Seidnotes: K/BB Ratio Redux

Well, last week was rather eventful, wasn’t it? I woke up to the tune of approximately 14 text messages from friends and family asking for my opinion of a supposed Cliff Lee-to-the-Yankees deal, only to find out after responding that no such deal had been consummated. Then, as I drove home from work, regularly scheduled Philly sports talk radio programming interrupted its continuous “If Lee was on the Phils" drivel to inform the listening audience that the lefty du jour had been traded to the Texas Rangers by the Seattle Mariners. It was a foregone conclusion that Lee would be traded around this time, and his very impressive numbers to date only increased the desire for his services from a bevy of potential suitors. This marks the fourth team Lee has pitched for since last summer—I suppose that’s what happens when a not-great pitcher signs a team-friendly deal and then becomes awesome—and while he doesn’t make the Rangers favorites to vie for the American League championship by any stretch, he brings a unique command of the strike zone that certainly aids their chances of playing deep into October.

Lee’s debut for the Rangers drew mixed reviews. On one hand, he pitched a complete game and continued to show the type of stuff that made his roster presence so important in the first place. On the other hand, he allowed six runs to the lowly Orioles. In the end, he didn’t walk a batter and struck out two, increasing his seasonal totals to 91 strikeouts and 6 walks, a rather preposterous 15.17 K/BB ratio. I say preposterous because that mark resembles one we might see after a week or two of the season, when small samples drastically skew rate statistics. Seeing as it is the middle of July, even though Lee missed the first month of the season with a abdominal strain, he has pitched 112 2/3 innings, hardly a sample small enough to deem his ratio faux-worthy.

In last week’s Seidnotes column, I discussed the K/BB ratio in general with Lee as the main character, as it is one of those metrics used so commonly that it is easy to forget why we use it in the first place or from where it was derived. In the process of examining the statistic, I offered single-season leaders with a 150 IP minimum to see where Lee’s 14.83 K/BB ratio at the time would rank. With the aforementioned filter, the highest K/BB belonged to Bret Saberhagen’s strike-shortened 1994 season, when he whiffed 143 and walked just 13 for the Mets, producing a 11.00. If intentional walks are removed, the honor goes to Greg Maddux’s 1997 campaign, in which he finished with a gaudy 12.64 K/BB ratio.

The analysis could be construed as a bit misleading, however, because I was comparing numbers compiled in half a season’s worth of work to those amassed over an entire season. While it is sexy to suggest that Lee could finish the season with the highest single-season K/BB ratio in the modern era, a more appropriate line of inquiry would compare his current rate to that of pitchers in the past through a similar span, be that innings pitched or date. For instance, how does Lee’s 15.17 K/BB through 112 2/3 innings compare to other pitchers in the past through their first 110-115 innings? An analysis like that would provide some telling information as far as whether or not any pitchers even fit this description and, if so, how they fared over the remainder of the season. That way, if nobody surfaces with a K/BB as high at that point in their season, we can more effectively conclude that Lee is on an historic pace.

To that end, I set up a table in my database that calculates the cumulative totals of several statistics after each game played. The benefit of using a tool like this instead of blindly computing numbers after a predetermined number of innings pitched is that we can compare pitchers after games have been played to one another as opposed to comparing Lee’s 112 2/3 innings and the associated K/BB mark to someone who reached 112 2/3 innings in the fourth inning of a game in which he pitched seven frames.

The following restrictions were used to come up with the potential comparables below:

1. Innings pitched from 108-115.
2. The pitcher had to be close to the midpoint of the season. For instance, Bret Saberhagen reached 110 innings in 1999 towards the end of the year, one in which he threw just 119 innings. How does his post-inclusion numbers help us here at all?
3. If a pitcher showed up twice on the list, the higher totals were used.
4. The earliest season was 1974.

Excluding Lee, here are the highest K/BB numbers through a similar point in the season:

Not too many surprises here aside from Wells, whose 2003 season did not surface on either of the tables last week. After his first 113 2/3 innings and an uber-impressive K/BB, Wells produced a 43/16 K/BB ratio—2.69 to be exact—over his final 99 1/3 to finish the season at 5.05. In fact, the table below shows the same 10 pitchers featured above but with their numbers after the point already shown:

Aside from Wells and perhaps the 2005 season from Brad Radke, everyone else posted elite ratios over the second half of the season, even if the latter portion of the split-half results were not as appealing as those from the former. The average in the second half was 6.23, down from the 9.04 in the first half but still more than respectable. Lee, at 15.17, has a higher K/BB ratio than anyone over the last 36 years as of this point in the season, and realistically that is the only period of time that matters for a study along these lines. The latter portion numbers for these 10 pitchers certainly suggest a dropoff in the rate is on the horizon, but not nearly one so severe so as to render moot the first-half numbers.

If we are conservative and project Lee to make another 13 starts at around seven innings per start, giving him 90 frames the rest of the way, even if he fell to the 2.69 rate Wells posted down the stretch in 2003, he would finish with a 9.62 rate. If he posted a 6.50 rate down the stretch he would finish at 11.31, setting the single-season record. In other words, nobody has done what he is currently doing, and he would really have to fall off of a cliff—get it?!—to not set the single-season record for K/BB ratio for those with 150 or more innings pitched.