Angling for Trout

One of the greatest stories of the 2012 baseball season—and for some, eventually one of the most infuriating—was the arrival of Mike Trout as a regular major league player. He had 40 games and 135 plate appearances with the Angels in 2011, keeping him just within rookie eligibility for 2012. His ’11 performance was honestly nothing special, producing a .220/.281/.390 batting line, but this scarcely damped the anticipation surrounding him, the confidence that he would be something special.

Was he ever. His rookie campaign, five months long due to an early stretch in Triple-A, beat even the most exaggerated hype. He batted a torrid .326/.399/.564, leading the American League in runs scored and in stolen bases, swiping 49 at a 91 percent success rate. He stroked 30 home runs, all from the leadoff spot, making him the most potent leadoff power threat since Alfonso Soriano in his pre-Cubs days, with the potential to become an even bigger one-spot bomber than Rickey Henderson, the all-time leadoff homer champ.

There is no major total-value system that didn’t have Mike Trout as the best player in the A.L. Not bWAR (Baseball-Reference), not fWAR (FanGraphs), not bWARP (Baseball Prospectus), not Win Shares. So it was a horror to many fans of analytics that Detroit’s Miguel Cabrera strolled off with the MVP Award on the strength of his Triple Crown season.

This article isn’t here to rehash that decision. Much.

If you want a devil’s advocate argument, though, there is this. Three times in history a player has won the Triple Crown and lost the MVP vote (in its current incarnation). It happened to Ted Williams twice, in 1942 and 1947, leaving him second both times to players from the pennant-winning Yankees. It also happened to Lou Gehrig, who won the Triple Crown in 1934 only to finish fifth in the voting.

Odds are, you read those lines and thought something like “Man, he got robbed.” Odds also are, MVP voters don’t want to be thought of as robbers by future fans, at least the casual ones who skim the major stats and don’t go mining for the advanced stuff. (Note: Williams and Gehrig did have the top bWAR and top Win Shares in the league each time their Triple Crown came up short. Man, they totally got robbed.)

There was likely also the notion that a Rookie of the Year award was recognition enough for Trout, at least this early in his career. “He’s 21,” some would think, “he’s got time to win his MVP.” One sees this voting mindset sometimes at the Academy Awards. I’ll let you draw the unflattering comparisons yourself.

The expectations for Mike Trout, previously stratospheric, have now sailed somewhere beyond the asteroid belt. After a historic rookie season, the assumption is that he can only get better, being so young. So is he going to be better this year?

The precedents, few as they are for a season of his outstanding level, are against it, but there is some hope for the wide-eyed optimists. I’m going to look at the matter from two converging angles: the performance of previous Rookies of the Year in their second seasons, and the follow-up performance of previous players to post seasons as good as Mike Trout’s—which is not a very big number.

The sophomore slump

One thing to understand before looking at next-year performance of Rookies of the Year (and the great-season players later) is the inherent selection bias. Voters were choosing what they believed to be the best players in the set called “rookies.” While they don’t always have the best value by advanced metrics, they tend strongly in that direction. As a high-performance group, they would have had good luck on average in achieving their positions, and thus been susceptible to regression to the mean, apart from any improvement (or decline) in underlying performance they put together year-to-year.

This would be a bigger problem if we were comparing RoYs to rookies in general. However, we’re comparing them to a specific, if highly unusual, member of their own group. Mike Trout has the same presumptions of both high ability and above-average luck. We should be comparing apples to apples here, even if one apple is remarkably large.

(I used Baseball-Reference’s bWAR metric for these calculations, primarily for its wide historical coverage.)

Previous to 2012, there have been 130 Rookie of the Year awards, including the ties: 97 given to position players, 33 to pitchers. Their average change in performance from their rookie to their sophomore seasons is a decline of about 1 bWAR. This comes from a base of 3.37 bWAR in their rookie campaigns, declining to 2.40. For comparison, Mike Trout registered 10.7 bWAR in 2012. If that number makes you dizzy, I don’t blame you.

(And yes, I said that just to set up a Dizzy Trout joke … but now I’m too ashamed to make it.)

