Chase Headley and the Art of the Comp

Reader TexPadre recently posed the following question:

When will Chase Headley show a decent amount of power? Or will he ever?

Headley, you may recall, was the subject of much hype in spring training 2008. He had just come off a monster season (.330/.437/.580) at Double-A San Antonio and people got a little silly, mistaking him for a future star rather than what he is — a solid support player. Headley is what I would call a classic #6 hitter on a championship team — not enough on-base skills for the top of the order, not enough power for the heart of the lineup, but enough of each to be useful in the bottom third.

This got me thinking about process and approach as it relates to the evaluation of individual players. It’s the type of thing baseball executives have to do all the time. The basic process is to gather inputs, analyze data, and reach tentative conclusions — always leaving room for adjustments as conditions change, new information is gleaned, etc. Then, based on this knowledge (which is necessarily imperfect; we are talking here about probabilities, i.e., likely outcomes), they must make decisions that involve large quantities of money and affect the future direction of the organization.

More recently, similar questions have become “important” to aficionados of fantasy baseball, which over the past quarter century has transformed itself from a hobby for ubergeeks into a lucrative industry. Careers have been made on the basis of projection systems that make various claims which sound impressive (my incredibly unsophisticated IVIE system, for example, promises to be “comatosely accurate,” whatever the heck that means) but which are basically educated guesses. Convince people that your guesses are good enough to be worth paying for, and you make money… but I digress.

So, we’re talking about Headley. We’re talking about projections and comps. Methodology and what-not. Okay, wait, before we get there, I want to show you one more thing. Here is what we (you and me) projected for Headley this year versus what he has actually done so far:

Chase Headley, 2009: Projected vs Actual
Statistics are through games of September 27, 2009, and are courtesy of Baseball-Reference.
IVIE 550 .270 .345 .448
Actual 591 .264 .342 .399

We were a little optimistic on the slugging percentage, but everyone will agree that this qualifies as comatosely accurate. If you don’t agree, kindly adjust your reality to conform.

Getting back to the original question, I responded as follows:

I could see Headley topping out in the 20-25 HR range, but yeah, he’s more of a doubles guy. If he tightens up his strike zone a bit, I’m envisioning a Kevin Millar/Lyle Overbay type of hitter, which is pretty valuable at third base.

Reader Tom Waits stepped right up (three for a dollar) with some thoughts of his own:

Corey Koskie is my current fave comp. He was consistently in the 115-120 OPS+ range in the AL, against a league average of 100; Chase looks to be headed for the 110 range against a league average of 94.

Koskie is another solid comp, and I especially liked the fact that Tom explained his reasoning. Because from where I sit, a well-considered thought process is at least as important as the ability to utilize a particular technique.

I don’t know how Tom came up with his comp, but I will tell you how I arrived at mine. I often forget to explain my process because these patterns of thinking become so ingrained in who I am that it’s hard to go back and break it all down into bite-sized steps. But maybe you will find my process interesting and/or useful; worst case, you will know that when I spout names of comparable players, some actual thought went into it.

And the reason I got into projections earlier is that — no matter what anyone might want to believe — there’s at least as much art here as science. There can be tremendous value in developing and deploying complex algorithms that take massive amounts of data and consolidate it into something manageable and useful.

It is important, however, to recognize that such algorithms provide us with a starting point, not an ending point. We use these tools to help us better understand the world around us, to begin an inquiry into something that sparks our curiosity.

Even the best algorithms cannot synthesize information and make judgments based on it. Algorithms are tremendously hard workers but they are stupid; they are tools (is this why we call stupid people “tools”?). Once the necessary data has been extracted using these tools, actual people must step into the process and decide what to do with it all in a way that makes sense and, with luck, adds to the greater body of knowledge and understanding.

Forgive me if this seems obvious, but folks screw it up all the time. You’d be surprised by the things people claim to know on the basis of what a computer told them. It doesn’t work that way. Somewhere along the line, you have to interpret the data.

Back to the discussion at hand, in seeking out comps for Headley, I first examined his entire minor- and major-league records to date. I won’t bother reposting all of that here (try the Internet), but the salient point is that in the context of Headley’s overall body of work, his 2007 campaign appears to have been an outlier. This doesn’t mean we discard that season entirely (after all, it happened), but using it as a baseline creates unreasonable expectations.

