As some of you may have noticed, there have been some changes in the Minor League EqA page.
Let’s start with the simple. When you go there now, you’ll get a short, simple, fast download, with what is essentially a page of links. The long list of every player in the minors? Not gone, but moved under its own link–so that only the people who really want it have to wait for it to download.
The main feature on the page is a list of all the leagues, along with their stats, sorted by offensive level. I’m always trying to remind people of the context of minor league statistics, and this is one more heavy-handed way to remind people that some leagues (near the top) favor the hitters, while others (near the bottom) favor the pitchers. Click on the league, and you’ll get the information that was on the old minor league page: a top-10 list for each league, a breakdown of league statistics by position (approximated by games played at each position), and a list of all players in that league, sorted by team.
Completely new to the page are the links below the list of leagues–the Future DTs and PDTs. These are the projections that one would reasonably make about how players (DTs) and pitchers (PDTs) will perform at their peak in the majors. These numbers are based only on 2004, and as the link says it uses the player’s performance, age, and level to generate the projection.
First of all, what do I mean by “their peak”? It does not mean their best season; I would expect most players to do better than what is shown for their best season. I am talking about the expected level of performance we would get from this player when he is between 27 and 31. This is a new projection scheme, developed by comparing players’ entire minor and major league careers (not just individual seasons). The resulting routine is substantially different from the ‘MjEqA’ found on the league pages. That value is pretty simple: equivalent average in minor league, times difficulty adjustment for league, equals major league eqa. But the Future DTs calculate the future EqA by using:
- Age. As most BP readers surely know, the age of a prospect can be as important as his actual performance. The more time a player has until he reaches 27, the more he is likely to improve. Moreover, the improvement is not linear; the average amount of improvement goes down as you get closer to age 27. A successful 20-year-old, on average, improves by more in one year than a successful 21-year-old does, and so on. There is an important caveat to that, however: while the youngest players, on average, improve by more than the older ones, the spread of their performance is also wider. The “peak” we are searching for is a combination of established ability plus expected improvement: the bigger the share of established ability (i.e., the closer to 27), the more certain the results.
- Age compared to league. Players who are old for their league–a 25-year-old in A-ball, for instance–have a strong tendency to do worse than expected when and if they get promoted. There appears to be something–I am guessing experience, something knowledge-based rather than physical, what you might call “guile”–that gives them an advantage when playing against younger players, beyond what the normal improvements from being older would create. More to the point, the advantage disappears when they face more age-appropriate competition. Surprisingly, though, I haven’t been able to document any effect from being unusually young for one’s league; the average improvement for them tends to always be based simply on their age.
- Differential changes in statistics. EqA essentially breaks “offense” down into four components: hitting for average, hitting for power, drawing walks, and stealing bases. What we call “normal aging” does not affect these components equally, and in fact most of what we call “normal aging” comes from improvement in power. The Future DTs treat the components separately; as a result, some players that we have gone a little too crazy over in the past based on higher minor league walk rates (ahem, Jackie Rexrode), don’t project nearly so favorably under this system.
- Size matters. This ties in with point #3, that the change in power is the single biggest driver in age-related improvements. Short players, under 6’0″, rarely develop as much power as their taller teammates (the Giles brothers being a major, current exception), and so a height adjustment is built into the projection. There is another issue related to size that is not yet included, because I haven’t yet studied it properly, but the logic goes like this: If improvement is largely driven by power, which in turn comes from the fairly common increase in strength (“filling out”) of players in their 20s, then it stands to reason that players who, at age 20, have already filled out–players like Calvin Pickering, Jack Cust, Prince Fielder–are not going to see the same sort of gains as their skinnier cohorts. Like I said, I haven’t built this factor in, since I first need to build a reliable database of player weights at different stages of their careers, assuming that “reliable database of player weights” isn’t a complete oxymoron.
- Strikeouts matter. Players whose translations suggest that they will strike out 130+ times tend not to develop well, although the data here can be misleading. If you look at players who had 650 plate appearances in the minors and majors who had a lot of strikeouts in the minors, you’ll find that they actually did better than expected, which would make you think that strikeouts don’t matter. The problem is that by selecting only those who have had 650-PA major league careers, you’ve created a selection bias–mainly, the ones who learned to cut down on their strikeouts and make it to the majors. Most strikeout-prone players don’t. This is true even though, as in the majors, a strikeout isn’t any more damaging than any other out. In a projection setting like the future DTs, it isn’t about the value; strikeouts are a proxy for a skill set, the skill of “making contact,” which is a very valuable skill for a successful major league career.
