Context Adjustments: Park Factors
The concept of “park factors” is thrown around frequently by fantasy writers and players, but rarely is it discussed intelligently or utilized properly. More often than not, talk will take the form of generalities, such as “Coors is a good park for hitters” or “PETCO is where home runs go to die.” Sometimes you’ll see someone cite a precise park factor (“U.S. Cellular inflated home runs by 54% last year”), but this can be incredibly misleading without calculating the factors properly and without considering context.
One of the most important points we need to consider when using park factors in our decision-making is that they are usually very unstable. It may seem counterintuitive that the way a park plays over an entire season is an unreliable indicator of how the park will play in the future, but this is the truth. When looking at park factors, at a bare minimum, we must use at least three years. If you hear someone cite a park factor that uses less than three years of data, you must disregard it out of hand or else risk drawing incorrect conclusions.
Ideally, you’d combine those three (or more) years of data with a regression to the mean component. This is because park factors can be so unstable that simply using extra data isn’t enough. Because we’re using a finite amount of data, there will still be some amount of random variation present, which we use regression to the mean to account for (if you’re unfamiliar with the concept of regression to the mean, there’s a great primer on it here).
If you want to go a step further, instead of simply comparing every team’s total home and road numbers like most park factors do, you can match up every team with every other team and compare, for example, how Team A did in their home park while playing against Team B versus how Team A played in Team B’s home park. Do this for every combination of teams, and you arrive at park factors that give us the following parks as the top five for hitting home runs:
And these as our bottom five home runs parks:
Ok, thanks Derek, are we finished here? Nope, not quite. In most articles about park factors, this is where the discussion ends. We talk about home runs and that’s it. But that’s not where park factors end. Not even close. You may not know but, but there are actually some very real and very significant effects that parks can have on other stats, like strikeouts and walks. At first, you may be thinking, “How can a park have any impact on strikeouts? That’s ridiculous!” Hear me out, though, because these effects are absolutely, without question, real.
While we may not be able to pinpoint the precise reason why park factors affect things like strikeouts and walks, we can absolutely observe them and tests show them to be significant. Plus, we have some pretty good guesses as to why we can observe them. The most likely reasons are atmospheric effects, park elevation, and humidity.
Accounting for multiple years of data and regressing to the mean, we get these as our top five strikeout parks:
And our bottom five:
And for the sake of being thorough, here are our top and bottom five for walks:
Hopefully this article has educated you a little bit about park factors and maybe even pointed some things out that you may not have known. When considering who to draft for your daily team, taking into account the park he’ll be playing in that day can be extremely important, especially for pitchers. The difference between Pitcher A starting in Florida and Pitcher A starting in Colorado can be as many as two strikeouts. That’s incredibly significant in the game we play, especially when you consider that these strikeouts will also have a positive impact on ERA, WHIP, Wins, and IP. So if you’re playing in the $25K MLB Grand Slam league or in the THT Fantasy/Fanduel MLB Salary Cap 35k contest (where you compete against me for cash; the league is free to enter), maybe you want to consider whoever’s pitching for (or against) the Marlins. Mike Pelfrey, anyone?
Derek Carty’s work can also be found at The Hardball Times Fantasy and DerekCarty.com. He is the youngest champion in the history of LABR, the longest-running experts league in existence, and is a graduate of the MLB Scouting Bureau’s Scout Development Program (aka Scout School). He welcomes questions via e-mail.