Exactly twice as many RoYs saw their performance drop in their sophomore seasons as saw it rise: 86-43, with one player staying level (John Montefusco, 1975-76). I checked for differing season lengths due to strikes, expansion, etc., and found two instances where this affected the trajectory. Willie Mays went into the Army early in the 1952 season, when his performance was otherwise on pace to better his 1951 rookie year. Fernando Valenzuela had a 0.1 bWAR uptick between 1981 and 1982, but his strike-truncated rookie season masked a substantial drop in WAR per game.

This leaves the overall ratio at 86-43, but it does affect the splits of position players and pitchers. Batters came out 34-63, which should really be 35-62, and pitchers were 9-23 with one tie, but ought to have been 8-24.

A Hardball Times Update
Goodbye for now.

As those ratios suggest, there are differences in the bWAR changes between position players and pitchers. Pitchers winning Rookie of the Year had a higher bWAR than batters (3.66 to 3.28), but fell off more in their sophomore seasons (2.13 to 2.50). Pitchers’ values were also more volatile: their average deviation from the mean fall-off for pitchers was 2.01, compared with 1.63 for position players. This variation holds when examining outright crashes. Eight of the 33 RoY pitchers actually had negative bWARs in their sophomore seasons, compared with a much smaller 11 of 97 ratio for position players.

For samples this small, I can’t call the numbers conclusive, but they are highly suggestive, and do fit in with the belief that pitchers are more uncertain in their future performance. For this reason, the batters-only numbers are probably a fairer comparison to Mike Trout than the overall numbers.

The hitch there is that the numbers are so far apart. Trout’s bWAR in 2012 is more than triple the average Rookie of the Year season, and over 2 WAR ahead of the second-best of the position players. It’s the best rookie performance, award or no award, whatever the position, since pitchers got moved back to 60 feet six inches. How can we draw conclusions for Trout from those much more mundane RoY seasons?

One conclusion stands out. We see frequent regression in even modest RoY efforts. The odds of seeing such a regression in a performance leaps and bounds ahead of them can only be that much greater. If we look at the 10 best bWAR numbers for Rookies of the Year, the conclusion stands out all the sharper. (In this case I will include, and identify, pitchers among the top 10.)

Player/Rookie Year         Age   RoY bWAR   RoY+1 bWAR  Gain/Loss
Mike Trout, 2012            20     10.7          ?          ?
(P) Mark Fidrych, 1976      21      9.3         2.3       -7.0
Dick Allen, 1964            22      8.5         6.1       -2.4
Ichiro Suzuki, 2001 (MVP)   27      7.5         3.3       -4.2
Fred Lynn, 1975 (MVP)       23      7.1         4.2       -2.9
Carlton Fisk, 1972          24   7.0 [7.3]      3.6     -3.4 [-3.7]
Mike Piazza, 1993           24      6.8      3.4 [4.8]  -3.4 [-2.0]
Tony Oliva, 1964            25      6.6         5.2       -1.4
(P) John Montefusco, 1975   25      6.6         6.6         0
Nomar Garciaparra, 1997     23      6.5         6.8       +0.3

Numbers in brackets are 162-game projections for strike seasons.

It takes until the seventh position player, Garciaparra, to find someone who raised his game (and only just) in his second full season. The next to do that is at No. 15, Johnny Bench (4.9 to 6.0). And we see the volatility of pitchers in microcosm: Montefusco held his value (only to lose it immediately thereafter), while Fidrych suffered injuries that wrecked not only his sophomore year, but his career.

Rookies within shouting distance of Mike Trout’s numbers almost all fell, and pretty far, in their next years; rising so high leaves that much more room to fall. This set of comparisons bodes ill for expectations that he’ll exceed his 2012 record—but there is a more promising one.

Last year was Mike Trout’s age-20 season (measured on June 30; he turned 21 in August). That’s a great place to be on the aging curve, a place where the decline of one’s body has scarcely begun, and where the experience one is gaining raises one’s skill level almost unimpeded. The average 20-year-old is expected to be better at baseball a year later, and while Trout’s 2012 was far beyond average, the principle remains. The momentum of natural improvement is at its greatest, and is the best chance he has to avoid sharp regression.

