Baseball fans used to spend February and March debating the merits of team predictions published in various national magazines. Coming up with a bottom line on entire teams seems pedestrian when compared to the modern phenomenon of precisely projecting how each player will perform.
There's going to be a lot of numbers in this column as we examine how the forecasters expect the Mets lineup to produce in 2007. But don't let the stats blind or bias you. Everyone who is serious about projecting players knows it's a very inexact science. In their best years, the various systems struggle to project even one-fifth of the players to within 10 percent accuracy. A very nice summary of various systems and their pitfalls can be found in the new book, A Mathematician at the Ballpark, by retired University of Oregon Ph.D. Ken Ross.
It's important to note that players change suddenly and dramatically and then back again. Baseball players are a lot like the quanta of quantum mechanics, whose next place in the atom can never be known but only guessed at probabilistically. Where a player was last year (or month or week) doesn't necessarily influence where he's going. All the data in the world allows only for slightly more educated guesses that will never be more than that. And, really, who would have it any other way? This, after all, is what makes the game so much fun to watch and follow.
As I did in my piece on the top three Mets starters, I'm using two dramatically different projection systems. The father of the sabermetric movement, Bill James, is represented as published in The Bill James Handbook. And Nate Silver of Baseball Prospectus is also cited via his PECOTA system, which finds each player's most similar players from the past and then projects him to perform as those past players actually did going forward.
Finally, I'll note other stats that I've come across that should be considered when trying to assess the likelihood of how each of the starting positional players and top reserves will measure up in 2007.
Leading off, PECOTA has Jose Reyes at .290/.335/.445 (batting average, on base percentage, slugging percentage) with 14 homers and 61 steals. James has him pegged at .289/.333/.448 with 13 homers and 58 steals.
I suspect that everyone thinks that the sky is the limit for Reyes. But it's very tough to project that because forecasters know two things: (1) projections are generally too generous by an average of 20 percent and (2) as noted by Ron Shandler of BaseballHQ.com, it's easier to stink than to be great. While there are just a couple of reasons for getting much better, there are virtually countless reasons for any player to be much worse.
I have no idea what Reyes is going to do. He could hit .330 or hit 25 homers or steal 80 bases. There's some chance he can do all those things. I'd look at the above projections as a floor. Reyes scored 40 percent of times on base, second only to Hanley Ramirez (41 percent) among MLB regulars. I know this has a lot to do with Carlos Beltran, Carlos Delgado and David Wright; but it's still impressive and makes his relatively low on-base percentage less important.
There's a big split on Paul Lo Duca in terms of at bats. Most catchers at Lo Duca's age and with his prior mileage start to fade physically, so PECOTA clocks him in with just 403 ABs. James looks only at the player and expects 516 at bats. I'd go with James here and agree that Lo Duca's slugging percentage will be within 100 points of his batting average. So, he better hit around .300 again or he's a problem.
Last year, I wrote that Carlos Beltran's power would bounce back significantly due to some bad luck in 2005 on turning flyballs into homers. But I never expected 41 Beltran Bombs despite missing almost a month of action. His homer rate on flyballs actually set a new career high. PECOTA and James both expect a decline to career levels (30 and 32 homers, respectively). I agree that Beltran's rate of homers on 22 percent of flyballs will decline by about four percentage points. Fun facts: Beltran was third best in pitches per plate appearance (4.2), sixth best at hitting changeups (1.082 on-base plus slugging percentage), second best with 26 road homers, third from the bottom in swinging at first pitches (12.9 percent) and the most productive road hitter in baseball (10.9 runs per 27 outs for a team of Beltrans on the road).
Carlos Delgado is given just 439 at bats by PECOTA. James gives him 582. PECOTA gives him 27 homers, James 37. Delgado was the 10th best hitter in the NL on sliders (.845 OPS) last year. I don't know if those splits on various pitches correlate from year-to-year (probably not very well), but I love 'em.
David Wright is given 29 homers by both projections systems. That is viewed as miserly by many fantasy players. But I think it may well be too generous unless his second-half power outage and homer rate of 12.9 percent of flyballs were flukes. Both projection systems also forecast a plus-.300 average and slugging percentage in the mid-.500s.
Moises Alou is a fascinating guy. PECOTA hates all older players. But players have started playing significantly longer and better relatively recently, so that may be a problem with the system. James gives Alou more than 200 more at bats (503) and 23 homers to PECOTA's 15. But Alou hit homers on 18.5 percent of flyballs last year (which would have placed him just behind Delgado and Beltran on the Mets). This is despite playing in San Francisco's AT&T Park, the toughest homer park for righty hitters in baseball by a million miles, at least in '06. Expect Alou to hit for more power than expected as long as he's healthy.
PECOTA again slams the veteran player, this time Shawn Green, while James is relatively optimistic. Note both systems project similar performance from Lastings Milledge if he's given playing time at Green's expense. PECOTA gives Milledge a plus-.800 OPS, James just under .800. Remember, Milledge, who's added 15 pounds of muscle over the winter, was a very young player in a pitching environment at AAA last year and performed much better there than league average.
Both systems forecast significant declines for Jose Valentin, the Mets' most extreme flyball hitter but one only very slightly above average in converting flyballs into homers. PECOTA thinks that Ruben Gotay (.708 projected OPS), who was once highly touted as a Royal rookie, is the best option should Valentin falter.
Similar declines are being forecast for Endy Chavez, but newcomer Ben Johnson is viewed as a capable outfield reserve by PECOTA (.775 OPS).
What does it all mean? PECOTA expects a significant decline in offensive productivity from the Mets in 2007 (relative to '06). James expects New York to again be among the handful of most productive offenses in baseball despite being handicapped by a pitcher-friendly park. Come October, I think James will get the fortune-telling prize.