June 24, 2010
Present-ing The Future
I don’t remember much about my baseball card-collecting days, but my overtaxed neurons still cling to a few isolated remnants. The way I divided the cards according to All-Star status, believing that an appearance in the Midsummer Classic conferred some indefinable air of greatness upon the likes of John Hudek, or signified that Scott Cooper was bound for Cooperstown. The times I toted around huge binders on family trips, bringing offerings of cardboard to an oracular cousin in Virginia (who, in retrospect, may well have been a false prophet) to be deemed worthy or unworthy of prominent display. And one particularly memorable series of cards, affixed with the tagline “The Future is Now,” composed of 15 players age 25 or under who represented the promise of innumerable productive seasons to come. Those cards rest not 10 feet from me as I write this—no, my mom never threw them out—but it would take me hours of flipping, sorting, and page turning to track them down. Fortunately, through the kindness of the strangers populating the internet, I don’t have to. After a Boolean search or two, I’ve located my quarry.
As it turns out, that group of 15 contained several Hall-worthy names, but it also included a number of players whose best work was already behind them (Steve Avery, Carlos Baerga, Rod Beck, Jason Bere, Aaron Sele) by the time the cards were placed into packs. That a third of Upper Deck’s picks failed (to varying degrees) to pan out, despite the fact that all of them had reached their mid-20s as established major leaguers before being featured, drives home the difficulty of forecasting performance. With that, I’ll conclude my detour into the domain of Josh Wilker—and to think, if I hadn’t wasted my youth in card collection, I could’ve spent my time on Rec.sport.baseball, working toward becoming the youngest founding member of BP.
The little trip down memory lane on which you’ve grudgingly accompanied me was prompted by the announcement earlier this week of this year’s rosters for the Futures Game, the mid-season showcase of minor-league talent that kicks off each All-Star weekend. The game is nothing more than an exhibition, but many members of its talented cast of characters (which includes 17 first-round picks, as well as a number of high-profile international signees) are about to go on to bigger and better things. By my count, almost half of the participants in last year’s game have already made their major-league debuts (nine from the World squad, and 13 from the U.S.), and a number of others are knocking on the door.
Still, if Upper Deck had a tough time projecting 25-year-olds (not that a card company had PECOTA at its disposal in 1994), we shouldn’t be overly confident in any of these players’ success. That got me wondering—what if these players didn’t have to wait any longer? What if, instead of playing for teams designed to showcase the major leagues' future, these players simply became part of its present? Could they hold their own without additional minor-league seasoning?
Using our Davenport Translations, listed somewhat deceptively as “minor league true average” under the “offense” heading on our statistics page, we can gain a fairly accurate idea of what these players are capable of in the here and now. The “regular” translations spirit each minor leaguer away to a league where hitters average a .260 TAv and pitchers allow 4.50 runs per nine innings, providing us with their expected stats in that arena based on their performance in whatever land and level they hail from. As Clay writes on the DT page, “Ideally, this is how the player would perform if he were called up to the majors right now - allowing for the difference between the real majors and the standard league.” All right, rookies: I’m placing the call you’ve all been waiting for. As the dearly departed Russell Carleton (nee Pizza Cutter) used to say: Warning! Gory methodological details ahead!
I gathered regular translations for each of the 50 players currently on this year’s Futures' rosters (which will likely change before the game is played because of injuries and promotions)—specifically, “true runs” for hitters, and earned runs (which I divided by .92 for an approximation of total runs) for pitchers. Then I divided those individual totals by each player’s actual playing time, to get rates of production per plate appearance and innings pitched, respectively.
With those rates in hand, I set about projecting playing time for our newly minted major-league teams. I awarded 150 innings apiece to the five starters on each squad whom I deemed closest to the majors (mostly determined by level and experience), and handed out 70 innings to the remaining five pitchers on each side. That brought us to 1,100 innings total, per team. The average major-league pitching staff threw 1,442 innings last season, so I credited a colorful-sounding fellow I dubbed “Replacement Pete” with the remaining 342 innings of .66 runs-per-inning pitching (I’d hate to see Replacement Pete’s PAP). Why .66? Primarily because Colin Wyers said so, but also because we use six runs per game as our replacement level for WARP purposes, and 6/9 equals approximately .66.
