Schrodinger’s Bat: Projecting Wily Mo

[Ed. note: Today we debut Dan Fox’s “Schrödinger’s Bat” column. Dan will write in this space every Thursday.–JDE]

“Prediction is very difficult, especially about the future.”

–Niels Bohr, Danish physicist (1885 – 1962)

That sentiment, while especially problematic in Bohr’s world of quantum physics, is also more than a bit perplexing in the Newtonian world of Major League baseball. That said, it can be argued that for both Major League front offices and performance analysts–not to mention fantasy league players–projecting performance accurately is a kind of Holy Grail.

It is in that context that the most interesting spring trade occurred when the Red Sox sent pitcher Bronson Arroyo to the Reds in exchange for outfielder Wily Mo Pena. It was essentially a straight-up deal since the Sox also included $1.5 million to compensate for the difference between Arroyo’s $2.75 million and Pena’s $1.25 million 2006 salary.

So in this, my inaugural column, we’ll take a closer look at the Red Sox side of the transaction, along with a bit of insight that their front office shared in making the deal.

A Peek Behind the Curtain

What was perhaps most interesting about the deal were the comments made by Red Sox General Manager Theo Epstein and quoted on MLB.com. In that article Epstein said the following about why the Red Sox were interested in Pena, who will be just 24 this season:

“We think his development is still going on. He’s a unique case. Because of his Major League contract, he was out of options by age 21…

Young players who show that much power at that young an age in the big leagues have pretty good projections in the future to become better all-around hitters. He’s going to strike out a lot. He strikes out on a rate-basis more than anyone else in the big leagues. That’s something we have to be patient with. But there’s a lot of potential there. [He’s] 6-foot-3, 245 pounds, and he’s an above-average runner with a plus arm.”

Epstein then compared Pena to a young Jesse Barfield, who at age 23 hit .253/.296/.510 with 27 home runs for the 1983 Blue Jays (the MLB.com article mentioned that Barfield’s 23 year-old season was 1982 but in fact he turned 23 on October 29, 1982 making his 23 year-old season 1983). By comparison, Pena hit .254/.304/.492 last season for the Reds.

Epstein went on to say:

“I’m not saying he’s going to be Jesse Barfield, obviously, there’s some difference defensively, too. [But] as a rule, really young hitters in the big leagues who have that kind of power develop into better overall hitters. They still strike out, but develop better walk rates, because as they advance and keep the power, pitchers tend to pitch to them more carefully.”

In addition, Epstein noted that Pena’s penchant for hitting left-handed pitchers to the tune of .276/.347/.526 (small sample size alert duly noted) over the last three years would be a benefit to this year’s BoSox. He’ll likely platoon in right field with Trot Nixon who hit–and I use the term loosely–.214/.305/.327 against southpaws in 2003-2005. Now that the Sox have acquired Hee Seop Choi to platoon at first base with Kevin Youkilis, one has to wonder if the Red Sox are taking a page out of Earl Weaver’s playbook.

As an aside, it’s revealing that the talks between the Red Sox and Reds didn’t begin until after the Reds hired new GM Wayne Krivsky away from the Twins on February 8th. As if more evidence was required, that simple fact is yet another indictment of the Dan O’Brien era during which the Reds sat on a surplus of outfielders for two years while trotting out a collection of suspect pitchers four days out of five. Krivsky, on the other hand, took action and likely improved his club by acquiring a solid, if unspectacular, arm in less than two months.

To recap, from the Red Sox perspective, Pena was attractive not only because of his raw tools–which include speed and a 6’3″ 245 pound frame–but because they feel his power at a young age is an indication that he’ll continue to develop as a hitter. Is that really the case? Let’s take look.

Searching for Comps

Obviously, here at BP the place to start when thinking about comparable players is our Similarity Scores listed on the PECOTA card. A quick glance at Pena’s top comparable players by season includes:

Name             Year  Similarity Score
Jesse Barfield   1984     44
Willie Stargell  1964     44
Pete Incaviglia  1988     42
Frank Howard     1961     41
Charlie Spikes   1975     34
Ivan Calderon    1986     34
Mike Marshall    1984     30
Nate Colbert     1970     29
Dick Gernert     1953     28
Kirk Gibson      1981     28

A score of 50 or higher indicates that the player in question was very similar. You can see that along with the aforementioned Barfield, the list is a bit of a mixed bag with the upside being Willie Stargell, Frank Howard, Barfield, and Kirk Gibson. However, it’s also heavy on players with old-hitter skills who did not enjoy particularly long careers.

Incidentally, Dick Gernert, while admittedly not a very close comp for Pena with a score of 28, is a bit of an anomaly on the list. Gernert showed a much greater propensity to walk as a 23 year-old Sox rookie in 1952 and in 1953 walked 88 times in 585 plate appearances. It’s doubtful, though, that Wily Mo will approach that kind of patience.