- Extreme performances tend not to be repeated. While it might, in fact, be skill, it is more likely to be the combination of a good skill level plus some random chance in the player’s favor. That is why, in addition to the regular adjustments for league and park effects, the future DTs incorporate a regression to the mean on the player’s rate statistics. For some performances, like Jeremy Reed‘s 2003 singles rate, this creates a substantial tempering of enthusiasm.
The effect of all these changes is, I believe, a better set of completely objective projections than I’ve had before, although still not without its share of misses. I’ve supplied lists of the top 20 minor league players in runs above replacement for each season from 1996-2003 (I’m missing strikeout and height data on a lot of players in the database before 1996). For a set of 654 players with at least 650 PA in the minors and majors since 1970–yes, I’m introducing a selection bias, and no, this isn’t every player, just every player in my database who meets these criteria–and only applying the regression to the mean adjustments to minor league data (or else I could regress completely to the mean and generate perfect results), the improvement between the test I would have used six months ago and today is remarkable (errors comparing minor league numbers per 650 PA with major league performance per 650 PA)
Old New RMS Error RMS Error Batting average .024 .017 Onbase average .029 .022 Slugging average .050 .040 Equivalent average .020 .017 Hits 14.6 10.5 Doubles 8.1 5.0 Triples 2.4 1.9 Home runs 6.9 5.4 Walks 11.1 10.5 Steals 6.4 5.8
Of the 654 players, 272 of them, 42%, had an EqA difference of 10 points or less. A total of 495 of them, 76%, were within 20 points. (The pitching DTs are a completely different story, again using performance, level, and age in combination, but we’ll cover that in a separate article.)
Note: Players with XX’s in their names are players who have not had their biographical information entered into the database, or who have changed organizations and haven’t been updated. The start of short-season play creates quite the hammer for data entry. Note that these players will all show an age of “23” because of a default entry in the program. Player positions and games played have been entered, but because we don’t have defensive information yet, all players will show a fielding level of “0.”
Top prospects, 2003 Projected future peak performance BA OBA SLG EQA EQR RAR 1 B.J. UPTON .302 .388 .528 .307 96 49 2 JEREMY REED .325 .396 .532 .314 92 49 3 JOSE LOPEZ .292 .352 .566 .303 98 48 4 PRINCE FIELDER .286 .373 .511 .299 91 43 5 ANDY MARTE .278 .367 .528 .300 87 42 6 JOE MAUER .322 .385 .486 .299 87 41 7 JOSH BARFIELD .296 .359 .511 .293 91 41 8 GRADY SIZEMORE .294 .363 .521 .297 85 40 9 JEFF FRANCOEUR .291 .341 .530 .290 87 38 10 ALEXIS RIOS .309 .361 .489 .290 81 35 11 MIGUEL CABRERA .321 .395 .609 .329 58 34 12 DAVID WRIGHT .259 .350 .493 .286 80 33 13 J.J. HARDY .280 .365 .505 .294 73 33 14 BRIAN MCCANN .292 .342 .545 .293 72 32 15 ERICK AYBAR .301 .346 .497 .286 