Prospectus Feature: Breaking Out

In an article that appeared last week on ESPN.com, Peter Gammons provided a list of 20 players whom respondants to an informal straw poll described as candidates for a breakout season. The list, derived from a survey of major league executives, included a mix of pitchers and hitters, five-tool talents and makeup guys, united only in their ability to tease hibernating fantasy leaguers into dreams of greener days ahead.

If one needs any reminder that lists like these are little more than a grownup's version of pin-the-tail-on-the-donkey, it's worth reviewing a similar list that Gammons produced last year. That list includes roughly equal representation of the good (Alfonso Soriano and Derek Lowe), the bad (J.D. Drew), and the ugly (Juan Uribe), as well as four players whose performances were so impressive that they made repeat appearances on this year's list.

Now, none of this is meant to be a knock on Gammons, or the lists he has compiled. Everybody likes to talk about breakout candidates this time of year, ourselves included (Eddie Yarnall, anyone?). Having formerly moonlighted as a daily team correspondent for another baseball website, I can attest to the fact that virtually every player provides at least some excuse each winter for gushing commentary, delusions of grandeur, or other forms of irrational exuberance.

As it happens, however, we're unrolling a new forecasting system at BP this year–one that is also preoccupied with the question of breakout candidates. The PECOTA system–short for Player Empirical Comparison and Optimization Test Algorithm–seeks to identify potential breakouts by comparing a player against a database of his historical peers. In so doing, it comes up with an objective estimate of the probability that a player will display marked improvement in the upcoming season (defined as an increase of at least 20% in his Equivalent Runs per plate appearance, or a decrease of at least 20% in his PERA, relative to a weighted average of his previous three years of performance). We refer to this estimate as a player's Breakout score. Readers interested in a more extensive treatment of the PECOTA system will find it in this year's book, and in the PECOTA glossary provided here.

One brief caveat: the PECOTA system is new technology. That doesn't mean that we stole it from the Raelians, or that we haven’t tested it thoroughly. But sometimes PECOTA provides us with definitive and unexpected answers, and we need to work backwards to try and explain why they came about. That's a bastardization of the scientific method, and I'll ask that you'll excuse me as I run through the hitters on Gammons' list.

Rank on Gammons List, Player, PECOTA Breakout Score

#2.     Nick Johnson, 1B, Yankees    27%
#17.    Corey Patterson, CF, Cubs    27%

Nick the Stick and Corey Patterson are virtual mirror images of one another: most every strength of Johnson's is a weakness of Patterson's, and vice versa. One of the things I've discovered in working with PECOTA is that the players who are most likely to improve are often those who have the most to improve upon, and Johnson and Patterson both fall into that category. Johnson's a big, strong guy who ought to be able to hit for more power, and even a marginal improvement to Patterson's plate discipline would make him a more valuable player. Each is a classic breakout candidate, and an appropriate selection by Gammons.

#7.     Carlos Lee, LF, White Sox    16%
#14.    Jay Gibbons, RF, Orioles     23%

Not a perfect match; Gibbons, at 26, is a year younger, and left-handed. Both players are offense-first corner outfielders who have been faulted at times for their plate discipline. The criticism is only half-correct; Lee and Gibbons don’t draw tons of walks historically, but neither do they strike out at a high rate for a power hitter. That's often an indicator of good bat speed, and the scouting reports for both Lee and Gibbons confirm that. I believe that players who make good contact stand to benefit more from taking additional pitches; that's exactly what Lee did in his excellent second half (.283/.407/.521). PECOTA seems to believe it too; both players have favorable breakout scores for their age.

#10.    Aubrey Huff, 1B, Devil Rays  16%

Huff is a similar hitter to the previous two, and could easily have been listed alongside them, except that he already performed at a very high level last season (.313/.364/.520). He's 26, so the question isn't so much whether he'll improve on last year's level of performance, but whether he'll manage to sustain it. PECOTA thinks there's a good chance that he will, probably trading some batting average for gains in his secondary numbers.

#3.     Erubiel Durazo, 1B, A's      14%
#13.    Preston Wilson, CF, Rockies  11%

Both Durazo and Wilson are likely to get more attention this season, but that will largely be an accident of context. Durazo will get more at bats under Billy Beane's watch, and should make an excellent middle-of-the-order hitter against right-handers. His inability to hit southpaws, however, is a problem that isn't likely to be resolved this late into his career, and that will inhibit any potential he has to become a superstar. After accounting for his move away from Bank One Ballpark, where he hit very well last season, PECOTA thinks there's only about a 15% chance that Durazo will match last year's .944 OPS, but he should turn in a performance in line with his three-year averages.

The thin air should do wonders for Preston Wilson's rotisserie value, but he strikes out far more than Dante Bichette or Vinny Castilla ever did, and unless he changes his approach at the plate, he won't take as much advantage of the environment as he could. Wilson's Breakout score, which discounts any improvement resulting from his change of address, isn't bad for a 28-year-old, but PECOTA puts the odds of his matching Gammons' prediction of 50 HR at only about 4%.