Next, I made a couple of assumptions:

  • Headley’s past performance is indicative of his likely future performance, inasmuch as the same holds true for any player, which is to say… somewhat
  • There is an adjustment period on reaching the big leagues; most players don’t dominate on arrival

I revisited Headley’s record and noticed a couple of things:

  • Wherever he’s been, he’s hit a ton of doubles (in the 31-43 range in every full professional season)
  • His strike-zone judgment has deteriorated as he’s moved up the ladder:
    Year Level BB/SO
    2006 A+ .771
    2007 AA .649
    2008 AAA .477
    2008 MLB .288
    2009 MLB .443

This latter finding should come as no surprise. The theory is that as a player advances, he faces better competition. Nobody exists in a vacuum. A guy doesn’t hit .330/.437/.580 simply because he wants it to be so. Other guys are trying to prevent him from doing just that, and those at Double-A tend to better at it than those at A-ball. Don’t get me wrong: The fact that Headley tore it up at Double-A is cause for excitement, but only in the larger context of his output. In other words, this excitement should have been tempered a little by knowledge of his performance at Lake Elsinore a year earlier.

We’ve seen Headley’s plummeting BB/SO ratios, but it’s not like he didn’t gain anything as a result. His ISO jumped more than 100 points from 2006 to 2007, and it held at the new level (around .250) in Triple-A the following year. That’s a good sign, but at some point, the grip-and-rip approach loses effectiveness. It’s that whole “other (more skilled) guys trying to prevent him” thing again.

Although Headley’s plate discipline fell apart when he first arrived in San Diego, he managed to hold his own as a rookie, hitting .269/.337/.420 despite playing a position where he was visibly uncomfortable (well, I was uncomfortable watching him; it is possible that I’m projecting my own feelings onto him, but I doubt it). That checks in as a 108 OPS+ (not the most sophisticated metric on the market, but a reasonable enough barometer for our purposes — if we had to base millions-dollar decisions on our findings, then we might want to reach for something a little more precise), which isn’t too shabby for a 24-year-old kid facing big-league pitching for the first time.

So, remember the bit a few paragraphs ago about an adjustment period? That’s what Headley is going through now. I’d feel a lot better if he were experiencing it at his natural position so he wouldn’t have to make two adjustments (one to big-league pitching, one to playing left field) at the same time, but what are you gonna do.

What I like about Headley’s 2009 campaign is that his overall line isn’t all that different from what he did as a rookie but he’s tightened his strike zone in the process. Yeah, the home run rate is down, but it’s worth recalling how Headley achieved his early homer binge in ’08. Here’s a chart that ran in the Ducksnorts 2009 Baseball Annual:

Chase Headley’s Plate Discipline, 2008
6/17/08 – 7/11/08 88 .276 .284 .506 5 0 27
7/12/08 – 9/28/08 280 .266 .354 .389 4 30 77

Do you think Headley could have sustained any amount of success with the approach that led to his initial power spike? Here’s a hint: No friggin’ way. Don’t believe me? Ask Rolando Roomes.

If we look at those first 88 plate appearances as sort of a “Hey, I’m here but I don’t really know what I can do yet” transition period, we can see the latter part of his rookie season as a more accurate representation of his ability as it existed at that point in time. In other words, here is a kid that hit a little, drew some walks, and showed flashes of power (though typically not manifested as home runs). We have the starting point for a young player that figures to improve — incrementally — as he gains experience… which is exactly what has happened this year.

Headley is making strides in both the plate discipline and power areas. They are very small strides, but he’s 25 years old and still playing (for the most part, until Kevin Kouzmanoff’s recent injury) out of position. Aside from the fact that some people bought into the misguided notion that Headley is a star in the making, there is absolutely no cause for panic. He is developing into what he will eventually become: A prototypical #6 hitter on a championship team.

Shoot, I keep forgetting to talk about Millar and Overbay. Again, this is at least as much art as it is science, so bear with me. The process I followed in coming up with those comps for Headley involves answering a series of questions and goes something like this (your specific questions and answers may vary; that’s fine as long as you have a process that works):

