Strikeout Proficiency (Part Two)
In A New Way to Measure Strikeout Proficiency yesterday, I introduced the concept of strikeouts per pitch (or 100 pitches) and proclaimed that "this stat just might be the best way to measure pitcher dominance, if not overall performance."
Well, as it turns out, strikeouts per pitch (K/P) explains runs allowed better than strikeouts per inning or strikeouts per batter faced. The stat measures dominance and efficiency, and I strongly believe that it is "the single greatest Defense Independent Pitching Stat out there."
Today's article is focused on the technical aspects of this argument. I refrained from getting overly technical yesterday because I wanted to share the idea without overburdening the reader with statistical terms such as correlation coefficients.
Among pitchers with 162 or more innings, I compared the correlations of three strikeout measures (by inning, batters faced, and pitches) against Earned Run Average (ERA), Runs Per Game (R/G), Component ERA (ERC), Fielding Independent Pitching (FIP), and Defense Independent Pitching Stats (DIPS).
ERC, FIP, and DIPS are metrics that estimate what a pitcher's ERA should have been, based on variables within his control (such as K, BB, HBP, and HR).
In addition to ERA, R/G, ERC, FIP, and DIPS, I am also going to use K/IP for strikeouts per inning, K/BF for strikeouts per batter faced, and K/P for strikeouts per pitch.
The correlation coefficient measures the strength of a linear relationship between two variables. The correlation coefficient is always between -1 and +1. The closer the correlation is to +/-1, the closer to a perfect linear relationship.
All of the correlations in the tables below are negative. A negative correlation is evidence of a general tendency that large values of "X" are associated with small values of "Y" and small values of "X" are associated with large values of "Y". Think of "X" as strikeouts and "Y" as runs.
Correlation Coefficients Matrix For 2005
ERA R/G ERC FIP DIPS K/IP -0.414 -0.436 -0.516 -0.615 -0.659 K/BF -0.503 -0.528 -0.614 -0.681 -0.720 K/P -0.534 -0.557 -0.656 -0.717 -0.755
As detailed above, K/P has the highest correlation in each of the five run measures. K/BF has the second-highest correlation and K/IP has the lowest correlation. In any other words, K/P > K/BF > K/IP.
Correlation Coefficients Matrix For 2004
ERA R/G FIP ERC DIPS K/IP -0.520 -0.527 -0.621 -0.637 -0.655 K/BF -0.581 -0.587 -0.673 -0.701 -0.704 K/P -0.592 -0.595 -0.703 -0.718 -0.736Once again, K/P has the highest correlation in each of the five run measures. K/BF has the second-highest correlation and K/IP has the lowest correlation. Just like in 2005, K/P > K/BF > K/IP. Conversely, DIPS has the highest correlation to each of the three strikeout measures. ERC and FIP switch spots with the former having the second-strongest correlation and the latter the third. R/G ranks fourth and ERA has the lowest correlation. That is, DIPS > ERC > FIP > R/G > ERA.
Despite claims to the contrary by some readers around the baseball blogosphere (including our site), the evidence is indisputable. K/P is not only a better measure of strikeout proficiency than K/IP and K/BF, but it has a stronger correlation to runs allowed than these other measures.
For whatever reason, many people are slow to accept new ideas. No matter how much proof one provides, there will always be naysayers who don't want to embrace the truth. But that is OK with me. You see, I'm not a member of the Flat Earth Society.