Thursday, April 29, 2010

Adam Dunn and "Old Player Skills"

Yesterday, we and about a million other people took the Phillies to task for the Ryan Howard contract. So with the concept of "old player skills" and contract extensions on everyone's mind, now seemed like a good time to look at Adam Dunn.

When Dunn first broke into the league, he had a bit of speed, but by his mid-20s he had come to epitomize the concept of the "young player with old player skills." And with his contract up at the end of the year, the Nationals need to make a decision soon about whether to re-sign him or trade him (or hold on to him for a pennant race, I guess).

For the uninitiated, the key old player skills are power and pitch recognition, which translate into lots of homers, walks, and strikeouts. "Young player skills" are speed, defense, and contact ability, which roughly translate into stolen bases and batting average.

To be clear, there's nothing wrong with a player who has old player skills--you can be a great offensive player with lots of homers, walks and strikeouts. Like Adam Dunn. The problem is that young hitters who display
only old player skills tend to age really badly.

A hitter with young player skills might lose speed as he gets older, but he can compensate for that physical decline by learning better pitch recognition and strike zone discipline. In other words, as their young player skills erode, they replace them with the development of old player skills.

A good example is Reggie Jackson. As a young player, Jackson was an all-around great offensive player. As he got older, the power and walks became a more important part of his game, as his speed and batting average declined. Here's a chart that shows pretty well what I'm talking about:

The red and pink lines show Reggie's batting average and stolen bases (young skills) and the green lines show his HR rate and walk rate (old skills). The blue line is wOBA, or his overall offensive value. All these numbers are presented as a ratio of league average. So for instance in his age 25 season, Jackson hit 2.5 times more homers per plate appearance than the average MLB hitter that year.

There are a set of things that jump out at you here. First, there's a lot of zig-zagging and criss-crossing of lines. That shows how he was able to use a wide variety of skills to be successful in different ways across his career. Early on, he has speed and power, but isn't particularly patient. The speed disappears fastest, and his batting average starts to fade a couple years later. Meanwhile, his home run rate actually peaks at age 33, and his very best walk season comes at age 39. This shows how a player with both y0ung player and old player skills can remain a dynamic player into his 30s.

In his last years, it's the batting average that clearly correlates with his steadily declining overall value. What usually happens with older hitters is that bat speed and reaction times slow, holes in their swings grow, and pitchers exploit those holes and gradually win more often, and in Reggie's case finally pulling him below average around age 40.

Now let's look at a player who aged just a bit less well: Richie Sexson. Admittedly, the comparison is a little inappropriate, because Jackson was overall just a way, way better player. But with that caveat, here's Sexson's chart:

First big thing: much less criss-crossing of lines. Like Jackson, Sexson's power was always evident, but he never had Reggie's speed and his batting average only barely edged above average for a year or two around his peak. Also, whereas Jackson didn't really maximize his strike zone discipline until his early 30s, Sexson's walk rate peaked in his late 20s. And a good thing, for him--if he hadn't figured out how to take those extra free passes, he never would have become an all star and never would have gotten that big contract from the M's, because he didn't have the skill set to be successful any other way.

But at age 30-31, gravity starts to kick in, as it does for all of us. Since Sexson had already maxed out his narrow skill set of power and walks, the inevitable reduction in bat speed and reaction time left him with no other skills to lean on and he went right off the cliff. By age 32 he was a below-average offensive player at a premium offensive position, and by 33 he was finished.

Now, let's look at Dunn
(note that the last season listed is his current year, and there's a ton of sample size noise there):

This chart certainly isn't a clone of Sexson's (for starters, Dunn is better), but it's a lot more like Richie's than Reggie's. Again, little criss-crossing: basically one pretty consistent set of skills, walks and power, driving his production throughout the productive portion of his career.

This season, his walk rate is as good ever. That makes sense. One would expect that the big dip in home run rate this season should correct itself shortly, unless Dunn is hurt. And at least for now,
he's still a solidly above average producer.

The line to watch is the red one--the one headed ever so slightly in the wrong direction. The speed of Dunn's decline will track with how fast that line falls, or rather how quickly he loses bat speed and reaction times, creating bigger holes in his swing for pitchers to exploit.

Drilling down one more level, we can look for a few other stats to tell us what's going on with that red line. First, there's BABIP. Batting average is heavily dependent on luck--whether the non-HRs fall in or not. Dunn's BABIP is .277. That's a bit low, but not hugely out of line. Career, he's a .291 BABIP hitter, and slow, fat guys who hit against the shift can be expected by be around there.

Looking at Sexson's collapse and two other well-known fast-faders, Mo Vaughn and David Ortiz, we can find a couple other leading indicators. Strikeout rate is frequently associated with batting average, but since these guys all had sky-high strikeout rates to begin with, those rates don't always jump a huge amount. In fact, it's not uncommon to see these types of players actually improve their K-rates in their collapse year.

What you see more often is an increase in the percentage of balls hit on the ground, a higher percentage of fastballs thrown to them, and the hitter making more outs on those fastballs. This is because they're getting slow on fastballs, start to guess, top more balls into ground outs... simply not driving the ball as much.

All this you can see on the Fangraphs player page--the batted ball section gives you GB% and the pitch type data gives you percentage of fastballs thrown, and the hitter's performance on the fastballs is measured in the linear weights. Sadly, Dunn is in fact seeing more fastballs and hitting more balls on the ground. He's doing ok on those fastballs though, so the data don't all line up... yet.

Bottom line, it's too early in the season to say whether he's just starting low or hitting the beginning of the end. But there are reasons to wonder, if not quite worry, and fans should be watching those key rates closely over the rest of the season.


Positively Half St. said...

Excellent lesson in deeper stat analysis. Thanks. I hope that Dunn can reverse the trend by picking up once the number of at-bats for the year is more statistically signifiant.

Harper said...

The hard part (for some team) is going to be figuring out when exactly the drop is going to happen and trying to get a contract around that. I don't think Dunn is at that "avoid at all cost" level but he's probably getting into the "no more than a 2-yr deal" area.

(by the way - showing percentiles rather than deviation from average might make the drop in average clearer.)

Eric said...

I like Dunn a lot and I'll wish him well when he heads to Boston (either via trade of FA), but I very much hope he isn't in the line-up the next 3-4 years.

I think Marrero will be ready soon, but he's a power righty very similar to Willingham and our roster is already short on power lefties. If only there was a power lefty phenom available in the draft...

Grover said...

My concern in letting Dunn go is the lack of power available to the Nats in the minors or free agency. We are likely to roll out something like a Strasburg-Zimmermann-Lannan-Marquis- Wang/Detwiler type of rotation next season. It would be a shame not to pair that with a lineup that can score some runs. If they can get Dunn to sign another two-year deal I'd be delighted.

niggledork said...

Enough said.

Wombat Rampant said...

What Grover said - if there's no good lefty slugger in the system, and there isn't a good one available for a reasonable salary in the free agent market, then it's obvious that we need to re-sign Dunn. Not for the retarded kind of contract the Phillies gave Howard, but maybe something loaded with incentives.

Steven said...

First base is the most loaded position in the league in FA: Adrian Gonzalez, Prince Fielder, Albert Pujols...

Wombat Rampant said...

Yes, Steven, it is. Do you really think the Lerners are going to spend the money it'll take to get one of those three if Dunn is still hitting 30-40 dingers a year, getting on base at a good clip, and showing even a mediocre glove at first? I doubt it, myself.