Beware of the Wilmington and Lexington effect on Royals prospects

Scouting the stat line is a very dangerous thing to do to judge a prospect and their future. It’s true that all good hitting prospects hit, and those that don’t, end up no longer being good hitting prospects (the same for pitching). I’m someone who is very involved in the cross section of statistics and scouting reports for prospects in regards to their current/future standings. But, you need both to really gauge a prospect. Not all good hitting performances means a player is a good hitting prospect just as a bad hitting performance doesn’t mean a player is a bad hitting prospect. Tools don’t universally make a prospect either because eventually they have to turn tools into performance.

One of the many dangers in scouting the stat line is the wackiness of minor league parks. For the majors, we have pretty good park factors that we can adjust a hitters performance, boosting it for hard to hit parks (like San Francisco) and discounting it for performance enhancing environments (like Coors Field). From those park factors we have metrics like wRC+ and OPS+ that give park and league adjustments.

For the minors, we don’t quite have that data, which can lead to some skewed numbers that need to be judged with proper modifications. MiLB parks offer a wide array of different park styles, on both extremes of the offensive environment. We have metric available that can adjust to the league quality (wRC+) but typically harder to find one that adjusts at park level.

This park level adjustment is particularly relevant for Royals minor league hitters and pitchers in the lower levels. When in Lexington (the Royals A-Ball affiliate), hitters live in an extremely good hitting environment (and the opposite for pitching). If/when they move up the ladder to A+ Wilmington, they go to both a dreaded home park (notoriously so for power) and a rained soaked league (Carolina League).

We can easily look at some numbers to help explain this. For instance, here are the Royals best hitting prospects in Lexington:

Seuly Matias

Home: 237/.321/.711 14 HR

Away: .228/.279/.532 6 HR

Matias has hit for very strong power on both sides of the schedule, but it’s notably much higher when at home in Lexington (as is the OBP).

MJ Melendez

Home: .292/.364/.629 6 HR

Away: .220/.265/.418 3 HR

Nick Pratto

Home: .301/.352/.496

Away: .183/.218/.269

The most extreme case here is Pratto, whose overall batting line this year .244/.289/.387 (which isn’t very good to begin with) is being propped up by his numbers in Lexington. When he is at home, he puts up near elite numbers. When he moves away from the friendly confines of Whitaker Bank Ballpark, he looks terrible.

Now let’s move up the ladder to Wilmington, the opposite extreme.

Khalil Lee:

Home: .255/.415/.402

Away: .316/.438/.490

Lee has had strong OBPs anywhere he’s at, and that’s due to his walk rate (which is park agnostic really). But when he’s left Frawley he’s hit for more power (via isolated slugging), as Wilmington’s home park is hell for power.

Chase Vallot

Home: .109/.315/.218

Away: .140/.290/.421

Now Vallot’s line is rough this year no matter the location of where the game is held, but his power has been decimated when he is at home (his OBP is also a function of his great walk rates). Vallot also played in Wilmington last year, so let’s see if the effect still held strong then too.

Home: .220/.388/.362

Away: .240/.374/.500

Almost a .120 point difference last year in isolated slugging depending on which side of the inning Vallot’s team batted during.

I mentioned that FanGraphs wRC+ metric adjusts for the league the player plays in, which is how Seuly Matias and Khalil Lee have nearly identical wRC+ results (153 vs 151, respectively - also a function of Matias’ power approach vs Lee’s OBP). It doesn’t however adjust for the park, which is more important than the league adjustment. There is only one site I know of that adjusts for both, and that’s StatCorner, a sabermetric site owned and operated as a hobby by a single person, Matthew Carruth.

Because Statcorner introduces park adjustments (called wOBA+ - similar enough concept to wRC+), we can get a different (or better) look at what a park adjustment can do:

*Note that while wRC+ and wOBA+ both scale a their respective lines to a league number of 100, the underlying figures are different so the two can’t be compared with each other from a +/- differential standpoint.

Through wRC+, Lee and Matias look basically equal, but when you add in a park adjustment (a boost for Lee and a discount for Matias), Lee comes out looking better. A same similar adjustment leads to Vallot looking equal to or better than Pratto. Like I said, Wilmington is hell for hitters.

Ideally I’d like to see FanGraphs wRC+ (I prefer it to wOBA+) adjust for minor league parks (the park data is relatively easy to find and then scale) so as to compare the difference between league + park and just league adjusted.

This doesn’t just only effect hitters, but Royals pitchers too. One classic example is Sean Manaea in 2014. At home in Wilmington, he had a 2.09 ERA. On the road, Manaea had a 4.44 ERA. This is another reason why you should use FIP over ERA mostly while making a mental note of the home runs allowed (ironically Manaea’s ERA was identical to his FIP in 2014: 3.11)

Surely there are teams out there that have this sort of data internally, looking to acquire players whose line looks artificially bad because of their home park, and trade players whose numbers look artificially good because of it.

This is all just a lesson in being cautious at looking at just raw minor league numbers. MiLB ballparks are just as goofy as some of the promotions they host. If we make league and park adjustments to them, we can get completely different (and more accurate in a sense) numbers.

*You should consider donating to Carruth’s site as a greater initiative to support and promote sabermetrics and the adoption of them to baseballs wider audience (I have no incentive or remunerations for this).