BaseballReference.com also includes similarity scores using the methodology Bill James created in the mid 1980s where 1,000 is a perfect comparison. Through age 23, the top 10 comparables are:

Jesse Barfield (976)
Rocky Colavito (955)
Pete Incaviglia (949)
Bobby Bonds (947)
Dave Kingman (946)
Willie Montanez (946)
Billy Conigliaro (944)
Willie Horton (943)
Harmon Killebrew (941)
Roger Maris (939)

Once again, Barfield tops the hit parade with Pete Incaviglia also making the list. This list is, if anything, even more skewed towards one-dimensional power hitters (although it’s always nice to have a Hall of Famer like Harmon Killebrew in your comps) with the exception of the elder Bonds. What Bobby Bonds and Pena have most in common, however, is a prolific number of strikeouts.

Besides Barfield, the player who has been most talked about in relation to Pena and the trade is conspicuously absent from these two lists: Sammy Sosa. Slammin’ Sammy doesn’t make the list of comps because when he was 20 to 23 years old he simply didn’t hit for much power, recording slugging averages of .366, .404, .335 and .393. He also showed more speed than Pena, hitting 10 triples and stealing 32 bases at the age of 21. It wasn’t until he was 24 in 1993 that he broke out with his first of two 30-30 seasons. That doesn’t mean that the comparisons aren’t in some ways appropriate, but only that Sosa developed from more of a speed profile while Pena is on the slugging outfielder track. And it should be remembered that, questions about PEDs aside, Sosa didn’t become a 60-home run hitting machine until he fundamentally changed his approach at the plate, which is obviously a difficult characteristic on which to project.

What this does confirm, if nothing else, is that the Red Sox front office is essentially seeing the same comparable players as others in the performance analysis community. And of course, it’s always nice to see insiders on the same page.

But while all of this interesting, let’s take a deeper look at the career paths of players like Pena.

Isolating Wily Mo

As mentioned previously, the fundamental premise in Epstein’s statements is that Pena’s power is an indicator of future all-around hitting potential. In order to see if that’s the case, we need a larger set of comparables and then use those to see how their careers developed.

To do this I created a list of all those seasons since 1955 where a hitter younger than 24 garnered 300 or more plate appearances and recorded a park adjusted Isolated Power (ISO) between 140% and 190% of league average. Using ISO allows us to focus on power at a young age and by park adjusting and scaling it to the league average–which we’ll call NISOPF–we remove biases for extreme parks and the roughly four different eras that have passed since the mid 1950s. Finally, the range 140 to 190 was chosen since Pena’s 2004 and 2005 seasons fit squarely here, coming in with NISOPF values of 174 and 151, respectively. We’ll also track batting average adjusted for league and park (hereafter NAVGPF), as well as strikeout and walk rate.

This gives us 180 total seasons ranging from Hank Aaron‘s 21 year-old season in 1955 through Pena, Miguel Cabrera, David Wright, and Jhonny Peralta in 2005. This list was then whittled down to 169 by excluding catchers, since the rigors of the position have a greater effect on career path. This removed Johnny Bench, Lance Parrish, Gary Carter, Ed Herrmann, Darrell Porter, Matt Nokes, Earl Williams, and Joe Torre, leaving 123 players represented in those 169 seasons.

To give you a feel for the company in which Pena finds himself, the following tables contain the seasons clustered around Pena’s 2005 and 2004 campaigns in terms of NISOPF.

Name                 Year   Age    PA  NISOPF
Willie McCovey       1960    22   307   178.8
Frank Thomas         1991    23   700   175.9
Cesar Cedeno         1974    23   688   175.7
Jim Ray Hart         1964    22   625   174.6
Wily Mo Pena         2004    22   364   174.1
Tony Conigliaro      1964    19   444   174.1
Jeff Burroughs       1973    22   605   173.3
Rocky Colavito       1957    23   544   172.8
Vladimir Guerrero    1999    23   674   172.8

Name                 Year   Age    PA  NISOPF
Ron Gant             1988    23   618   153.1
Jason Thompson       1977    22   668   152.5
Cal Ripken           1982    21   655   152.4
Dusty Baker          1972    23   503   151.7
Wily Mo Pena         2005    23   335   151.6
Luis Gonzalez        1991    23   526   151.6
Rick Reichardt       1966    23   361   151.5
Gary Gaetti          1982    23   565   151.5
Ron Santo            1963    23   687   151.3

Clearly, those players grouped with the 2005 version of Pena are not quite of the caliber of those in 2004, which highlights his slight regression last season.