#9.     Adrian Beltre, 3B, Dodgers   12%
#18.    Joe Crede, 3B, White Sox     13%

Remarkable as it might seem, Beltre is nearly a full year younger than Crede. He looked like a budding superstar two years ago, after a .290/.360/.475 season as a 21-year-old in Chavez Ravine, but since then his growth has stalled. The Breakout score listed above is deceptively low; PECOTA, which projects five years of performance going forward, gives him a good chance to have a breakout season in 2004 or 2005. Beltre is a small player by today's standards (5' 11'', 165 lbs.), too small to be a great power hitter, and he might do well to follow Miguel Tejada's example and hit the weight room. If not, he's more likely to follow the long but unremarkable career path of his most historically comparable player, Aurelio Rodriguez. Crede is a notch below Beltre. He should be a solid third sacker for the White Sox for the next few seasons, but PECOTA doesn’t think he has much star potential, and slow third basemen tend to have short careers. Think Bill Melton.

#16.    Toby Hall, C, Devil Rays     13%

At first glance, Hall's breakout score looks reasonably high, but he’s a player that PECOTA doesn't care for much at all. In addition to Breakout rates, PECOTA tracks a companion metric called Collapse; it is the exact opposite of Breakout, estimating the percent chance that a player's productivity will decrease by at least 20%. Hall's Breakout score is 13%, but his Collapse score is 31%, which is very high for a mid-career player.

Catchers are a funny breed. They are more likely than players at other positions to develop late, but also more likely to fade early; both tendencies are represented in Hall's forecast. He's slow as molasses, but puts the ball in play frequently enough that he manages to sustain a decent batting average; with some luck, he could be the new Don Slaught. But he's a funny choice for a breakout candidate, and it's more likely that he'll go on to have an undistinguished career.

#4.     Jeremy Giambi, 1B, Red Sox    5%
#6.     Bobby Kielty, OF, Twins       6%

I was expecting Little G to fall into the same category as Durazo: a stathead favorite who might be somewhat misplaced on this list by virtue of the likelihood of his picking up more playing time. Instead, PECOTA renders a very strong judgment against Giambi; his Breakout rate (5%) is the lowest of any player on Gammons' list, and his Collapse rate (33%) is the highest of any hitter.

The notion has been tossed around that Giambi is a breakout candidate because he's shown a growth curve similar to his older brother. Giambi the Elder does appear on his little brother's comparables list–he's Jeremy's 77th best comp, well behind such notables as Sixto Lezcano, Kevin Maas, and Mike Epstein. Sure, bloodlines might count for something, but that obscures the fact that the two Giambis aren't tremendously similar players. Jason is two or three inches taller, depending on who is doing the measuring, and he played much more regularly than his brother did early in his career, which is normally a positive developmental sign.

More importantly, though, there are important differences in their approach at the plate:

        JEREMY                  JASON
Age     BB rate K rate          BB rate K rate
25      12.3%   20.2%           8.5%    15.9%
26      17.0%   18.7%           9.4%    15.1%
27      19.9%   23.7%           12.3%   15.5%

Jason, while always a patient hitter, did not walk nearly as often as Jeremy did early in his career, nor did he strike out nearly as often. For a player in mid-career, Jeremy's patience borders on the absurd. He saw, on the average, 4.5 pitches per plate appearance last season; no regular player topped 4.3. About 44% of Jeremy's PAs last season ended with a strikeout or a walk, a figure that almost exactly matches Rob Deer's career average. Giambi has "old player's skills" to an extreme, and the PECOTA program thinks that players with that sort of profile don't age very well.

Why is that?

Giambi doesn't put very many balls into play, and when he does, he's one of the slowest runners in baseball. Poor speed and a high strikeout rate are both drags on batting average, and generally a combination to be avoided. Certainly, there are exceptions; some of the greatest sluggers in recent memory have put together spectacular careers with just that collection of skills.

But those hitters had substantially more power than Giambi, and provided greater disincentives for pitchers to challenge them. The worry is that Giambi's approach will cease to be effective if pitchers simply resolve to throw him more strikes. He hasn't displayed enough power, or a consistent enough ability to make contact, to suggest that he'd be able to compensate fully for a decline in his walk rate by improving his contribution in other areas.

It may not be a coincidence that Giambi's tenure in Philadelphia was the most successful period of his career to date; he was in a new league, and by virtue of Larry Bowa's infatuation with Travis Lee, his exposure to pitchers was irregular. The PECOTA system insinuates that, given repeated trials against the same set of pitchers, the weaknesses in Giambi's approach will be exploited. He'll provide the Red Sox with a multifold improvement over the most recent vintage of Tony Clark, but it's almost certain that Giambi's rate of production will be well off from last year's level.

Kielty is a better athlete than Giambi, but shares some things in common with him–his walk and strikeout rates are relatively high for a player with only average power. He also reached the majors at a late age, which normally correlates with a more rapid decline phase. Kielty isn't treated quite as harshly by PECOTA as is Giambi–his Collapse rate is 23%–but he is unlikely to match his .890 OPS from a year ago, let alone to improve upon it.

#20.    Kevin Mench, OF, Rangers     34%

The last player on the Gammons list is, at least according to PECOTA, the best breakout candidate of the bunch. Kevin Mench gets picked on by scouting types because of his stocky frame, and forgotten about by statheads because of a lost 2001 campaign that was undermined by a wrist injury.

Mench's power is for real, and has the chance to exhibit even more of if he can improve upon his plate discipline. Appearances aside, he isn't a bad athlete, and he shows no lingering effects from the injury. PECOTA doesn't expect him to emerge as a world-beater, but he compares well to other big-bootied sluggers who had consistent and productive careers, such as Dave Henderson and Bob Allison. He's one of a number of young hitters who should help the Rangers to make up some ground on the rest of their division.

Coming up next week: The Pitchers.