## Roto-Relevant Research: Expected Walk Rate

Sometimes, finding nothing is interesting. It’s obviously not exciting as finding *something*, but it’s not a waste of time.

For example, Blake Murphy went looking for an expected walk rate for pitchers given different plate discipline peripherals. You’d think, if given a pitcher’s percentage of pitches in the zone, and batters’ propensity to reach (and miss) for pitches outside the zone, you might be able to get a sense of their future walk rate. After all, not walking a guy is as simple as throwing the ball in the zone, and getting batters to turn a couple walks into strikes for you, right?

Guess not.

Murphy couldn’t predict walk rate any better than previous walk rates. This, after finding some success predicting strikeout rates using velocity and swinging strike rates (FanGraphs’ Michael Barr had similar findings). That might be because there are confounding factors that aren’t easy to quantify — repeatability of mechanics, consistency of release points, and the severity of movement on the pitches — or it might be because we need to weight each of the variables differently to get there. Control (and command) is a difficult thing to suss. Even guys with similar walk rates have different levels of control and command.

But this wouldn’t be a good column if I just shrugged and said — eh, we can’t know! (Who knows if it’s a good column anyway, but there’s no need to give the opposition more evidence.)

So instead, let’s point to the research Murphy and I have both done that came to the same results: first strike rate is hugely correlated with walk rate. It’s the best single peripheral to look at, and it explains almost half the variance in walk rate.  Unfortunately, zone percentage, which is a per-pitch metric, isn’t that well correlated, and we’re stuck with a per at-bat rate. That probably means that it doesn’t stabilize that quickly. On the other hand, strikeout rate has stabilized by now, and that’s a per-PA stat.

It makes a lot of sense that the first pitch strike is meaningful. Batters have put up a .269/.381/.445 line after a 1-0 count this year, and that drops to .222/.263/.343 on an 0-1 count. That’s huge. And yet it’s just one strike.

So let’s look at the pitchers whose first-strike rate is out of whack with their walk rate. If you have a bad walk rate and a good first strike rate, it stands to reason that one would regress to the other. And in cases where the walk rate diverges from a career walk rate, we might have our perfect candidates for positive regression.

Here’s our list of pitchers that have an above-average first strike rate (average is just a tick over 60% this year) but are showing a walk rate that’s worse than their career rate:

Name K/9 BB/9 ERA FIP F-Strike% BB% c BB% BB% +/-
Kris Medlen 5.98 3.44 3.44 4.79 0.691 8.50% 6.30% 2.20%
Kevin Slowey 6.57 2.01 2.55 3.40 0.690 5.50% 3.90% 1.60%
Ian Kennedy 6.62 3.58 4.83 4.78 0.653 9.40% 7.70% 1.70%
Jeremy Guthrie 4.97 2.82 2.82 5.54 0.641 7.60% 7.00% 0.60%
Madison Bumgarner 9.06 2.18 2.18 2.75 0.638 6.30% 5.70% 0.60%
Travis Wood 6.08 2.87 2.03 3.65 0.636 8.40% 7.90% 0.50%
Hiroki Kuroda 6.04 2.31 2.31 3.33 0.631 6.40% 5.70% 0.70%
Josh Beckett 8.52 3.12 5.19 4.63 0.631 7.70% 7.30% 0.40%
Justin Verlander 9.99 3.16 1.93 2.10 0.626 8.50% 7.40% 1.10%
Clay Buchholz 9.20 3.22 1.69 2.26 0.626 9.30% 9.20% 0.10%
Jake Westbrook 4.38 4.15 1.62 3.50 0.615 10.80% 7.50% 3.30%
C.J. Wilson 8.69 5.18 3.88 4.27 0.613 12.60% 10.10% 2.50%
James Shields 8.22 2.17 2.48 2.91 0.613 6.20% 5.70% 0.50%
Jason Vargas 5.96 2.98 4.03 4.01 0.613 7.60% 7.30% 0.30%
Ricky Nolasco 6.58 2.19 4.39 3.86 0.612 5.80% 5.50% 0.30%
Gio Gonzalez 9.40 4.20 4.20 3.77 0.606 11.20% 10.80% 0.40%
Justin Masterson 8.57 3.43 3.14 3.11 0.606 9.30% 9.20% 0.10%

Kris Medlen’s walk rate isn’t terrible this year — it’s basically league average — but he’s traditionally had great control. And among qualified starters, he’s got the fifth-best first strike rate in the league. So perhaps his walk rate will start regressing towards his career average. Kevin Slowey’s showing great control. Looks like it could be even better. And since he’s in the National League and calls that spacious park in Miami home, he’s an even better pickup. Ian Kennedy completes our trifecta of nodoubters, since everything lines up for him.

The rest of the list down to C.J. Wilson and Jake Westbrook are showing walk rates a bit too close to their career rates to expect much of a difference going forward. Westbrook is hurt, and Wilson is sort of meh. But you could expect better control from Wilson in the future. It’s tempting to say that Gio Gonzalez could do better, but he’s had a history of high walk rates and his first strike rate is only a tick above average.

This all started with us learning that we didn’t learn anything, but maybe we learned something after all.