Quantifying the Quality of Opponents
Many baseball analysts strive to create “context neutral” stats, so they can compare players stats, while minimizing biases. A player’s traditional stat line–batting average, home runs, OBP, slugging percentage, OPS, etc.–are a product of more than the player’s talent level and luck. For example, they are also impacted by the handedness of the pitchers they face and the ballpark in which they play . These factors can be accounted for with evaluating platoon splits or park-adjusted stats. Perhaps because it is more difficult to measure, advanced stats are seldom adjusted for the quality of opponents. Adjusting for the quality of opponents is particularly important when interpreting starting pitcher stats, since we deal with a small sample size of 25 to 35 starts per year, with a start every 5th day, regardless of who is on the schedule. Since talent tends to be clustered in certain Divisions, the unbalanced schedule produces a skewed distribution of opponents. A pitcher in the AL East is likely to be playing a different game than one in the NL East.
I use a simple measure to provide capture the quality of opponents a starting pitcher has encountered over the course of a season–the OPS of the offenses he has faced. More specifically, the OPS of the teams he has faced, against the same handed pitchers. In other words, for David Price, I take the opponents for each of his 27 starts for 2013 and use the OPS against LHP as my measure. If a RHP, faced the same teams, their opposition would be categorized based on their OPS against RHP. For example, Texas mashes LHPs to the tune of a .798 OPS, while they maintain a modest .705 OPS against RHP. On the other hand, the Cardinals hit RHP at a .753 clip, while batting only .675 against lefties. So, it’s not enough to say a pitcher faced the Cardinals or Rangers. It’s also important to distinguish his handedness.
I use a quality of opponents factor in ranking starting pitchers each year. Several years ago when I developed my SPR, I wanted to reduce the context bias that is in our everyday pitching stats, like ERA or K/9, etc. By adjusting for both the ballpark a played pitched in and the opponents he faced, we get closer to “context neutral” in evaluating how well a pitcher performed. Let’s take a look at the landscape for 2013. When we look at pitchers who faced the toughest (and weakest) competition, we find patterns. We tend to find pitchers clustered in the same division and often even on the same teams. Esmil Rogers of the Toronto Blue Jays holds the distinction of pitching against the toughest opponents in 2013. The Yankees and Astros dominate the top 20 pitchers who faced toughest opponents, for two reasons. First, they play in offensive divisions. (It is rare to see a NL pitcher near the top of the rankings, due to the lack of DH in the NL.) Second, both the Yankees and Astros were weak hitting teams in divisions with decent offensive prowess. (When was the last time we could say the Yankees were a weak hitting team? In 2013 they batted .676 against LHP and .686 against RHP, nearly 30 points below the league average.) Five Yankees populate the top 20: Ivan Nova #2, Hiroki Kuroda at #4, CC Sabathia at #11, Andy Pettitte at #12, and Phil Hughes at #19. Houston also had 5 of the top 20–Dallas Keuchel #3, Eric Bedard #6, Bud Norris #9 (including his time with the Orioles), Lucas Harrell #10 and Jordan Lyles #18.
The other end of the list–pitchers that faced the easiest competition in 2013 are dominated by NL East hurlers. Of the bottom 15, twelve are from the NL East. Mike Minor, Julio Teheran, Kris Medlen, and Tim Hudson from the Braves, along with Jordan Zimmermann, Dan Haren, and Stephen Strasburg of the Nats are all bottom 10, along with the Mets’ Matt Harvey. The Mets also have Hefner, Niese and Wheeler in the bottom 20.
So, how wide is the range in the quality of opponents? The top pitchers face offenses with OPS about 4% greater than the league average, while the bottom pitchers tend to face offenses that are 3-4% weaker. This amounts to about 25 to 30 OPS points difference. Below is a list of the top 20 (faced toughest opponents) and bottom 20 (faced weakest opponents).
An interesting analysis, as much for what it says as what it doesn’t. While observing that no NL pitchers make the Top 20, only 1 AL pitcher makes the Bottom 20, Sanchez at #20. If the middle 100 pitchers are distributed in a similar manner, doesn’t this analysis really demonstrate the difference the DH makes in the game played by the AL and NL. To truly compare all starting pitchers, doesn’t one have to minimize the distortion created by the DH?
Karl–my thinking is the opposite of your point. The DH creates a different reality, not distortion, which often is under-appreciated in evaluating pitchers. This shows the significance of the impact of the DH and why NL pitchers are consistently facing easier lineups. I remember a conversation I had with an AL GM about a certain starting pitcher. His comment to me was, “he wouldn’t work for us…he’s a classic NL pitcher”. The guy he was referring to was a 4th/5th starter type, but I can appreciate his comment. My use of this info is to adjust a pitcher’s stats to reflect the quality of the opponents (as well as the ballparks in which he pitches). It seems to be a useful mechanism to adjust for context, including the impact of DH/no DH.
“OPS of the teams he has faced, against the same handed pitchers”:
1. For what time period?
2. For what batters?
I had thought about this when I considered analyzing the quality of lineups for no-hitters. But it quickly became clear that there was more to it than simply taking batting average or on base average for perfect games. I would need to look only at the batters who actually batted in those no-hitters and weight their stat by the number of plate appearances in the game.
Then it gets really tricky. For what period? This season? What if it’s opening day or simply early in the season? For this and last season? For the last calendar year? The last three years? What about rookies and those with few appearances?
So, how do you handle those issues? Thanks for another interesting post.
This is a very interesting blog. I’m super into numbers and baseball. I’m basically studying Bill James. I’m definitely looking forward to reading more! I was wondering if you could check out my blog, Bleacher Boy.
Come and read a kid’s view on all things baseball!