up to him?

Nate Silver: Sometimes there isn't a sexy answer to this kind of question. We expect some development out of Kotchman, but fundamentally it's very hard to be a first baseman to be valuable when he doesn't hit for plus power. And based both on his PECOTA and his scouting reports, there's little evidence that Kotchman is going to do that.

Jesse (Los Angeles, CA): A lot of crazy comparables on Jose Reyes' card--Garry Templeton, Barry Larkin, Luis Aparicio, a bunch of other greats though with similarity scores below 30. Do us Mets fans have reason to hope after all (now that we've all but given up), or does the low similarity index mean it all should be taken with many grains of salt?

Nate Silver: Jesse,

There were some obvious issues with Reyes' numbers last year, but fundamentally it's quite unusual for a player to be performing at major league caliber at Age 22. In fact, Reyes didn't only do that last year, but also as a 20-year-old in 2003. Provided that he can avoid injury, he should at least be a solid regular for the next 8-10 years, and has some higher upside career paths on top of that.

That said, the comparables may overstate the case somewhat, since a lot of those guys were truly elite defensive shortstops (at least early in their careers), and Reyes hasn't been. PECOTA thinks his defense could get better because he's young and very athletic, and anecdotally you'd hope him to settle down after being jerked back and forth from second base. But he's not going to be Louie Aparicio out there.

azhitnik (Easton, MA): I was looking at the PECOTA card for Nomar and noticed that he was identified as "out of baseball" for 2010. Not that I necessarily disagree with you, but I was just wondering how that determination was made. I haven't looked at enough other cards to see a similar forecast, but I was just curious.

Nate Silver: The "out of baseball" designation simply means that more than 50% of the comparable players didn't play a lick of baseball at that age and dropped out of a player's comp set. What's interesting is when we get this result for a young player - look at Anthony Gwynn or Greg Golson, for example.

Eric (Jamaica Estates): PECOTA is downright bullish on David Wright. Are his counting stat projections grounded on him hitting in a highly favorable spot in the batting order (3rd, 4th)? It sounds like Willie "moneyball" Randolph is going to be batting Wrightey 5th, with (gulp) Reyes and LoDuca in the 1-2 holes.

Nate Silver: We're not trying to be quite as cutesy about projecting batting orders this year, since those can change so frequently over the course of the season. But, certainly the fact that the Mets offense could be pretty schizophrenic from an OBP department harms him.

Ryan (West Hills, CA): The Diamondbacks clearly have the best farm system in the game, right? I realize the pitching is thin. And I realize the Dodgers have incredible depth. But when has a system boasted so many upper-tier bats? Quentin, Jackson, Upton, Drew, Young, and Carlos Gonzales is just sick.

Nate Silver: The Diamondbacks have a great system, but so do the Dodgers, The O.C., and a couple of others that I'm forgetting. I do think that the Quentin/Jackson duo is a little bit overrated. The D-Backs have some very favorable hitters parks in their system, and to reiterate, the bar is set very high for a guy playing first base or corner outfield. Drew's numbers, frankly, weren't all that impressive for a very polished 22-year-old. But still, those guys are Top 50 prospects, and that plus Young and Upton makes for a pretty awesome far system.

Ali Nagib (Chicago): I know that this is more of a "mainstream" media type question, and one that's gotten bandied about a fair amount in the last 3 months, but given that you actually live here, do you think that Chicago could ever be a "White Sox" town again, and what would it take (and how long) for that to happen?

Nate Silver: Ali,

I'm hoping to do an article on this once some of the PECOTA stuff settles down. People tend to see the Cubs' dominance over the White Sox as being immutable, when in fact the teams have traded positions many, many times over the past fifty years or so. Certainly, the underlying momentum favors the White Sox, between the championship, the improvements made to the Cell (I might be in a minority, but I've always enjoyed that facility), and some of the PR hits that the Cubs have taken. But there was a lot of pent up demand for Cubs tickets and it might take a couple of years to manifest itself.

phanatic (philly): Ryan Howard, possible MVP candidate based on PECOTA. Doesn't this feel like one of the more optimistic projections.

Nate Silver: Not especially. He can absolutely murder the baseball, and he should draw a few more walks this time around the league since he's going to be pitched a lot more carefully. Last year, he was one of those guys where PECOTA said he was either going to be a star or flame out completely, and with the progress he's made its now thoroughly in his camp.

Evan (Vancouver, BC): Man Aaron Hill's picture makes him look old. Why is PECOTA so down on him, by the way? It's like it thinks his loss of power switching to the majors was real, but his improved OBP wasn't.

Nate Silver: The thing is that the power and OBP issues aren't independent of one another. So long as he doesn't represent a longball threat, opposing pitchers are going to throw him strikes, and if he can't get the power numbers back up than both his OBP and SLG are going to suffer.

Adam J. Morris (Houston, Texas): I did a double-take when I first saw Hank Blalock's PECOTA projections...weighted mean EQA of .290 or so the next few years, on the heels of a .262 EQA season in 2005? Particularly since he's never had a .290 EQA before? What are your thoughts on this? Any explanation as to why PECOTA expects Blalock to be so much better in 2006?