77 32 16 VICTOR DIAZ .291 .346 .503 .285 77 32 17 CHASE UTLEY .289 .362 .489 .290 71 31 18 AARON BALDIRIS .295 .367 .450 .282 76 30 19 BOBBY CROSBY .267 .353 .473 .284 75 30 20 JEFF MATHIS .282 .351 .487 .284 73 30 2002 1 JOSE LOPEZ .337 .381 .591 .320 103 58 2 ANDY MARTE .278 .349 .552 .298 88 41 3 TRAVIS HAFNER .298 .407 .499 .311 78 41 4 MIGUEL CABRERA .284 .349 .539 .296 86 40 5 WILL SMITH .296 .342 .533 .292 91 40 6 BRENDAN HARRIS .295 .355 .525 .295 82 38 7 JEFF MATHIS .288 .351 .527 .294 84 38 8 JASON STOKES .290 .374 .564 .311 71 37 9 JOSE REYES .289 .347 .495 .285 91 37 10 JUAN TEJEDA .286 .354 .489 .286 86 36 11 ROCCO BALDELLI .309 .355 .528 .295 80 36 12 BRAD NELSON .269 .337 .514 .285 84 34 13 JUSTIN HUBER .278 .371 .487 .293 75 34 14 VICTOR MARTINEZ .286 .363 .508 .294 76 34 15 JASON KUBEL .290 .352 .528 .293 72 33 16 SHAUN BOYD .286 .348 .478 .282 82 33 17 HEE CHOI .258 .367 .458 .285 79 32 18 SHIN-SOO CHOO .280 .376 .445 .286 78 32 19 GRADY SIZEMORE .290 .378 .452 .288 72 31 20 MARK TEIXEIRA .280 .371 .539 .305 62 31 2001 1 HANK BLALOCK .315 .388 .559 .315 101 54 2 ADAM DUNN .292 .403 .562 .322 77 43 3 ADRIAN GONZALEZ .294 .365 .524 .298 92 43 4 WILSON BETEMIT .304 .360 .522 .296 87 40 5 MIKE CUDDYER .269 .360 .498 .290 89 39 6 JUAN RIVERA .296 .349 .530 .292 86 38 7 JUSTIN MORNEAU .289 .364 .500 .294 81 36 8 KELLY JOHNSON .269 .368 .498 .295 78 36 9 MARCUS THAMES .269 .352 .486 .285 87 36 10 SEAN BURROUGHS .315 .384 .510 .304 70 35 11 CHRIS SNELLING .304 .377 .468 .291 74 32 12 JASON LANE .263 .341 .486 .281 83 32 13 JOSE REYES .307 .352 .532 .295 70 32 14 GARRETT ATKINS .286 .372 .447 .284 77 31 15 CARLOS PENA .250 .362 .468 .286 73 30 16 JASON BOTTS .283 .372 .464 .288 70 30 17 WILL SMITH .284 .335 .492 .278 79 30 18 BRANDON PHILLIPS .280 .355 .465 .282 73 29 19 MARLON BYRD .276 .345 .464 .279 79 29 20 JESUS COTA .291 .386 .543 .311 53 28 2000 1 ALBERT PUJOLS .300 .361 .579 .309 94 48 2 HANK BLALOCK .301 .372 .533 .304 96 48 3 AUSTIN KEARNS .269 .363 .518 .296 89 41 4 KEVIN MENCH .278 .362 .521 .297 89 41 5 CARLOS PENA .261 .359 .466 .285 86 35 6 CARL CRAWFORD .302 .348 .479 .283 87 35 7 JOSE ORTIZ .303 .354 .503 .290 82 35 8 JOE CREDE .283 .350 .482 .283 85 34 9 TONY TORCATO .310 .357 .498 .290 79 34 10 AUBREY HUFF .282 .357 .522 .294 70 32 11 JOSE CASTILLO .284 .336 .505 .281 81 31 12 SEAN BURROUGHS .300 .384 .478 .295 68 31 13 JASON HART .268 .338 .474 .277 82 30 14 COREY PATTERSON .260 .335 .504 .282 73 29 15 HEE CHOI .259 .343 .490 .282 75 29 16 KEITH GINTER .269 .368 .444 .283 73 29 17 VAL PASCUCCI .259 .354 .455 .279 76 29 18 ADAM DUNN .252 .365 .447 .283 72 28 19 BRIAN COLE .272 .325 .472 .273 80 28 20 BRAD WILKERSON .253 .359 .451 .280 70 27 1999 1 NICK JOHNSON .307 .456 .523 .338 102 63 2 SEAN BURROUGHS .330 .418 .517 .319 90 50 3 ARAMIS RAMIREZ .301 .396 .536 .314 89 48 4 D'ANGELO JIMENEZ .308 .373 .499 .296 89 41 5 MIKE CUDDYER .283 .376 .499 .298 84 40 6 VERNON WELLS .299 .360 .518 .297 85 40 7 JASON ROMANO .292 .359 .539 .299 82 39 8 STEVE COX .291 .363 .498 .292 87 39 9 DEE BROWN .286 .377 .488 .296 81 37 10 