"What is the evidence, and what does it mean?" Bill JamesThis article was posted to the alt.sports.baseball.tor-bluejays newsgroup April 1, 2002 before the Blue Jays' first game.
Welcome to my 6th-annual "Blue Jays Projected Record" article, where I add up individual player projections for each Blue Jays player to project a win-loss record for the Jays *before* the season starts. (See disclaimer below on accuracy.)
For each player, I projected his Equivalent Average (EqA). For those new to EqA, an average major league hitter has an EqA of .260, and an EqA of .300 represents excellence. Because of the DH rule, the average American League hitter has about a .264 EqA. EqA is directly related to how much the player helps the team score runs.
The EqA projections were primarily made using the method described in the year 2000 article (link provided below). For players with a lot of recent minor league time (there are a *lot* of those this year!), their input EqAs were calculated as described in the year 2000 minor league stats article (link below). For older minor league players (e.g. Tom Wilson), I sometimes lessened the age adjustment compared to what that article said.
For projected Plate Appearances, I picked a number in line with the player's last few years, though sometimes lowered it so that the total plate appearances for each position would add up to the fixed number (6165 PA for the team, or 757, 739, 721, 703, 685, 667, 649, 631, 613 per lineup slot respectively). If the Jays succeed in increasing their on-base percentage, the team will get more Plate Appearances than estimated. But my offense estimate justs depend on the proportion being approximately right.
Unlike some previous years, I never looked at the STATS Inc. projections this year, though I did use the STATS books a lot for minor league data.
The individual projections are
Age ProjEqA
Stewart OF 28 .282 670 PA
Hinske 3B 24 .276 520 PA
Mondesi RF 31 .272 610 PA
Delgado 1B 30 .315 700 PA
Cruz OF 28 .279 670 PA
VWells OF 23 .257 590 PA
Fletcher C 35 .236 330 PA
Bush 2B 29 .237 300 PA
FLopez SS 22 .254 460 PA
---- ----
.273 4850 PA
Woodward IF 26 .254 250 PA
Wilson C 31 .264 220 PA
Berg IF 31 .246 200 PA
Lesher DH 31 .254 280 PA
Lawrence IF 25 .239 75 PA
Phelps C 24 .262 100 PA
Hudson 2B 24 .252 150 PA
Johnson OF 25 .239 40 PA
---- ----
.254 1315 PA
I weight the EqA projections by plate appearances (which isn't quite mathematically right, but should be close enough, and saves me from estimating on-base percentages). The team offense works out to a .269 EqA. Remembering the 4% adjustment for the DH league, we get
(.269/.260)^2.5 / 1.04 = 5% above league average
For example, if you figure the average team will score 5 runs per game, then the Jays would project to score 5*1.05 per game (5.25).
Notes:
For innings pitched, I picked a number that the pitcher in line with what the pitcher has done in the past (the main exception is that I stretched out Prokopec a little, to 175 innings this year). The total innings were made to add to 1443.
For opponents EqA, I picked a number in line with past performance. Unlike for the hitters, I didn't rely on an automatic projection system to produce most of the numbers. Inspired by McCracken, I did look at what the pithers' past EqAs would have been if Hits Allowed was replaced by (0.300*(AB-HR-K)+HR) (this isn't quite how McCracken would do it; link to his article below).
In principle one should adjust for team fielding if you think it's quality will be different than most of the pitchers are used to. I don't expect a big quality difference this year. The outfield defense, which is the most important, seems likely to be better. The infield defense may not be as strong. Overall, I think it will be about the same as before.
For those not used to opponents EqA as a pitching stat, remember that .260 is average (like having a 4.50 ERA, or a .500 pitcher). The A's front four last year (Mulder, Hudson, Zito, Lidle) had opponents EqAs of .232, .237, .238, .241 respectively. "Replacement level" (what the better triple-A pitchers can probably do) would be about .275.
The individual projections:
Starters:
Age OppEqA
Carpenter 27, 215 IP, .250, 32 GS
Prokopec 23, 175 IP, .260, 29 GS
Halladay 25, 195 IP, .255, 30 GS
Lyon 22, 190 IP, .260, 28 GS
Eyre 30, 100 IP, .275, 16 GS
Loaiza 30, 90 IP, .270, 15 GS
Parris 34, 60 IP, .275, 12 GS
---- ---- ---
1025 IP, .260, 162 GS
Pen:
Escobar 26, 90 IP, .245
Plesac 40, 45 IP, .240
File 25, 80 IP, .255
Borbon 34, 50 IP, .260
Heredia 27, 60 IP, .270
Thurman 23, 70 IP, .275
Cassidy 26, 23 IP, .280
----
418 IP, .259 pen
Overall that works out to a projected .260 opponents EqA for the staff, or a league average staff. There's no need to make a DH adjustment because we assume pitching quality is equal in the 2 leagues.
Notes:
To estimate the projected team winning percentage, we use the Pythagorean formula with an exponent of 1.83:
1.05^1.83 / (1.05^1.83 + 1.00^1.83) = 0.522 wpct -->85 wins, 77 losses
So that's my official projection this year: 85 wins. This assumes a full season is played, which admittedly looks unlikely at this point. If the season is shortened, then the official projection is a 0.522 winning percentage for the portion played.
85 wins would be 5 more wins that last year's team achieved, and 6 more than I projected for last year's team. I don't apologize for not including payroll in the projection system; contrary to most media analysis, it just doesn't belong.
Every year I point out that it wouldn't take a lot of things to go right (e.g. bigger years than projected from the starters, young hitters have breakout years, etc.), for the Jays to win a lot more games than projected. It's still true. This team has a real chance to make the playoffs.
This team also has more flexibility for in-season improvement than last year. For example, if the regular shortstop hits worse than one of the backups, the Jays can give more time to the backup and send the regular to AAA.
The Jays seem likely to make more trades. The new management's trades, even though they've had a long-term focus, have tended to also help the team in the short-term, so I'm not too concerned about trades throwing off the projection. (I don't think there's a need to try to predict the trades.)
The biggest area of vulnerability is the starting rotation, but that's true of almost every team.
Toss a fair coin 162 times, and your best bet is to predict 81 heads, but there's a less than 6% of chance of that happening, even though it really is the best bet. There's a 5% chance that the fair coin will produce either 68 heads or fewer, or 94 heads or more.
So even if this article has accurately estimated the Jays' probability of winning each game, there's a less than 6% chance of the projected win total matching the actual. The 95% confidence interval for an "85 win coin" is approximately 73 to 97 wins.
(But we're not claiming to even be 95% sure of a 73 to 97 win season, because we can't be sure the errors in the individual projections will all cancel out, and we can't be sure that all the other assumptions of this article are valid.)
If we do assume an 85-win quality team, the Jays have a 22% chance of winning 90 games or more, and a 6% chance of winning 95 or more. Of course, this team might be better than 85-win quality. The playoffs are definitely a possibility.
Previous year's articles are on the web at
Only the '98 projection produced exactly the right number of wins (88). Last year's projection was 1 win too low.
Sources:
-- Stephen Tomlinson http://www.stephent.com/jays/ mailto:stephent@magma.ca Ottawa, Ontario "What is the evidence, and what does it mean?" (Bill James)
Last Updated: 2002 Apr 2 (just to add HTML tags)
Comments are welcome at comments@stephent.com.