Knowing When To Surrender the Walk: Run Expectancy and The Free Pass

Today I want to begin a new winter project of mine that involves examining which pitchers succeed in sequencing their component peripherals. Using Retrosheet's event files and Tom Tango's Run Expectancy table, I've measured the change in RE for each plate appearance from 1993-2011. I want to focus first on non-intentional walks this morning.

For each pitcher from 1993-2011 with at least 5,000 Batters Faced, I've added up the average growth in run expectancy per NIBB. The average increase in RE for an NIBB over that period was .323 runs, which jives with Tango's findings here. There were 175 pitchers in the sample, with a standard deviation of .013.

The pitcher who gave up the least damaging change in RunEx with at least 5,000 BFP was Rick Reed at .258 RE/NIBB, a full .065 better than the league in that time. Reed scores just ahead of Gregg Maddux at #2 with a change of RE .041 better than the league.

Here is the top ten:

I always prefer the eye test when viewing a measurement for the first time, and I'll admit I was extremely happy to see Greg Maddux's name show up as #2. Of course, I would have preferred #1, but it's important to remember that Maddux maintained this walk-efficiency over the course of a much larger sample of over 15,000 BF compared to that of Reed, who just edges out the minimum requirement.

Reed wasn't nearly the type of dominant pitcher that Maddux was, but I don't suppose you have to be particularly elite to be efficient with your free passes.

Aside from Maddux, we have two additional elite 'command type' pitchers in Schilling and Sheets, who we could easily imagine exhibiting situational control over their walks. And Mike Mussina at #10 has always carried the label of being especially 'cerebral' on the mound. Not to mention the unearthly Mariano Rivera at #7.

So far so good, but let's move on to the 10 worst just to be sure we're on the right track.

This is certainly not the worst list of pitchers I have ever seen. In fact a number of them have had some excellent seasons over the course of their careers-- Kerry Wood especially. I don't think anyone would mistake Ryan Dempster or Wandy Rodriguez for a pair of baseball's best pitchers, but the two certainly managed to put together a handful of all-star-caliber seasons over the years.

But, it should be noted that the "top ten worst" were not nearly as bad as the "top ten best" were good, as most of the players from the former table had RE/BB's just barely outside the standard deviation. This is to be expected with a decent sample size-- if a player truly had a problem walking players when he wasn't supposed to, it is unlikely he would have survived very long in the majors.

I don't expect there to be a 100% correlation between pitcher performance and sequencing skills. Instead, I believe we are looking for a particular type of pitcher in this and the up-coming RunEx posts. I'm looking to find a type of pitcher that ultimately prevents runs based on sequencing more than the average pitcher. So, this is not to say that Ryan Dempster, Wandy Rodriguez or any pitcher demonstrating inefficient sequencing based on run expectancy are not effective, they are just not utilizing that particular skill in their arsenal.

Final Note

I'll admit to being partially inspired to run this data by recent internet discussions about Tom Glavine and the ongoing pursuit to uncover the nature of his strange ERA/FIP discrepancy. Yet, interestingly, Glavine placed 72nd of the 175 pitchers sampled, just .065 better than average. It is encouraging, however, that his RunEx on Home Runs Allowed was significantly better than league average (18th overall, min 5000 BFP), a topic which we will look at next week.

All data from retrosheet, 1993-2011, includes post-season.

Follow @JDGentile on twitter.