Measuring a Pitcher's Strikeout Rate: Why K% Is Better than K/9

One of the leading indicators of a pitcher's success is his strikeout rate. For years, strikeout rate has been represented by strikeouts per nine innings (K/9). Many television broadcasts use this stat. Even the Fangraphs dashboard for pitchers shows K/9 as opposed to K%. Unfortunately, there's a problem with this widely used stat. K/9 can be misleading.

Even though pitchers haven't come close to averaging a complete game per appearance for fifty years, K/9 scales a pitcher's strikeout rate to nine innings. It is also based on a pitcher's innings pitched total. Here's the formula for K/9:

K/9 = 9 x (K / IP)

Because K/9 is based on a pitcher's innings pitched total, it can be warped by a pitcher's luck on balls in play. If opposing batters see more of their batted balls fall for hits against a pitcher, his Batting Average on Balls In Play (BABIP) will rise. This means the pitcher will face more batters per inning, since more men are reaching base. If a pitcher faces more batters per inning, he has more opportunities to notch strikeouts. This can warp a pitcher's K/9.

The volatility of BABIP means that the better strikeout rate is K%. This stat is more straightforward in its calculation. The formula for K% is:

K / Total Batters Faced

K% is unaffected by a pitcher's batted ball luck. A pitcher's innings total (or outs induced) doesn't come into play. Instead, K% is a straightforward calculation that simply looks at the percentage share of opposition batsmen the pitcher strikes out.

St. Louis Cardinals ace Adam Wainwright provides us with a case study in how K/9 can be misleading and why K% is a better stat for measuring a pitcher's strikeout rate. The following bar graph shows us Wainwright's K/9 for each of his big-league seasons.

We're focusing on Wainwright's K/9 over the past few seasons. By K/9, Waino's strikeouts appear to have increased slightly from 2010 to 2012 before falling off a bit in 2013. Looking at the Wagonmaker's K% for his career, we see that his K/9 is misleading.

Wainwright actually struck out a lesser share of opponent batsmen faced in 2012 than in 2010. What's more, Waino K'd a larger share of the batters he faced in 2013 than he did in 2012. K/9 suggests the opposite took place between 2010 and 2012, as well as between 2012 and 2013.

The reasons for the opposite trends in Wainwright's K/9 and K% are twofold. Part of it is caused by a change in batted-ball luck. Wainwright's BABIP surged in 2012, driving up his K/9 even though his K% actually dropped. In 2013, his BABIP dipped a bit while he also significantly reduced the number of walks he issued. The drop in BB% and BABIP contributed to a drop in Wainwright's K/9 and an increase in his K%.

Wainwright's stat line is Exhibit A in the case for using K% over K/9. By basing a pitcher's strikeout rate on innings pitched, K/9 is susceptible to changes in BABIP or BB% creating a misleading impression of a pitcher's strikeout rate. This is why K% is better than K/9 when measuring a pitcher's strikeout rate.