This introduction is a setup. Don’t fall for it. I’m going to present you with two stat lines and ask you to silently compare them. Your job is going to be to determine which player had the better season at the plate. Remember, it’s a trick.
- Player A: 697 PA, .372/.463/.698, .476 wOBA, 42 HR, 59 2B, 103 BB, 61 K
- Player B: 716 PA, .323/.432/.557, .423 wOBA, 27 HR, 39 2B, 110 BB, 136 K
If I hadn’t primed you, it would be hard to suggest anything other than that Player A had the better season. He’s leading everything, except for a slight disadvantage in walk rate. Player A had the better season, right? It’s obvious. Even though I told you it was a trick, you’re still struggling to find a way to argue the opposing side. I’m telling you that Player B actually had the better season, but that’s because I have more information. I know a couple of important pieces of information that you don’t have and it makes a world of difference.
I know the year in which each season took place and the park in which each hitter played. Those are two really important bits of information when evaluating a baseball player and without them you’re left to assume that Player A had the better season, which simply isn’t the case. Played A, if you haven’t figured out, is Todd Helton in 2000 and Played B is Mike Trout in 2013.
It’s an extreme example used for illustrative purposes. More runs were scored per plate appearance in 2000 than in any season since 1930.In 2013, that number was dramatically lower. In 2000, teams scored 5.14 runs per game. In 2013, they scored 4.17 runs per game. That’s a huge shift in the run environment. Put another way, the value of adding a run in 2013 was substantially higher than adding a run in 2000.
If you want to compare players across time, you need a statistic that accounts for the differences in run scoring. Maybe that’s an easy thing to understand if you want to compare Ted Williams and Joey Votto, but it can actually make a pretty big difference over just a decade or two. The early years of Derek Jeter‘s career, for instance, were much different than the later years of his career when it comes to league-wide run scoring.
There’s another fun little quirk about baseball that doesn’t show up in the other major American sports; The playing fields are different in all 30 home parks. Not only that, but because baseball is played outdoors and is heavily dependent on the flight of the ball, different altitudes and climates play a big role. In other words, it would be harder to hit a home run in San Diego than in Colorado even if the pitcher threw you identical pitches in both cities.
The bottom line is that identical results in separate contexts do not necessarily indicate equal performance, skill, or talent. This is a fundamental aspect of sabermetrics and advanced analysis in baseball. You need to do everything you can to control for context or you may wind up with the wrong answers to your questions about the game.
At FanGraphs we have many statistics that attempt to do this. One of the most visible and notable ones is Weighted Runs Created Plus (wRC+), which is an all-encompassing hitting statistic that weights each offensive action properly (like wOBA), but also adjusts performance based on park effects and league average.
So while Todd Helton’s raw production looks more impressive than Trout’s, Helton actually had a 162 wRC+ in 2000 while Trout had a 176 wRC+ in 2013. You can read in detail about wRC+ here (including the precise formula), but it tells you how much better or worse a player was than league average, which is set to 100 during each season. It is also park adjusted, so not only can you compare production across time, but you can compare players from vastly different home parks.
For example, if a player has a 120 wRC+, it means their offensive production is 20 percentage points better than league average during that year after controlling for their home park. Sure, Todd Helton out slugged Trout by .140, but he did so in a park that was more conducive to offense and during an era in which the overall offensive baseline was much higher.
It’s very simple to understand, but extremely important nonetheless. Sabermetrics has many goals and important principles, but one of the most critical is the effort to control for the context of events. If we want to make comparisons among players from vastly different eras and evaluate their relative contributions, we need to strip out anything and everything outside of their control that impacted the raw output we observed.
If you’ve been reading sites like FanGraphs for a long time, I’m not telling you anything you don’t already know, but for those who are new to advanced statistics in baseball, controlling for context is a cornerstone of the entire enterprise. We use statistics in an effort to communicate the best possible representation of the truth. Once you’ve learned the basics of wOBA, making the jump to wRC+ is relatively easy and it provides a more accurate picture of what a player is doing at the plate than any other rate stat we have to offer.
Have questions about wRC+ or controlling for context? Ask in the comments!
Print This Post