Walk Rate Trends Under Mike Maddux

Not long after the season ended, I wrote an article here at VEB that tried to evaluate how Mike Maddux was doing as a pitching coach. The results were pretty positive. The Cardinals have displayed the ability to consistently outperform their computer projections in multiple categories – an improvement that we can somewhat credit to coaching. Ks rose. ERA vs. FIP improve. The pitching staff has performed better.

This came despite a walk rate that pretty consistently underperformed. The numbers and the strategy they implied were obvious enough that I found myself quipping (sarcastically), “Maddux don’t care about no walks, though!”

Now a quarter of the way through the 2021 season, I am wondering how much truth was in that throw-away statement. The Cardinals are walking batters. A LOT of batters. A near-historic rate of walked batters. I did a custom search at Fangraphs of stats relevant to walk rate. Here are the results:

The Cardinals are pretty much last across the board. BB% - last. BB/9 – last. BB+ – walk rate rated to league average – last. ERA – not last!

Yes, even though the Cardinals are reaching All-Star levels of walking ineptitude, the pitching staff is doing ok in terms of runs allowed.

The Cardinals under Mike Maddux have been masters of outperforming their peripherals. Some of that’s by design, but even Maddux has to be concerned about the current rate of free passes. Doesn’t he?

My question today is whether the club’s propensity to give batters free passes is a product of roster construction and strategy – and likely to continue – or just poor execution by individuals and small sample size volatility – and likely to change.

Let’s talk method. Below are the Cardinals’ top-10 pitchers by IP for every season since 2018. I then compare actual performance to expected (ZiPS) performance for walk stats — BB% or BB/9 (based on what ZiPS used that season) — calculating the percentage of improvement or decline.

(Note that the colors of these images are reversed from what is used at Baseball Savant – blue is good, red is bad; dark blue is very good, dark red is very bad.)

Since Maddux was installed, Cardinals pitchers have seen a relatively consistent increase in walk rates versus expected. In 2018, only three of the top 10 pitchers by innings improved their walk rates over expected. One of those three was Miles Mikolas, whose projections were based on shaky data translations. Five pitchers regressed by 28% or more in walk rates.

In 2019, the change was less extreme. The rate of decline slowed – only two pitchers were in dark red – but the majority of pitchers still underperformed. Flaherty beat projections, which is no surprise considering his historic run in the second half. Projections still were not able to account for Mikolas. Then there was Gallegos, who defied the odds to become one of the best relievers the Cards have had in a while.

That’s two years of full-season data and one consistent trend: extreme performers (Mikolas, Flaherty, and Gallegos) were the most likely to outperformed their expected walk rates. Otherwise, the staff generally saw walk rates go up under Maddux’s guidance.

In 2020 those trends were more balanced. Five pitchers saw an increase in BB rate. Five pitchers saw a decrease. That said, most of the improvement in walk rate versus projections was limited to the back of the roster – Woodford, Webb, and Oviedo. Two of those players were rookies with pretty sketchy projections. Hudson improved; credit to Maddux there. Wainwright rediscovered his old self. Flaherty, Kim, Ponce, and Gomber all underperformed.

Sample size is a consideration for the season, as well as pitching/playing environment. So, 2020 was better but it’s pretty easy to poke holes in. The trend largely continued.

Let’s look at 2021.

Huh. Considering the Cardinals’ horrendous league rank in walk rates, I really expected that chart to look a lot worse. Six of the club’s top-10 pitchers have improved on their projected walk rates compared to preseason ZiPS projections. Four have declined, and those four have REALLY declined. Dark red doesn’t do justice to the terrible that is Reyes’ and Gant’s actual walk rates compared to projections.

But, as usual, Maddux has made it work. Before Tuesday’s game, Gant had an ERA under 2. Reyes had an ERA under 0.5.

If so many pitchers have improved their walk rate, how can the club rank so low in team walks? What’s going on here? There are several reasons:

1. The walk problem is amplified by the back of the roster.

Using the top 10 pitchers was a method that worked well for full-season evaluations of Mike Maddux. It’s limiting here because it removes the margins of the roster. At the same time, it really helps identify the source of the walk problem,

Among the core of the Cardinals’ pitching staff, the walk issue is isolated to three high-use arms who have struggled with control issues their entire professional careers: Gant, Reyes and Helsley.

(Note: Kim is also showing a 17% decline from expected. This is misleading. He currently has a 2.73 BB/9. That’s very good but ZiPS thinks it should be much lower. This looks like a translation problem for KBO stats and not a “real” underperformance. Kim is exactly on pace with 2020. It’s best to just throw him out of this conversation for now.)

