The formula for park factors assumes that fielders and pitchers are of major league caliber or are, at the very least, replacement level and major league ready. The difference with the minor leagues is that, well, it’s the minor leagues. The levels between the instructional leagues to High-A are geared to be more developmental than anything else.
Pitchers in the lower levels are more hittable because they’re working on developing new pitches and don’t necessarily attack the zone the same way a AAA pitcher would. Hitters can expose the missed targets and connect the bat to the ball at a higher rate. This means more bloop hits, more doubles, more home runs.
Fielders also aren’t as polished in the minor leagues. Some were promoted due to their ability to hit the ball well, but still need to work on their footwork or their glove. There’s also the DH-type of player who sometimes has to field, and since the minor leagues use a DH regardless of their parent organization’s league, they don’t get the same experience fielding as the every day defenders.
By using a park factor formula that only uses runs scored (RS) and runs allowed (RA) to analyze minor league parks, it’s allowing inflated numbers to be included and doesn’t reflect the skill level of the players and the fact that lower level leagues focus on a player’s development more than the results.
One way to neutralize inflated numbers in a park factor formula could be to incorporate FIP, or elements of it. In the minor leagues, FIP will show if a pitcher’s outperforming the defense behind him, as it does in the major leagues, but there’s more weight to it in the minors because of the skill level. This can be flawed, however, because of the home run numbers and the fact that many parks — and leagues for that matter — are considered launching pads. But it will give an idea of what kind of defense plays at a home park. It’ll also give an idea of where a pitcher is with working on command. More strikeouts can show that he can control the ball better/throws more swing through stuff, which can contribute to a better FIP number.
BABIP is another, because hitters often exploit the development of pitchers and bad defense for higher contact rates. Regression is not necessarily the same as it is at the major league level because MiLB hitter regression is often because of adjustments made at the plate. Neutralizing BABIP or elements of it into park factors could also help with minor league park factors.
Then there’s ISO. Power is not really measured correctly due to all factors above. All it shows in the minor leagues, especially at the lower levels, that hitters are doing what they’re expected to do, especially if it’s a hitter’s league. If there’s a park factor formula that can neutralize inflated numbers, then stats such as ISO can be neutralized in order to get an accurate sense of how a player is doing in the minor leagues.
In order to do this, a new formula must be created that includes all the variables and being able to recognize which numbers have more weight than others. Sorting out home splits and road splits for a team is also another thing, which isn’t easily available online (as far as I know; I’ve checked all the usual suspects for minor league gamelogs and stats).
If there were a way to take the traditional park factor formula and modify it to include variables that are, or involve elements of, BABIP and such, it could be a start to having statistics that can more properly reflect the constant development in an organization’s farm system.