August 4, 2010

Seidnotes

Let's Go for a Free Pass

by Eric Seidman

Printer-friendly

Contact Author

I've been thinking an awful lot about walks lately—not the kind I try to take each night before bed, but rather the kind that Cliff Lee now avoids with startling regularity. My last two Seidnotes columns focused on his fantastic season in an attempt to deduce whether or not anyone ever matched his potentially historic pace. Additionally, I used his numbers to illustrate the differences between the more common strikeout-to-walk ratio and the strikeout-minus-walk differential. Today, I frame his walk-averse campaign in a slightly different light. Entering his most recent complete-game loss, Lee issued seven walks while surrendering nine home runs.

Yes, the man had walked fewer batters than he had allowed home runs! An out-of-character, two-walk performance on Sunday tied the numbers, but I began to wonder how rare it is for dingers to exceed walks. With that in mind, my goals today are to explore this very phenomenon, and to discuss walks and walk rates on a very basic level, as the numbers are used very frequently, yet leaders in the respective categories are not exactly common knowledge.

First, how often has a pitcher with 150 or more innings in a season allowed fewer free passes than McGwire Specials? Dating back to 1974—and excluding the 2010 season for now—a measly 14 pitcher-seasons met the criteria. As we might expect, some of the names are repeated, making the feat all the more rare. In order of fewest walks allowed, here are Lee’s potential contemporaries:

What has made Lee’s year so outstanding is that, in addition to virtually never walking anyone, he isn’t even being hit hard. Well, what’s the least amount of home runs a pitcher has surrendered in a season with the same innings stipulation? An initial run of the query explained that almost all of the leaders in this area did so in past eras. Further restricting the sample to seasons in the wild-card era, a four-way tie for first place emerges between Chris Carpenter’s 2009, Clayton Kershaw’s 2009, Pedro Martinez’s 2003, and Andy Pettitte’s 1997. All four of those pitchers allowed a mere seven home runs over the course of the season; however, none of them are anywhere near Lee’s walk rate. Nobody in the wild-card era who has pitched as much as Lee will and gone on to finish the year with fewer than 20 walks and fewer than 20 home runs.

Then again, these are raw totals, which can be rather wonky when trying to compare players or solve problems like this due to playing time. A pitcher with 50 walks in 250 innings has more raw walks than one with 30 walks in 130 innings, but the rate of the former is likely to be superior.

With that in mind, let’s rephrase the question: Has anyone ever finished a season with a walk rate (BB/PA) and home-run rate (HR/PA) in the same vicinity of Lee’s? To answer, we must first determine Lee’s rates. Since he has allowed both nine home runs and walks, and faced 572 batters, his rates are each .016. Here are the lowest walk rates and home-run rates since 1974:

Moving on, I want to touch a little on the differences between rates and raw statistics, and between rates themselves, as I am asked questions along these lines quite frequently. Rates and raw tallies are the two basic building blocks of baseball statistics. When asked Lee's penchant for giving up home runs, we can answer with nine, which is the literal amount of balls that have left the yard, or .016, which is the rate at which batters he faces hit home runs. Both have their place and neither is technically wrong. What matters is the level of accuracy. For instance, if I’m comparing a group of pitchers with between 200-210 innings pitched, raw numbers are likely fine as there won’t be much in the way of playing time issues.

If, however, I’m comparing someone with 322 1/3 innings to another pitcher with 187 2/3 innings, the raw numbers will lead to inaccurate conclusions. The higher-innings hurler is likely to have larger raw tallies by virtue of being out on the mound more often. That does not make him better or worse. Perhaps the latter pitcher struck out 14 percent of his opponents while the former whiffed only 10 percent; the former is much more likely to have a larger tally in spite of the less effective rate.

But there are numerous types of rates as well, which muddies the waters a bit. The more commonly used rates are shown on a per-nine basis. K/9 or BB/9 are the most frequently quoted “offenders” and the numbers essentially attempt to tell us the raw total that would occur if the pitcher were to have thrown a complete game. See the problem there? First, we’re using a rate to give us a raw number, and it doesn’t actually tell us anything characteristic of the pitcher. Knowing a pitcher has a 7.7 K/9 is somewhat helpful, but I have no idea how that translates to batters faced, which is a more effective baseline.

Don’t believe me? There are multiple ways to skin a cat, just like per-nine rates can be arrived at in different fashions. Consider these two innings:

1. Strikeout, Double, Single, Strikeout, Home Run, Walk, Strikeout
2. Strikeout, Single, Strikeout, Strikeout

 Both innings featured three strikeouts, meaning that, on a per-nine basis, each achieved a 27.00 K/9 for the frame. However, the first inning is obviously far worse than the second, as three runs scored and a walk was issued. The per-nine basis doesn’t care. Using percentages, the first inning has an SO/PA of .429, while the second checks in at .750. Choosing between raw tallies and rates can be a bit confusing, but choosing between the types of rates can be maddening. However, when trying to show how Cliff Lee is in the midst of a season for the books, or any other similar exercise, it’s important to use the right numbers to make the point, or else you risk not making the right point at all.  

Eric Seidman is an author of Baseball Prospectus. 
Click here to see Eric's other articles. You can contact Eric by clicking here