We can see the aging curve pattern at work in Rookie of the Year numbers. The ages of position-player RoYs break down conveniently into something close to fifths, and the bWAR changes into their sophomore years follow the pattern smoothly.

Age Cohort        19-21    22      23     24    25+    All
Number              18     20      21     18     20     97
Avg. bWAR Change  -0.17  -0.365  -0.58  -1.12  -1.64  -0.78

The regression for 25-plus rookies is bad, though there are exceptions. The oldest Rookie of the Year ever was Sam Jethroe, the 1950 winner in the National League at age 33. “Jet” was looking like a Negro League lifer until the original Rookie of the Year changed everything. Jethroe posted a 2.9 bWAR for the Braves in 1950, then a 4.3 in 1951. He fell off badly from there, but not before he had boosted the average a little for the elder RoYs.

Great youth doesn’t immunize against regression, but it is some protection. The rookie figures may end up being the greater comfort, not to Mike Trout, but to Bryce Harper. Winning his Rookie of the Year award in his age-19 season, and with his bWAR of 5.0 much closer to the mean, he’s in less danger of a bad regression. Trout would still be the favorite to have a better 2013, but one can easily envision a case where a six-WAR Harper is hailed a sophomore success, while an eight-WAR Trout is bemoaned as jinxed. Expectations are rough that way.

The 10-WAR club

Trout need not be examined solely as a rookie. He can also be analyzed as a player who had an extraordinarily good season. By Baseball Reference’s numbers, only 47 position player seasons since 1901 have accumulated a bWAR of 10 or greater. Mike Trout’s was the 47th.

He’s also the 20th, as a mere 19 players before him compiled those 46 10-plus years. It might be tedious to name them all—you can find the list at the foot of this article—but I can boil down the essence. There is not a single position player who achieved a 10-plus bWAR year in the 1900s who is not in the Hall of Fame, as a player. Perfect, 16 for 16. As for Trout’s predecessors in the 2000s—Alex Rodriguez, Sammy Sosa and Barry Bonds—I will just say they have not exhausted, or begun, their eligibility with the BBWAA yet.

(It’s fair to note here that, of the four primary rating systems, bWAR is the most generous toward Trout. Both fWAR and bWARP put him in the nines, though still tops in the bigs, and while Win Shares doesn’t map directly onto the WAR(P) systems, Trout was only tied for second in 2012 by that metric, level with Buster Posey and behind Andrew McCutchen. The pantheon that Trout joined with his rookie work may not be as exclusive as bWAR numbers imply: it’s just really fun to think that it is.)

There have been 38 pitchers since 1901 who have managed the feat, but I am excluding them as a poor comparison for Trout. Nearly half of them come during the 1901-1919 deadball era, when the 19th-century dominance of pitchers in WAR numbers had not yet faded. They also show a steeper drop-off in follow-up seasons, an even 4 bWAR compared with 2.60 for the position players.

So what do the remaining comps say? Is it possible for a player who reaches these remarkable heights to maintain altitude, or soar yet higher, next season? The answer is yes, though it’s an achievement so far reserved only for the greatest players of all time.

Of the 46 10-bWAR seasons, eight have been followed up with higher bWARs the next year—which considering the narrow band of room left for improvement before one hits the ceiling of human athletic ability, is quite remarkable. The names of the eight are exactly what you’d expect: Ty, Babe, Babe, Ted, Ted*, Mickey, Willie, Willie. Three other times, a player has exactly matched his 10-plus bWAR the next season: Babe, Willie, Barry.

* I count Ted Williams’ 1942 and 1946 seasons as consecutive, as he served in the Army during the three intervening years. In that context, it is amazing that he improved, and to imagine what the three lost years would have been like.

Two others have had a drop-off of less than one WAR the following season—Honus and Rogers—while a further three have had a larger decline, yet stayed above the 10-WAR mark—Babe, Rogers and Yaz. That’s 16 of the 46 who more or less maintained their greatness the next season, just over a third.

