True talent: Jason Kendall, Sammy Sosa, David Ortizby Sal Baxamusa
March 10, 2008
How do a player's skills change as they get go from fresh-faced rookie to cagey veteran? It's a natural question for a baseball fan, and an important one, too. The usual way of answering this question is to dial up Baseball-Reference and check out year-by-year stats. And that's a pretty decent way to look at a player's career.
But when we do so, we're looking only at how a player's performance changed over time, not how his true talent level changed over time. All performance data are samples from a player's true talent. During times when that sampling happens to be favorable we will get a falsely favorable opinion of their talent; if the sampling happens to be unfavorable we will get a falsely unfavorable opinion of their talent. Is that any way to look at a career arc?
Fortunately for us, the answer has been known for some time now. Don't look at performance data—look at projections instead. Player projections, such as what you might find in the THT Preseason Preview, are simply estimates of a player's true talent at any given time. So it stands to reason that if we want to know a player's true talent at any given time, whether directly before a season or in the dog days of summer, all we have to do is generate a projection.
Easier said than done. Forecasting is tricky business, but the task if far more tractable if we use the Marcel projection system. I've discussed the relative merits of the Marcel system previously, but the chief reasons to use it are 1) all the necessary formulas are in the public domain, 2) all the necessary data are available from the usual suspects (hello, Retrosheet!), and 3) it's pretty darn easy to compute.
So today, we're going to look at a few players and see the changes in their true talent by generating running projections throughout their careers. This way, we'll see how their true talent changes over time without the messiness of the random sampling that is their performance data. You can read more about the method in my previous articles here and here.
Oh sure, he's the butt of jokes involving the word "suckitude" now, but how many people remember that Jason Kendall was on a Hall of Fame career track through age 26? A catcher capable of posting an .850 OPS with his eyes closed while lapping the league in innings caught? I'll take two, please. His truly great years were never appreciated, since he toiled for awful Pittsburgh Pirates. By the time he landed with a contender, he was a shell of his former self. Let's go to the chart:
(As usual, the blue line indicates his performance over the past 365 days and the red line indicates his projection at that point in time. The years are demarcated at the beginning of each season. So on April 1, 2000, his OPS over the previous year was about .940 and his true talent (Marcel projected) OPS was about .875.)
See if you can spot Kendall's catastrophic thumb injury. In 2001, Kendall hurt his thumb badly enough that he required "reconstructive thumb surgery," which—if Tommy John surgery is any indication—involves relocating a toe from the foot to the hand.
If we just look at the performance data (the blue line), it looks like Kendall bounced back from his thumb injury, but the true talent trace (the red line) shows a steady and serious decline to a new level of performance in the .750 OPS range between 2003-2005. His tenure with the A's coincided squarely with a further, and likely irreversible, decline into irrelevance.
The thumb injury most seriously affected Kendall's already iffy power. Here are his career traces for isolated power, which is the difference between slugging and batting averages.
Looks like my stock portfolio.
Given the rivalry between the White Sox and the Cubs, I'm surprised that South Siders don't get more guff for having traded Sammy Sosa for a few years of a pretty bad George Bell. Sosa was just a tools goof with no sense of the strike zone in those days, but of course we remember Sammy for the fearsome hitter he eventually became.
First, his power:
Through the 1997 season, Sosa's true talent isolated slugging looked to be settling into a comfortable plateau in the .200 ISO range as he moved through the generic age 26-29 peak years. And then...whoa. Looking at his true talent trace, his power really shot upward in May of 1998, and continued to rise until it peaked in early 2002 at a remarkable .335.
And what came first, the chicken or the egg? Slammin' Sammy is an oft-cited example that a hitter can develop patience. I've always wondered if Sosa's tremendous power simply caused pitchers to be more cautious, which manifested itself in more walks. Let's check out Sosa's isolated patience, which is the difference between on-base and batting averages.
The performance data (blue line) indicate that in 1998, Sosa began walking at the same time as his power spike. But look at his true talent isolated patience (red line): it barely budged in 1998. In fact, it was only in 1999 that Sosa's newfound patience was reflected in his true talent. I would argue that Sammy's power spike came first, with the patience lagging behind by a full year.
I suppose we had better acknowledge the elephant in the room. Whether or not this kind of analysis can help you spot users of performance enhancing drugs is an open question. I'm skeptical, but the always thoughtful Dan Fox suggests that it could be part of the flagging process. (I'm seeing now that Dan has already profiled Sammy Sosa, so consider my plots to derive from his already fine work.) I'll let you draw your own conclusions.
I considered profiling Jeff Kent or Magglio Ordonez, but I eventually settled on David Ortiz. THT could badly use the traffic, so I figured pandering to the rabid fans in the Northeast would be helpful.
Ortiz has improved his performance every year for the last eight and for some reason has a reputation as being somewhat of an above-average hitter in clutch situations. He sounds like quite the ballplayer; one of these days I'll remove my nose from these spreadsheets and watch him play.
In truth, Ortiz's true talent OPS didn't change much during his Minnesota tenure. I suppose that's partially due to the fact that he had sporadic playing time and wasn't able to accumulate lots of performance data. But his remarkable progression from Big Poopy to Big Papi didn't begin in earnest until 2003. You can see his true talent OPS really taking off in his first season with the Red Sox.
In Ortiz's case, it was the steady increase in his power that landed him where he is today. In fact, his true talent isolated slugging started marching uphill as early as 2001 and has continued apace since then.
Notice that his OPS and ISO traces have essentially the same shape.
Of course, Ortiz isn't just a large mammal who hits the ball a long way. He also gets on base a litte. Was he always this patient?
As with Sosa, Ortiz's patience lagged behind his power. Ortiz's true talent isolated patience was essentially static until 2005, when—whether due to pitcher fear or a discriminating hitting palate—the difference between his on-base and batting averages started to climb steadily. That was four years after the first, modest, increases in his true talent isolated power.
References and Resources
I gathered game-by-game data via the Baseball Musings Day by Day Database, which is a tremendous resource for fans and researchers. David Pinto, proprietor of Baseball Musings and fine all-around gentleman is holding a pledge drive this month. If you enjoyed what you saw here, please consider dropping a dime in his bucket.
Sal Baxamusa is a graduate student in chemical engineering. He can be reached here.
<< Return to Article