Extreme environments

It’s a truism that numbers aren’t the whole story when evaluating amateur prospects. If we’re going to look at college stats at all, we need to treat them with care, adjusting for the inconsistency of the game at the NCAA level. Even if we limit ourselves to the top schools, it’s still an important step.

In the first five rounds of this year’s draft, 93 Division I college players were selected, representing 60 different schools. Now that those players are on the brink of entering the professional ranks, more fans than ever will be perusing their stats. I’m interested to discover which of those stats are the most misleading.

Fun with parks

In all of Division I, there are some downright silly stadiums. We tend to think of Petco and Coors as extreme parks, with park factors of 89 and 107, respectively. By contrast, Hawaii’s field has a three-year weighted park factor of 64, while both Arkansas-Pine Bluff and Northern Colorado come in over 150.

Limiting ourselves to the first 60 schools from which players were drafted does keep some of that in check. The Braves drafted RHP David Hale from Princeton in the third round. His numbers aren’t particularly impressive (4.43 ERA in seven starts, though 47 strikeouts in 40.2 innings is encouraging), but they are even less so when you consider that Princeton’s park factor is an extremely pitcher-friendly 78.

We can say much of the same about Rhode Island’s Eric Smith, who was selected by the Diamondbacks in the second round. Smith posted an ERA of 4.08 along with 56 strikeoutss and 29 walks in 70.2 innings in a park just as pitcher friendly as Hale’s. The only difference? Rhody played a more challenging schedule. We’ll get to that in a bit.

On the flip side, we have Ohio, the only one of these 60 schools with a park factor above 125. Four picks after Smith, Arizona took Marc Krauss, a left fielder from the OU Bobcats. Krauss’s .402/.521/.852 line is impressive, but the hitter-friendliness of his home park takes away a bit of the luster. His OPS at home was 350 points higher than on the road, and his home run was nearly twice as high.

Competition adjustments

In all of Division I, strength of schedule factors can be nearly as extreme as park adjustments. The SWAC, the MEAC, and some conferences with a presence in the Northeast just don’t play at the same level as the ACC and the Pac-10.

When we’re focusing on the alma maters of the top 100 college picks, most of those extreme cases fall out of the picture. However, we’re still left with some numbers that need to be taken into consideration.

The hitter who played the weakest schedule was Ball State center fielder Jeremy Hazelbaker, selected by Boston in the fourth round. Against modest competition, he posted an impressive .429/.550/.724 line. His competition adjustment is equivalent to a park factor of 110, but that only brings him in line with an average level of D1 competition. Since most of the elite college players come from a handful of stronger conferences, we’d have to adjust his numbers more aggressively to put him on a par with, say, Dustin Ackley or Grant Green.

Speaking of Ackley, the second overall pick on Tuesday, North Carolina’s schedule is at the other extreme. Their schedule was 15 percent more difficult than the average D1 competition, in part because they’ve spent the last two weeks taking on high-end competition in the NCAA tournament. The Tar Heels play in a slightly pitcher-friendly park, so while Ackley’s OPS of nearly 1.300 doesn’t need any help from me, it should be interpreted as more like 1.400.

Building an uber-multiplier

Once we have park factors and strength of schedule, it’s easy enough to combine the two for one multiplier. Let’s start by looking at the friendliest circumstances for hitters.

Ohio tops them all (and by a substantial margin) with their 127 park factor and below-average level of competition. Next up are Western Kentucky, former home of Tigers pick third baseman Wade Gaynor, and West Virginia, where the Twins found fifth rounder Tobias Streich. Both of those schools have hitter-friendly parks and roughly average levels of competition.

At the other end of the spectrum for offense, we have an odd trio, including Texas, Miami, Fla., and Rhode Island. Texas and Miami play in pitcher-friendly parks against elite competition; Rhode Island plays in a really pitcher-friendly park against middle-of-the-pack opposition.

Thus, we can look at Texas’ Brandon Belt and Miami’s Jason Hagerty as if they played in sub-90 hitting environments. As for Rhode Island, their pitching was viewed as more pro-ready; no Rams hitter was selected until the Rays took Dan Rhault with the 799th overall pick.

Pitching extremes

Of course, we can look for the same friendly and unfriendly environments for pitchers as well.

Once again, we return to Rhode Island and Princeton. Neither plays a difficult schedule, so the extreme park factors I noted above made those two of three most favorable environments for pitchers. Rounding out the top three is Kennesaw State with their very pitcher-friendly park and middling competition.

Using Recurrent Neural Networks to Predict Player Performance
Technology is rapidly advancing possibilities in decision-making.

That’s a notable finding, especially compared to most of the third and fourth round names I’ve been mentioning so far in this article. Pitchers Chad Jenkins and Kyle Heckathorn were both chosen from Kennesaw State within the first 50 picks, and two more were chosen after the first day.

The three toughest parks/schedules for pitchers were those of Mississippi, Tennessee, and Missouri. From Mississippi, the Cardinals selected hurler Scott Bittle in the fourth round while the Pirates took Nate Baker twenty picks later. Between hitters’ parks and challenging schedules, those two guys should be treated as if they pitched in stadiums with park factors above 120.

We can feel almost as much pity for Tennessee’s Bryan Morgado (third round pick of the White Sox) the 22nd overall selection, Mizzou’s Kyle Gibson. Gibson’s numbers are the pitching equivalent of Ackley’s: His OPS against was about .625, but we can treat it more like .550.

Context in context

I don’t mention these numbers to criticize or applaud any of the picks. I highly doubt that when Boston took Hazelbaker or Arizona took Krauss, anyone in the draft room said “Ooh! Shiny power numbers!” and pulled the trigger. Some teams may not have strength of schedule spreadsheets handy, but just about everybody making these decisions knows that Princeton plays an easy schedule and North Carolina plays a tough one.

Most of all, these are just one more factor to add to our toolbox when evaluating draft picks and considering how they will perform this summer and beyond. While someone like Bittle or Belt might be more ready to take on the minors than we think, all those Kennesaw State pitchers may well take longer to adjust to pro competition.

The numbers aren’t the whole story, but they can be a useful component when we know what they can and can’t tell us.

References & Resources
Craig Burley has previously examined park and strength of schedule adjustments for college players for The Hardball Times in his articles “Adjusted NCAA statistics: introduction and top 100s,” “2004 NCAA adjusted statistics” and “NCAA adjusted statistics top 250 hitters.”

Print This Post

Comments are closed.