In that context, Mike Trout has a halfway decent chance to roughly repeat his 2012 performance—if you count his 46 predecessors as a valid comparison group. That is to say, if you accept Mike Trout as already being one of the greatest baseball players who ever lived. It’s awfully tough to say that after one (almost) full season. Then again, he did have that awesome season two years younger than any other position player ever did, and the same age as the youngest player ever to hit 10 bWAR.

The position player was Ted Williams. The pitcher was Dwight Gooden. And there is the problem in a nutshell: Whose later performance will Trout’s most resemble? This isn’t to say that a fall-off foretells an entire career of unfulfilled promise (it only might). It does say that plenty has to go right for him to sustain this level of success even into a second year.

Conclusion, sort of

Even if Trout is a historic talent whose name won’t look out of place among the other 10-plus bWAR players once his career is complete, the odds of a repeat performance are against him. The Rookie of the Year numbers put him at about one in three to sustain or improve his performance, not counting the immense risk of regression due to being so far above the rest of the RoY class. The 10-bWAR numbers likewise put him between one in four and one in three, not counting the lack of a sustained track record to show that he belongs in that elite group of immortals.

And there’s a real life complication to add to all the historical comparisons. The Angels are moving Trout to left field in 2013, making way for Peter Bourjos to patrol the central patch. Being shunted to a less critical defensive position will cut into Trout’s defensive value. Even if he outperforms the league’s left fielders by a wider margin than he outperformed its center fielders, which is quite likely, he won’t make up all of that loss. His playing value isn’t lowering, it’s being lowered, and for the Angels, that’s all part of the plan.

There are glass-half-full facts for a little balance. Trout may be a two-to-one shot at best in both categories, but nobody has ever before been in both categories at once. His performance level is literally unique, at a level where comparisons break down. He is also young enough to outrun a good proportion of his expected regression. With some luck, he could outrun it all. All that said, if you made me plunk down my money this minute, I’d have to bet against a 10-WAR 2013 from Mike Trout.

Admittedly, I’ve gone the long way ’round the barn to say something you knew, or could have guessed, from the start: Don’t get your hopes up. There are two justifications for my running these comparisons (aside from it just being fun). First, it’s always healthy to check your assumptions. People with a good grounding in baseball analytics will assume that a super season will be followed by a regression. It doesn’t hurt to run the numbers and confirm that assumption, as well as see what the chances are that the general rule won’t hold this time.

Second, it helps restrain one’s own enthusiasms. I’d love for Mike Trout to have another monster year, to show those voters what they missed the first time. (Even though the probable response would be “See, nothing lost in making him wait a year. Fully vindicated in voting Cabrera MVP.”) It’d be great fun for Trout to be a once-in-a-generation, even once-in-a-lifetime player, tearing up the league like it’s a scale-model movie set and he’s in a giant lizard suit. Fire-breathing optional.

That’s enthusiasm—and that needs a little restraint. Seeing how far both his fellow rookies and the superstars of the sport slid back after Trout-like seasons provides that tug on the reins to keep imagination from running away.

And yet … the numbers aren’t saying “definitely not,” they’re saying “probably not.” And that probability isn’t 99.9 percent, but more like 90 percent, or 80, or even 70. Mike Trout has a decent chance to knock aside all those obstacles, to break the sophomore jinx on a greater scale than it has ever been broken before.

And what would that mean? Well, there will be time enough to analyze history if and when Trout makes it. Good luck, Mike.

References & Resources
Baseball-Reference provided the bulk of statistical underpinning. FanGraphs, Baseball Prospectus, and the Bill James Handbook 2013 also contributed.