With all those innings accounted for, I assigned 605 plate appearances to my projected starters at each position, and awarded 200 plate appearances to the remaining non-pitchers. That gave me a total of 6,240 PA, very close to the MLB average of 6,236 last season. With all the playing time parceled out, I multiplied the players’ established rates of production by their projected playing times, which gave me each team’s projected runs scored and allowed. Plug those numbers into the Pythagenpat equation, and—presto!—expected winning percentage. Multiply that by 162, and—presto!—expected wins. Subtract those wins from 162, and—OK, you get the point.
Chances are you didn’t care about anything I just wrote, so let’s get to the results:
It’s clear that the U.S. team is a much more talented bunch than the World team, especially at the plate. However, lest you take these numbers as the conclusive evidence you’ve been looking for that the U.S. is, in fact, better than the rest of the countries in the world combined, bringing all your xenophobic fantasies to fruition, know that this is not the normal state of affairs for the Futures Game, and doesn’t reflect the true U.S.-World breakdown of minor-league talent. A replacement-level club would be expected to win approximately 48 games, so congratulations are in order for the U.S. squad on surpassing that mark.
There are a number of caveats here. For one thing, we’re assuming that these teams would be average defensively. For another, we’re dealing with some selection bias. Roster selection for the Futures teams is a collaborative process, involving Major League Baseball, the Major League Scouting Bureau, MLB.com, Baseball America, USA Baseball, and the 30 organizations that make up MLB, which can choose whether to withhold picked players from the game. According to Kevin Goldstein, who’s been intimately involved in the proceedings, the whole production is akin to the MLB All-Star Game, in that a good deal of weight is given to reputation and past performance. However, plenty of importance is also placed upon in-season play. Since many of these players were chosen based on their superlative first-half performances, and since our translations are based on those statistics alone, we might expect some regression in store, both for the players on an individual level, and for the imaginary full-season squads we’ve constructed (although given these players’ youth, the passage of time should be a boon to them in other ways).
Let’s add a new table to the discussion, keeping the top row from our last one (remember, all winning percentages and win totals are Pythagenpat estimates based on runs scores and allowed):
Sorry, Pirates and Orioles fans—you probably saw this coming. Our estimate of the U.S. Futures team’s performance may be on the optimistic side, but a solid seven wins of daylight are shining between it and the Pirates’ expected record. If the Bucs had traded places at the beginning of the season with the players currently composing the U.S. Futures team, it’s quite likely that your 2010 Pittsburgh Pirates would be better off—not just at some later date, but also in terms of present success.
Not only that, but Pirates owner Bob Nutting would be rolling around in additional piles of hoarded cash. Since all of the players on the Futures teams would make the major-league minimum salary of $400,000 if promoted en masse, the sight of the lot of them wearing the same major-league uniforms would constitute a frugal owner’s wet dream; if they actually succeeded in winning 51 games, they’d have done so for a meager $196,000 per win. For what it’s worth, I sanity-checked these results with Goldstein, whom I hear knows a thing or two about prospects. After studying the roster, he pronounced his agreement: “This team could win 50-plus games.” I neglected to ask him whether he believed that to be true of the Pirates.
The average age of the World roster, as currently composed, is 21.6, and its average “level rating” (which I found by assigning 4 to AAA, 3 to AA, 2 to A+, and 1 to A) is 2.6. The U.S. team’s average age is 21.4, and its average level rating is 2.7. What would these teams look like if we sped forward to a time when their average level ratings were 5 (MLB), and their average age were, oh, say, 27-ish? Fortunately, there’s an app for that—The “Peak Translation,” which, in Clay’s words, “Applies a typical aging pattern to the regular translation, to try and assess how good the player will be at his peak.” I came up with two versions of the peak projection for our squads: one using the same breakdown of playing time I used for the regular translation version (with players closer to the majors given more prominent roles), and another (the “Adjusted-Peak” version) giving the most playing time to the best players available:
According to these translations, the World team’s unadjusted peak would barely be higher than the U.S. team’s regularly translated present. The adjusted-peak U.S. club, on the other hand, would be a borderline playoff team. I’ll get back to you on that in five or six years. You’ve no doubt noticed that in standard sabermetric fashion, I haven’t accounted for one whit of team chemistry in the course of running all these numbers, so my analysis is woefully incomplete. Of course, that’s why they play the games—or in this case, play one game, and call it a season.