  1. Based on what you’ve seen of Headley so far, what do you think his current level of production is?
    This is easy because his first two seasons have been nearly identical. I think he’s a .265/.340/.405 hitter with improving plate discipline.
  2. Based on his age and experience, do you expect him to improve in the future? To what degree?
    Yes, but not by a lot. He is 25, which suggests additional growth may be possible but not as much as we might expect from a 22- or 23-year old.
  3. Are there mitigating factors?
    Petco Park hurts his offensive game. Headley hit one out to dead center in Pittsburgh the other night that would have been a fairly routine fly ball in San Diego. Also, he is playing out of position, although it is not certain how much offensive gains he would make if he moved back to third base full time.
  4. Given the above information, what do you expect his peak to be and how long will he play at that level?
    Well, he won’t make huge strides but if he continues to improve his plate discipline and learn his strengths and weaknesses as a hitter, he should do a little better than what he’s doing now. I’ll say 4-5 years of .280/.350/.450 production, give or take.
  5. Are there other players in recent history that have put up similar numbers?
    A quick check of Baseball-Reference’s indispensable Play Index tool (I alluded to this method in my recent Everth Cabrera article at Hardball Times) reveals that there are several such players.
  6. Which of these are best fits for Headley, in terms of position, “shape” of skills (e.g., ISO, BB/SO), etc.?
    I’m eliminating anyone with batting average higher than about .290 or lower than about .270. That weeds out guys like Al Oliver at the high end and Nick Swisher at the low end, cutting my list from 41 players to 21. I’m looking at guys in the Hal McRae/Michael Cuddyer range now.
  7. Can the list be further refined?
    Carlos Guillen, Cal Ripken, and John Valentin spent a lot of time at shortstop… Bill White retired 40 years ago, and conditions have changed since then… Ryne Sandberg was a second baseman, Cesar Cedeno stole a boatload of bases… Jeff Conine isn’t a bad comp… Neither is Millar… Or Ken Caminiti, or Michael Cuddyer.
  8. Hey, what about Overbay?
    Okay, when I made my original response and did this exercise the first time, I came up with some peak numbers for Headley in my head that generated a list which included Overbay. I have since forgotten what my original parameters were, but he can’t be far off my list (and now I see that Overbay has a career .363 OBP; my upper limit this time was .360 — I must have given Headley a few extra OBP points earlier).

Well, let’s stick with Millar and Overbay since those are the names I committed to originally, and since Overbay’s numbers are within shouting distance of my current parameters. Those guys didn’t establish themselves at the big-league level until age 27, so they aren’t perfect comps (there are no perfect comps; that would involve cloning), but they’re good enough to give us a rough idea of what Headley’s future might look like.

I hem, I haw… I’m not one of those people that are comfortable pretending to have knowledge they do not have. I use qualifiers not because I doubt myself but because I doubt everything.

So, I look at the career numbers of Millar and Overbay:

Kevin Millar and Lyle Overbay, Career
Statistics are through games of September 27, 2009, and are courtesy of Baseball-Reference.
Millar 5367 .273 .357 .451 110 169
Overbay 3815 .279 .363 .450 111 101

Millar is 37 years old, Overbay is 32. If you told me that Headley’s numbers would look something like that at those ages, I’d find it to be eminently reasonable.

How about their peaks (conveniently, both men did their best work at age 29):

Kevin Millar and Lyle Overbay, Peak
Statistics are through games of September 27, 2009, and are courtesy of Baseball-Reference.
Millar 2001 495 .314 .374 .557 140 20
Overbay 2006 640 .312 .372 .508 125 22

The batting averages seem a little high, and Millar had quite the power spike in ’01, but anything can happen in a given season (if Aaron Hill can hit 34 homers one year, then I see no reason why Headley can’t hit .314; we’re not talking about what is likely, but what is possible). Yeah, I could see Headley putting together that kind of line once or twice in his lifetime. I wouldn’t expect it on a regular basis, but he’s demonstrated — even if it was an outlier at Double-A (an outlier can happen more than once, eh?) — that he’s capable of hitting the ball with authority over relatively short periods of time (a single season is fairly long, but it is a subset of the larger unit called “career”).

Now we have some idea of what Headley’s career might look like. And we arrived at this picture by following a roughly defined process (sort of like composing a song — it has a definite structure, but leaves room for improvisation) that makes a fair amount of sense.

Could we devise a better process? Probably, and it’s good to remember this as we start asserting our new-found “knowledge” about Headley and his possible future career paths. Actually we should bear a couple things in mind:

  • There is always a better mousetrap waiting to be built
  • Despite what I have just demonstrated, Headley is not Millar or Overbay; he is Headley

Don’t become overly attached to these comps. Yeah, a lot of work went into deriving them, but this isn’t the type of exercise you do once and then let rest. What if Headley goes Aaron Hill on us next year? Then we have to feed this new information back into the system and recheck our assumptions, possibly adjusting our expectations. (For example, at various times, I have compared Adrian Gonzalez to the likes of David Dellucci, Brad Fullmer, and Chad Tracy. Last I checked, none of them ever hit 40 homers in a season.)

Ask questions. Search for answers. Wash, rinse, repeat. If you feel like you’ve got something nailed, you don’t. Go back and dig deeper. Eventually you will come to understand that you don’t know squat.

This is a good sign. It means the real fun is about to begin.