For those who were wondering, these are the 20 position players who have had seasons with at least 10 bWAR: Honus Wagner (1905, ’08), Ty Cobb (1910, ’11, ’17), Eddie Collins (1910), Babe Ruth (1920, ’21, ’23, ’24, ’26, ’27, ’30, ’31), Rogers Hornsby (1921, ’22, ’24, ’25, ’29), Lou Gehrig (1927, ’34), Jimmie Foxx (1932), Ted Williams (1941, ’42, ’46), Lou Boudreau (1948), Stan Musial (1948), Willie Mays (1954, ’58, ’62-’65(!)), Mickey Mantle (1956, ’57, ’61), Carl Yastrzemski (1967-68), Joe Morgan (1975), Robin Yount (1982), Cal Ripken Jr. (1991), Alex Rodriguez (2000), Sammy Sosa (2001), Barry Bonds (2001, ’02, ’04), Mike Trout (2012).

A writer for The Hardball Times, Shane has been writing about baseball and science fiction since 1997. His stories have been translated into French, Russian and Japanese, and he was nominated for the 2002 Hugo Award.
Newest Most Voted
Inline Feedbacks
View all comments
Pat G
11 years ago

I though fWAR had him at 10.7

11 years ago

Trout could literally see his WAR (by any measure) drop in half and he’d still be awesome.

Greg Simons
11 years ago

Very nice job, Shane.  While I agree Trout likely will not have another 10-WAR season, even if the position change from CF to LF wasn’t going to happen, he does have another month worth of playing time to add value in 2013.  So even if his performance on a rate basis drops 15%, he still would have similar cumulative value based solely on having more PT.

Bob B
11 years ago

I am always amused that these loving tributes to Trout always completely avoid the issue of his BABIP and the fact that his insanely great July stats obviously boosted his overall numbers, especially since he arrived in May.

Trout’s BABIP in June was .438 and July’s was .416. He came back to reality in Aug. with .328, increasing in Sep/Oct to .378. With his speed his BABIP will presumably be in the .340-.360 range but do you really believe he is going to consistently hit over .400?

Since he is a rookie his hot start in June/July followed by a cooler Aug/Sep/Oct is entirely predictable as he initially shows opposing teams how well he can hit a fastball. On his second and third times around the league scouts have identified whatever weaknesses exist and he gets a nice steady diet of those pitches.

Trout is a great player but he is not the once in a lifetime player many claim. He is a very good player, MikeC sums it up nicely, and his first full year will be a much better representation of his long term performance than Jun/Jul of 2012.

Joe W
11 years ago

Bob B – You’re going to bust on him for 20-40 points of BABIP? First of all, his BABIP in the minors was .406 in over 1100 ABs. I agree it will drop a good amount from that, but a .360 BABIP is not out of the question.

Second, let’s assume a .340 BABIP (low end of the proposed range) for last year. He slashed .326/.399/.564 with a .383 BABIP. With a .340 BABIP (assume 2B/(H-HR) and 3B/(H-HR) remain constant and the same number of HR), his slash line is .292/.369/.521, which looks an awful lot like Miguel Cabrera’s age 21 season in 2004 of .294/.366/.512 which had a BABIP of .335. Not an all-time great year (3.2 bWAR, he was below average in the field), but I think Miguel Cabrera may be making the Hall of Fame someday. (Trout slashes .304/.380/.535 for .360 BABIP, btw)

I don’t think you can say that Trout is an all time great after one year, but I don’t think you can absolutely rule it out either. He’s only 21, and his career could take several different trajectories. However, saying that BABIP regression and one good month point to the fact that he’s never going to be once in a lifetime is just as short-sighted as pointing to one year as a sign of greatness. Lots of great players have insanely good months, but they’re still great players.

Trout’s all around game (good plate discipline for 21, great speed on the basepaths and in the field, above average power) point to the potential for a once in a lifetime player. But to say one way or the other after a little more than one full season’s worth of games is jumping the gun.

Bob B
11 years ago

I’m not necessarily busting on Trout for a high BABIP but simply pointing out how all of the stat heads that sing his praises completely overlook this red flag. Picking and choosing which stats you use to predict future performance is little different than not using the ‘latest information’.

You also cannot simply average out his season. He had one of the truly amazing MLB debuts in June and July where teams were more concerned about Pujols, Hunter, etc. They pitched to him like a rookie and he punished them. Later in the season the teams were smarter, treated him like a threat, pitched to his weaknesses, etc. He came back to earth but still put up very respectable numbers. I think this year will see numbers more like his Sep/Oct line of 289/400/500. His BABIP was .378 but his numbers won’t change much if that drops a little.