ADAM PIATT .256 .360 .475 .286 81 33 11 MIKE RESTOVICH .276 .359 .482 .286 80 33 12 TONY MOTA .297 .369 .557 .308 65 33 13 DAVID ECKSTEIN .286 .387 .409 .284 77 31 14 PAT BURRELL .267 .362 .479 .286 74 31 15 ANGEL SANTOS .267 .348 .481 .282 76 30 16 AUBREY HUFF .270 .344 .490 .282 77 30 17 MIKE LAMB .278 .341 .483 .279 81 30 18 COREY PATTERSON .278 .323 .530 .283 73 29 19 LUKE ALLEN .285 .341 .478 .278 79 29 20 RICO WASHINGTON .277 .361 .445 .279 77 29 1998 1 CALVIN PICKERING .278 .391 .508 .305 95 47 2 ERUBIEL DURAZO .300 .398 .566 .321 84 47 3 ERIC CHAVEZ .290 .356 .556 .302 95 46 4 MITCH MELUSKEY .295 .402 .502 .311 76 40 5 JOE CREDE .287 .363 .508 .294 85 39 6 DOUG MIENTKIEWICZ .279 .375 .470 .291 86 38 7 GABE KAPLER .278 .345 .515 .288 88 38 8 MIKE CUDDYER .271 .352 .509 .290 85 37 9 PETER BERGERON .293 .373 .450 .287 89 37 10 RUBEN MATEO .295 .359 .538 .301 76 37 11 NICK JOHNSON .289 .420 .517 .321 65 36 12 TROY GLAUS .270 .366 .538 .302 75 36 13 ALEX ESCOBAR .272 .355 .522 .296 75 35 14 SHAWN GALLAGHER .269 .347 .492 .284 84 34 15 DERNELL STENSON .259 .365 .452 .283 82 33 16 LANCE BERKMAN .257 .362 .471 .286 80 33 17 MICHAEL BARRETT .296 .342 .529 .290 73 32 18 TROT NIXON .278 .362 .461 .283 80 32 19 WILTON VERAS .297 .336 .535 .289 75 32 20 CARLOS FEBLES .277 .377 .444 .289 72 31 1997 1 BEN GRIEVE .293 .400 .554 .319 97 54 2 ARAMIS RAMIREZ .271 .370 .514 .298 89 42 3 JUAN ENCARNACION .294 .365 .532 .302 87 42 4 NICK JOHNSON .275 .386 .501 .303 85 42 5 BRENT BUTLER .289 .366 .523 .299 85 40 6 ADRIAN BELTRE .283 .372 .517 .301 80 39 7 DERNELL STENSON .275 .373 .494 .295 84 38 8 MIKE LOWELL .283 .361 .504 .292 84 37 9 PAUL KONERKO .279 .366 .508 .295 79 36 10 RUBEN MATEO .305 .362 .567 .307 72 36 11 CHAD HERMANSEN .267 .361 .488 .290 81 35 12 DAVID ORTIZ .281 .340 .512 .285 86 35 13 ROBERTO PETAGINE .266 .372 .485 .293 76 34 14 RICHARD HIDALGO .291 .347 .496 .284 82 33 15 MIKE DARR .293 .354 .469 .282 78 31 16 ADAM JOHNSON .265 .330 .508 .280 79 30 17 CALVIN PICKERING .268 .349 .492 .285 74 30 18 DARYLE WARD .294 .359 .479 .286 72 30 19 MARK KOTSAY .274 .360 .477 .287 71 30 20 SEAN CASEY .324 .389 .538 .313 58 30 1996 1 VLADIMIR GUERRERO .323 .397 .601 .327 106 62 2 PAUL KONERKO .295 .394 .558 .317 99 54 3 GABE KAPLER .279 .354 .557 .302 100 49 4 ADRIAN BELTRE .281 .359 .556 .302 91 44 5 ANDRUW JONES .290 .370 .539 .304 83 41 6 CHAD HERMANSEN .270 .363 .513 .296 87 40 7 RICHARD HIDALGO .297 .354 .526 .294 85 39 8 DERREK LEE .255 .343 .530 .291 87 38 9 RUBEN MATEO .289 .342 .539 .293 85 38 10 TODD WALKER .281 .348 .499 .286 86 36 11 BEN GRIEVE .282 .358 .490 .288 83 35 12 EDGARD CLEMENTE .279 .352 .498 .288 80 34 13 SCOTT ROLEN .292 .384 .500 .301 71 34 14 MIKE RENNHACK .291 .352 .518 .291 74 33 15 RICKY LEDEE .267 .345 .496 .284 79 32 16 FRANK CATALANOTTO .273 .356 .464 .281 78 31 17 MARIO VALDEZ .286 .383 .482 .297 66 31 18 BRENT BREDE .282 .379 .430 .284 74 30 19 DANTE POWELL .267 .347 .465 .279 81 30 20 DARIN ERSTAD .304 .380 .501 .301 62 30
Thank you for reading
This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.
Subscribe now