The Greatest of Market Inefficiencies Part II: Stats

Stats are a complex beast at the Minor League Level. In the Majors, if a guy has an .800 OPS you can say, oh, that's pretty good for a Shortstop, or that's awful for a First Baseman. It's fairly cut and dry. The minors laugh at that simplicity. An .800 OPS is really good for a Shortstop, unless he's 23 in Low A. Or, he's age appropriate in the PCL, the most notoriously hitter friendly league in baseball, and clearly .800 is bad for First, unless it's a 17 year old playing well above age in the pitcher friendly FSL.

Minor League stats require a whole lot more context than their Major League equivalents. Yet, the Majors have a lot of stats to allow for equivalency across varied park factors, league averages, run environments, etc. with OPS+, wRC, ERA+, wOBA and a slew of others. The Minors you get Avg, OBP, OPS, ERA, W-L, K/IP, K:BB and maybe at one or two sites you can get FIP, xFIP, SIERA and maybe some batted ball stats (LD%, GB%, etc.). It's virtaully impossible to find anything on pitch types, swing %, or any kind of Defensive stats.

If you want to know anything about how a player is as a Defender you essentially have Errors, Passed balls and Caught Stealing, as applicable. Even there, Anthony Hewitt had almost 2 dozen errors in the Outfield last year. How the hell do you even do that? Well, you can't find much from the stats. You'd have to look at the game recaps or go to scouting reports (as it turns out Hewitt apparently has some yips with his OF throws and has missed cut-off men, bases, infields, etc.). Even for Catcher stats which should be cut and dry, well, they aren't. Obviously, Gabriel Lino having 28 Passed Balls in 2012 is bad with no other context and it's hard to imagine what other context would mitigate it. Caught Stealing though, does require other context. It goes without any kind of contextualization needed that a Catcher throwing out 40% of Base runners is doing a stellar job (assuming we're talking full season stats and not just throwing out 2 of the 5 guys in a few games you played in)). Does that mean a guy throwing out 28% of runners is srtruggling? Probably, but it's not that cut and dry. Maybe the Pitcher is working on his delivery, so he won't pitch out of the stretch with runners on. Maybe the Pitcher has a habit of getting distracted by the pressure, so coaches (and development guys much higher up) want him to ignore the guy on first and just worry about pitching, giving the runner a large lead and free release (conversely both the guy with 40% and the guy with 28% may both be catching runners who are still developing their technique as well and get late breaks, or run on fastball counts or keep their leads too short, etc.).

So, are any stats usefull, or is the minors a barren wasteland where only scouts can be trusted? Well a few stats are pretty reliable on their own with very little context required. A High K rate by itself is always a red flag for a hitter. I don't care if the guy hits 40 Home Runs, if he strikes out 35% of the time there's a strong chance he won't be a successfull Major Leaguer. Of course, there are notable exceptions. Ryan Howard struck out in nearly 30% of AB's in the Minors. That's a big red flag, which was mitigated somewhat by a roughly 10% BB rate and prodigious power. Delmon Young, meanwhile, only struck out ~20% of the time, but only walked about 6% of the time. That brings us to point 2 in the paragraph, it's important to look at K and BB rates. People werre pretty high on Delmon Young because he was a good enough defender (yes there was a time when Young could run), a future power hitter, who didn't strike out the way Howard, Dunn, Reynolds, etc. do. Sadly, no one seems to have paid enough attention to the fact that he also doesn't Walk like those guys did. I still hold the opinion that a 30%+ K rate is likely a death knell for any prospect if it continues for more than a season or two early on. Meanwhile, any prospect who can't take a Walk is going to have a hard time sticking as an everyday player in the Majors, regardless of how infrequently they may or may not strike out. Our own Joecatz recently related a story about a prospect who was instructed to only swing at breaking balls during his Triple-A season. I've heard other stories about guys who were too quick to swing at everything being instructed to take 2 pitches before they were allowed to swing, didn't matter if it was right down the heart of the plate for pitch 1, they weren't allowed to swing.

It's the same story for Pitchers, it's all about the K:BB Ratios. No matter how good your stuff is, if you walk 5 guy per 9 innings, you're ceiling is likely reliever, at best. High walk rates are often related to guys with breaking pitches they just can't quite get control of or guys whose mechanics are inconsistent (Aumont) leading them to varied release points and balls intended for the outside corner rolling to the backstop. Of course, there are caveats here as well. A Pitcher may have a plus fastball an above average changeup and below average slider and cutter, but the team believes that even just an average cutter or slider will keep Major League hitters from just waiting for the 4 seam fastball, so the Pitcher is told to pitch solely, or largely, with the slider and cutter for several weeks. Sometimes following minor league stats is like being M.C. Escher the stairs go up, unless the guy on them is going down to the next floor, in which case the may be going up, but who knows.

One reliable stat for Pitching prospects in Ground Ball rate (GB%). This is largely the result of the type of pitches they throw. Guys with heavy downward motion often generate a higher GB%. This will typically carry to the majors and is pretty important, especially for guys without high strike out stuff.

Of course, it's an oversimplification to say only a few stats are reliable. They're all usefull in context. The problem is the context part. If you're a GM, tracking the progress of your own prospects is a little easier since you know what they've been instructed to do, but what if you're trading a star player for another team's prospects? It won't be as easy as simply pulling some numbers to determine that player's future. For that matter, even with your own players, yes you know the context of what they've been instructed, but how do you make the stats valuable, while avoiding bias? I'm not a statistician, so I don't have a great answer for that, but it seems like any correction you may add o the stats to provide context would introduce bias.

In the end, stats alone aren't an answer. Not yet, anyway. As interest grows in the Prospect world I expect we'll see more work from Baseball Prospectus, Minor League Central, Fangraphs and others to develop some level of equivalency stats or universal prospect rating scores. A few of these are out there, but we'll need a few years to see how predictive they are.

Of course, none of this addresses the elephant in the room... the MLB amatuer player draft. College stats can be projected somewhat to the Minor/Major Leagues. A guy who sucks at walking in College will likely suck at walking after he's drafted too. A guy who struggles his second time through a college lineup can be expected to face the same issue as a pro (of course the Pitcher can be taught a new pitch. It's hard to teach a wild hacker not to swing at a tasty circle change he thinks is a fastball). For High School players though all bets are off. Stats are pretty useless, as even the best High School leagues have a lot of kids who won't even be good enough to play in College. High School players are heavily projection. How good an athlete are they? How are their mechanics? Are they fast runners? Do they have a good glove and insticts for the position they're playing? More on all these factors next time.