That kind of line is certainly not a bust or a flash in the pan. However, he has needs to put a couple of complete seasons together before he’s anointed as ‘once in a lifetime’.

Joe W
11 years ago

Picking is Sept/Oct numbers and presenting it as “true talent” is just as bad when it comes to picking and choosing as selecting his July numbers. You could have said after last April that Albert Pujols was never going to hit a home run in 2012 based on using the “latest information”. Most recent does not mean most accurate. This is why I used his full season numbers as an example, his July still happened and needs to be considered as *part* of the overall picture.

However, I think that we are in agreement that we should wait to pronounce Mike Trout either once in a lifetime or flash in the pan.

11 years ago

When I look at Trout’s season, I always check to “see what Junior did”, as Griffey Jr is the “model” for amazing talent at a young ML age.

In his age 21 season, Jr had ~7 bWAR and age 22 ~5.5 bWAR and then ~8.5 bWAR at age 23.

The only way Trout DOESN’T decrease is if we view his “true talent” at being 10+ WAR. I don;t think ANYONE’s “true talent” is 10 WAR.

So, I think he pretty much HAS TO decrease his performance. Which brings up the interesting questions, “What is the highest WAR a player has had in his 1st 2 ML seasons?”  (Pujols?)


When looking at where/how Mike Trout could perform realistically worse, I think it’s:

1. Hitting fewer Home Runs (HR/FB% = 21.6)
2. Lower BABIP (via lower LD%)
3. League adjustments*.

He faced a high percentage of FB from opposing pitchers.

Basically he seemed to have a “Joe Mauer Year” where basically you set career highs in HR/FB, BABIP, etc all in the same year and put up a ridiculous season (even for a very talented player), and even that player may not be able to ever duplicate that season, despite putting up really good seasons.

The scary thing is that Trout could ADD to his WAR total because he’ll actually play a full season in 2013.

11 years ago

The potential for “overvaluing” Trout’s value is that his bWAR is based on 2.1 WAR from defense and another 1.0 just for his baserunning.  This means 29% of his total value came from outside the batters box.  He was great on both, but to say that just his defense and baserunning alone provide 1/2 the total value of Alex Gordon (who was top 10 in all of baseball last year in bWAR) or was worth 50% more than Derek Jeter’s total value just seems wrong.

Also note he will be changing positions this year, which keeping everything else constant will hurt his defensive and overall WAR totals as CF get a much bigger defensive positoin bonus than corner OFers.

Jack Hughes
11 years ago

Fantastic premise.  Very fascinating analysis.  I tend to think that a few of the commenters are on to something by looking at the monthly numbers.  It would be interesting to see if the variation in WAR from month to month is a good predictor of future performance. 

I’ve never seen it analyzed before but the way people talk about adjustments you would expect to see a players monthly WAR going up and down from month to month or week to week as he adjusts and then the league adjusts until he settles in to his career performance.  Once in a while there might be a sharp step up or down after the number settled in, corresponding to him figuring something out, injury, or hell even PEDs. 

One can imagine a scenario where a rookie hits the league with a bang then has his numbers come down as teams figure him out.  The length of time before he responds (the period) would be a factor as well as the amplitude. 

Not to be too nerdy but you would expect performance to look like a dampening wave.  At least, anecdotally that’s what people seem to be describing. 

If that kind of a curve is real, you could use it as a predictor, and you could also help a player recognize when they are slumping not from their own changes but from other people “figuring them out.”

Paul E
11 years ago

    How about Shoeless Joe Jackson with 18.3 WAR in his first two seasons (ages 23-24)? As far as I can tell, he’s it with the two best seasons.

  Not for anything, perhaps in lieu of the “title” Rookie of the Year, you should have merely included all rookies with WAR >5 or 6? For instance, Rico Carty and Jim Ray Hart had phenomenal rookie seasons in 1964, however, they lost out to Dick Allen and were not included in this discussion…..But, thank you for the research