The team walk issue is then amplified by almost universally poor performances from down-roster pitchers. If I reversed the chart above to sort by lowest innings to most innings, you would see a lot more dark red ink. For example:

Bernardo Flores - 0 IP, 18.00 BB/9
Junior Fernandez – 1 IP, 18.00 BB/9
Jake Woodford – 8.2 IP, 7.27 BB/9
Kodi Whitley – 9.2 IP, 4.66 BB/9
Jordan Hicks – 10.0, 9.00 BB/9

Throw Tyler Webb (10.80) and Daniel Ponce de Leon (6.10) onto the list as well. These players account for a relatively low percentage of team innings but a relatively high percentage of the league-worst walk totals.

2. The walk problem is not an issue of roster construction.

Yes, the Cardinals expected players like Alex Reyes and Ryan Helsley to have high walk rates. That’s probably also true of Gant, Ponce, and Woodford. The club has internal systems that aren’t all that different than ZiPS. They rely on these projections in roster construction.

Frankly, the Cardinals front office had no way of knowing that so many back-end arms would underperform their projected walk models as dramatically as they have. They also did not expect – and should not have expected – major walk problems from control arms, like Kodi Whitely or Tyler Webb. They can and do account for volatility in performance, but not this much underperformance over this number of players.

In other words, yes, the club built this team knowing that they would have some pitchers who would struggle to keep the ball in the zone. They did well to balance that with control specialists across the roster. That’s true of the rotation depth: For Martinez (who has been fine with walks but has struggled in the past), there was Mikolas. For Ponce or Gant, there was Kim. And of the bullpen: For Reyes and Helsley, there was Gallegos and Whitley. For Cabrera, there was Webb.

The walk problem is not a problem of design. It’s a problem of execution and volatility.

3. Is Mike Maddux part of the problem?

As time has gone on, more and more of the top-inning pitchers on the Cardinals have met or exceeded their walk projections. Underperformance early in Maddux’s tenure has now stabilized for much of the pitching staff’s core, particularly in the rotation.

What do we make of that for Maddux?

Not much, to be honest. For two seasons, the club has constructed the core of their rotation out of largely reliable, control-oriented arms. Flaherty, Wainwright, Mikolas, and Kim fit this mold. Hudson doesn’t but he was bound to improve his walk rates as he grew more comfortable in the majors based on his minor league profile alone.

We have seen the rotation arms bounce back and forth relative to projections, but by and large, the ones they have kept are the ones that are relatively stable. Maddux uses this command/control ability to its full advantage. The starters pitch to the edge of the zone, generating a large percentage of weak contact and acceptable whiff rates. That generates both Ks and weak contact. In Busch with this defense, Ks and weak contact lead to outs and high strand rates. The rotation routinely outperforms its FIP. This is Maddux’s way.

I’m not sure how well this strategy plays for the rest of the roster. The core of the bullpen is built around guys with “stuff” – arms that have ridiculous movement but weren’t given the chance to stick in the rotation largely because of a lack of command and control.

That seems to be the way the front office wants to operate. Throw strikes and go deep into games? You’re a starter. Have crazy movement and velocity but don’t know where it’s going? To the pen with you!

Maddux has improved K rates throughout the roster, but in the pen that has clearly come at the expense of added walks. When Reyes, Helsley, Cabrera, Ponce and others pitch to the edges of the zone, they can rack up impressive K totals. And painful walk rates.

Because of their physical inconsistency and the small sample sizes involved in the bullpen and spot starts, these kinds of arms are the most susceptible to wild swings in rate stats.

That’s what we are now seeing.

Maddux is not a big part the walk problem. But I do think that his strategy can lead to elevated walk totals. Not THIS elevated, but more than the club might have under the leadership of an average pitching coach. At the same time, I’m very sure that the average pitching coach would not be able to coax consistent overperformance in other areas, particularly ERA vs. FIP.

At the end of the day, runs prevented is what matters the most for a pitching staff.

For now, I think the walk problem mostly comes from volatility in execution and sample size from the margins of the staff. Reyes, Helsley, Gant, Whitley, and a few others should normalize. Others – Webb, Ponce, and Woodford – have already seen their roles reduced and that will likely continue until they can display better command. The result should be an improvement in team walk rankings. It might not translate to more wins, since the team has done well to minimize the impact of all those bases on balls so far. But the overall performance will be better, more consistent, and significantly more watchable.