Nate Silver: The short answer is that I think people are reading way too much into that 2005. For an older player, a big one year decline can often be the sign of a guy who has lost his game, but that's very rarely the case for a young player. PECOTA still sees a guy who posted 285 EqAs as a 22- and 23-year old major leaguer.

Dave (Providence, RI): Peralta! Where's the love???

Nate Silver: There are a couple of questions like this about Jhonny Peralta, and another very similar one about Chase Utley. The answer is the counterpoint to the Hank Blalock issue. Just as young players can have flukishly bad seasons, they can have career years too. That said, Peralta still projects as one of the 10-20 most valuable commodities in baseball.

mikemangiore (St Louis): PECOTA like Albert Pujols. Do you take into account the age thing, or has this been put to rest?

Nate Silver: One of the things that surprised me in looking at the new valuation metrics that PECOTA is producing - though it probably *shouldn't* have surprised me - is just how far ahead Pujols is of any other player in baseball. MORP has him as being a $20-$23M/year player over the course of the next five seasons, and I wouldn't have any problem paying him that kind of money if I was a GM.

I don't take the age rumor very seriously. There are a lot of guys I've heard something backchannel about over the course of the years, things that are sometimes true and sometimes aren't, things that sometimes become public record and sometimes don't. But Pujols has never been one of them, and you'd think that given all the scrutinty given to agegate cases over the past couple of seasons, there has been ample due diligence on the issue.

tcfatone (NYC): What's the story behind Huston Street's absurdly high Collapse/Attrition rates?

Nate Silver: The issue is that the collapse rate stuff is keyed off of ERA, whereas the projections are keyed *mostly* off a pitcher's peripheral statistics. Street is a great pitcher, but his ERAs have been way ahead of his peripherals since he turned professional. On top of that, he only gave up 3 HR last year in spite of having an about average groundball rate, and that's likely to work against him too.
Finally, there might be something to the "second time around the league" thing, as a lot of his comparables backtracked after fidning some success initially.

Scott (Renton, WA): Hey Nate, So, a guy with 30,10,1 stats opens in MP1, and I have A....wait, wrong chat. Can you talk a bit about the new career path graph, which includes the red and yellow lines, to go with the different blue ones for the different percentiles for WARP? What sort of valueable trends do you see in this? What should we be looking for when we glance at that graph? Any interesting projections or outliers?

Nate Silver: Scott,

The main thing I look for in that chart is the really high upside career paths. Delmon Young is a fine example of this - just a 10 or 20% chance of being a true HoF type player can have a profound impact on a player's valuation.

TGisriel (Baltimore): My limited review of PECOA gives me the impression that it is not really impressed with prospects until they show something in the majors. Do you think this comment is valid?

Nate Silver: Ceraintly, in terms of risk/reward metrics, the risk goes *way* down once a player has demonstrated that he can perform at a competent-to-good level in the major leagues. This is especially the case for a pitcher (where we're as much looking for his arm to hold up over 200 innings as anything else).

This is really not a new finding - Bill James has been talking for years about how important reaching the major leagues early can be as a predictor of future success. But PECOTA sees this more clearly now with the introduction of the 'level' adjustment.

rick (boston): Papelbon bad peripherals?!? He walked three guys in 27 AAA innings while striking out 27. His lowest K/9 in the past 2 yrs was 8.59. I just don't understand the hate.

Nate Silver: Ryan Vogelsong has had stretches of 27 innings where he looked like the next Juan Marichal.

The argument I can buy in defense of Papelbon, frankly, is the "stuff" argument - that he's more likely to realize one of his higher percentile forecasts because he has a couple of really outstanding pitches. PECOTA won't be able to account entirely for something like that. But PECOTA is very good at weighing the various factors that go into interpreting a pitcher's past statistics, and the totality of Papelbon's record is way out of line with the public perception of him.

JayhawkBill (Farmville, VA): Nate, any thoughts as to why the translations for hitters' performance with the Pawtucket Red Sox are so modest in MLB equivalents? Could it have anything to do with the small sample size over the past couple of years, as well as a few big drops (Shoppach and Petagine coming to mind)?

Nate Silver: Jayhawk,

Pawtucket usually plays at a park factor of about 97, in the much tougher offensive environment of the two Triple-A leagues. And Portland, Maine is one of the toughest hitting environments in the entire minor leagues.

I think people have gotten pretty good at making mental adjustments for major league park effects, but that understanding lags behind when it comes to the minors, and it's a *bigger* issue for the minors because you have such profound differences between the different leagues too. It's no surprise that the trade-off for Paps and Lester getting "bad" projections is Pedroia getting a very good looking projection.

Sam (lake forest): Great chat.

Nate Silver: Thanks, Sam, and on that note we'll wrap things up.

Nate Silver: Great questions, and look for the PFM and other fantasy materials to come online *very* soon.