Finally, 97 of the 123 players are no longer active and so we’ll restrict ourselves to that group in the analysis that follows.

First, let’s take a look at the career progression of the 97 players:

A note on both this graph and the one that follows: the walk and strikeout rate use the Y-axis on the left-hand side of the graph, while NISOPF and NAVGPF use the axis on the right.

As you can see, the players in this group lowered their strikeout rates a bit, though they remained free swingers. They also raised their walk rates as they got into their early thirties. They also slightly increased their ISO relative to the league, but largely treaded water in the batting average department. As a group, by the age of 23 these players had almost reached their maximum power and batting average while they still had a bit to go to reach their optimum strikeout and walk rates.

What is most encouraging, though, is that these players averaged 15.3 years and 6,390 plate appearances, with almost a third of them still active at the age of 37.

It should be noted that the graph shows only those players who were active at each age and so there is some selection bias going on, as players with lesser skills are pushed out of the league as we move to the right.

At first glance it appears that Epstein is right in thinking that Pena’s future is pretty bright.

In taking a second look at the list of qualifying players, what jumps out at you–and what was mentioned by Epstein–is that Pena recorded the highest rate of strikeouts per plate appearance of any of the 169 seasons with his 2005 performance (.346). His 2004 season (.268) placed 8th. His walk rate was also among the lowest, which means his strikeout-to-walk ratio placed him 163rd and 164th.

Given these facts it’s probably appropriate to group the 97 inactive players by K/BB ratio as well, as shown in the following table:

K/BB Rate   Players   Career   PA  NISOPF  NAVGPF
0 - 1.75         33    15.9  7002   142.5   107.5
1.75 - 2.5       25    15.3  6467   154.6   108.1
2.5 - 3.5        25    14.4  5609   145.5   102.9
> 3.5            14    15.2  6205   151.6   106.0

Total            97    15.3  6390   147.6   106.0

While lower strikeout-to-walk ratios are indicative of longer careers, the effect seems to dissipate once the ratio reaches the 2.5:1 range. In other words, young hitters with a strikeout-to-walk ratio above 3.5:1 don’t have careers that are much different than those with ratios of 2.5:1. Indeed, their isolated power and batting average turned out to be a little bit higher. The dividing line, if there is one, appears to be somewhere between a ratio of 1.75:1 and 2.5:1, where those with lower ratios average a couple more years and around 1,000 more plate appearances.

Among those in the > 3.5:1 group with Pena are, not surprisingly, PECOTA comparables Barfield, Incaviglia, and Stargell, along with Andre Dawson, Jim Rice, Nick Esasky, Cory Snyder, and Juan Samuel, among others.

The following graph shows the career trends for that group of 14 players:

One of the things this graph indicates is that, as opposed to the group as a whole, this admittedly small subset’s walk rate did not increase very much, while their strikeout rate decreased dramatically from age 23 to 26. In addition, although ISO spiked at age 25, overall these hitters didn’t show significantly more power nor did their hitting ability seem to improve as they aged, as measured by batting average relative to the league.

Because there were fewer players in this group, those who were still active in their mid thirties–such as Stargell and Dawson and Tony Perez–are responsible for the upward trend as we move to the right. This is an indication that players in this group walk a finer line (literally in part because of their marginal walk rates) and either tend to flame out early (Snyder, Esasky, and Dave Roberts–the one who played for the Padres in the early 70s) or adjust and go on to have success (Stargell steadily increased his walk rate while Dawson hit for a higher average).

Overall, this picture seems to indicate that hitters like Pena don’t necessarily tend to become more patient over time, although they do cut their strikeout rates and continue to put up good power numbers while maintaining their batting average.

Another factor one might be tempted to consider is Pena’s platoon split (he’s hit .237/.286/.454 against right-handers in his career). However, given that he’s garnered only just over 900 career plate appearances, one probably shouldn’t read too much into his splits. This is supported by one of the interesting points made in the recently published The Book:

“A right-handed hitter needs around 2000 plate appearances against left-handed pitchers before his measured platoon split can be considered reliable (in other words, using the measured platoon split is more accurate than assuming the player has an average split).”

Wrapping it Up

As much as we love to wring every drop of insight out of the numbers, all players are, of course, individuals; the vagaries of personality, coaching, and fortune both good and bad all conspire to create the career path of young players like Wily Mo Pena. As a result, our conclusions must always be digested with a generous helping of salt.

That said, given the history of players with similar profiles, it seems Red Sox fans have reason to be optimistic that Pena will enjoy a relatively long career. And although he may not hit the top of his comparable list a la Willie Stargell and instead fall somewhere in the Jessie Barfield range, with adequate playing time and patience, he’ll likely be a productive hitter for at least the next few years.