Ball%: Simple, Underutilized, & Highly Effective
A few days, I got into a spat looking into Tanner Bibee.
I had him at the top of my "K & BB Going to Regress" query coming into the game.
In 3-2 counts, he only had K's this year but historically, there are more walks than strikeouts at 3-2 (see Singer's & Manoah's regression).
Second, his 38% Ball% equates to a 10% BB%. pic.twitter.com/t6GSvAYeKf
— Jeff Zimmerman (@jeffwzimmerman) May 8, 2023
My issue was that even though Bibee’s walk rate was good in 2022 (combined minor league rate of 1.8 BB/9) there were signs that his walks could be an issue once this season started. While Bibee had some luck in 3-2 counts is one issue, I’m just going to focus on Ball% (Balls/Pitches).
First off, there has been some great work with Ball% by Steve Wiemer in this year’s BHQ’s Forecaster (p. 72) on relievers. So check it out. My focus today is to just show the correction of Ball% to BB/9 and BB% and the predictiveness of Ball%. First off, here are the comparisons of a pitcher’s two walks rate (BB/9 and BB%) to Ball% over the past two seasons (min 60 IP). Included in the graphs are the r-squares and the best-fit lines.
Far from perfect correlations, but still extremely useful.
Using the formulas from the graphs, here are the walk rate equivalents for different Ball% values. The values colored yellow are greater than the criteria from my previous article on what walk rate is too high.
“Focusing on the original table, I will lower my acceptable walk threshold a bit. I went and averaged the 20 values around a 3.5 BB/0 (9.1% BB%)…”
| Ball% | BB/9 | BB% |
|---|---|---|
| 30% | 1.5 | 4.1% |
| 31% | 1.7 | 4.6% |
| 32% | 1.9 | 5.1% |
| 33% | 2.1 | 5.7% |
| 34% | 2.3 | 6.3% |
| 35% | 2.6 | 6.9% |
| 36% | 2.9 | 7.6% |
| 37% | 3.2 | 8.4% |
| 38% | 3.6 | 9.2% |
| 39% | 3.9 | 10.0% |
| 40% | 4.3 | 10.9% |
| 41% | 4.7 | 11.8% |
| 42% | 5.2 | 12.8% |
| 43% | 5.7 | 13.9% |
| 44% | 6.2 | 15.0% |
For a simple rule, pitchers wanted to keep a Ball% under 38%.
For a second bit of information, I compared the April and May walk rates, Ball%, and walk rate equivalents to see what is more predictive. For this study, I examined pitchers from 2010 and on who threw at least 20 IP in each month. Here are the r-squared values.
| BB% | BB/9 | Ball% | |
|---|---|---|---|
| BB% | 0.13 | ||
| BB/9 | 0.12 | ||
| Ball% Equivalent | 0.15 | 0.15 | 0.31 |
The Ball% is easily the most sticky and the two Ball% formulas calculated above do a better job at predicting the walk rates than the original stat.
Now, going back to Bibee, here are some of his recent Ball% and walk rates.
| Season | Level | BB/9 | BB% | Ball% | eqBB/9 | eqBB% |
|---|---|---|---|---|---|---|
| 2022 | A+ | 2.0 | 6% | 32% | 1.9 | 5% |
| 2022 | AA | 1.7 | 5% | 34% | 2.4 | 6% |
| 2023 | AAA | 4.7 | 13% | 39% | 3.9 | 10% |
| 2023 | MLB (first 2 GS) | 0.0 | 0% | 37% | 3.1 | 8% |
| 4/26/2023 | COL | 0.0 | 0% | 34% | 2.4 | 6% |
| 5/2/2023 | @NYY | 0.0 | 0% | 39% | 4.0 | 10% |
| 5/8/2023 | DET | 9.8 | 20% | 45% | 6.6 | 16% |
There are no issues with Ball% or walks in 2022. In 2023, he was struggling with walks in AAA with all his values being unacceptable. His first MLB start was a major improvement from AAA but the second start had some major red flags. Combined, both starts didn’t enter the Danger Zone, but were right on the edge.
While no one expected the blowup that happened in his third start, his control was already a major issue, and the regression came fast.
The simple lesson is that while Ball% is not readily available, it does a great job of determining a pitcher’s walk rate going forward, even better than the original walk rates.
Jeff, one of the authors of the fantasy baseball guide,The Process, writes for RotoGraphs, The Hardball Times, Rotowire, Baseball America, and BaseballHQ. He has been nominated for two SABR Analytics Research Award for Contemporary Analysis and won it in 2013 in tandem with Bill Petti. He has won four FSWA Awards including on for his Mining the News series. He's won Tout Wars three times, LABR twice, and got his first NFBC Main Event win in 2021. Follow him on Twitter @jeffwzimmerman.
This is great info. I had been looking at Zone% and First Pitch Strike% to “check” the validity of pitcher’s walk rate, but I would imagine Ball% is more predictive than Zone% in a large sample. I know you said it’s not readily available, but where can I find Ball%? Do you just tabulate it from individual game logs?
Ball%: https://baseballsavant.mlb.com/statcast_search?hfPTM=&hfPT=&hfAB=&hfGT=R%7C&hfPR=ball%7Cblocked%5C.%5C.ball%7C&hfZ=&hfStadium=&hfBBL=&hfNewZones=&hfPull=&hfC=&hfSea=2023%7C&hfSit=&player_type=pitcher&hfOuts=&hfOpponent=&pitcher_throws=&batter_stands=&hfSA=&game_date_gt=&game_date_lt=&hfMo=&hfTeam=&home_road=&hfRO=&position=&hfInfield=&hfOutfield=&hfInn=&hfBBT=&hfFlag=&metric_1=&group_by=name&min_pitches=0&min_results=0&min_pas=0&sort_col=pitch_percent&player_event_sort=api_p_release_speed&sort_order=desc#results
I know that’s a long link, but I didn’t want to hide it behind a link shortener because I wanted people to be confident that there were no shenanigans afoot.
Thanks! I just added a filter for >200 pitches and looks like we’re cooking with gas.
https://baseballsavant.mlb.com/statcast_search?hfPTM=&hfPT=&hfAB=&hfGT=R%7C&hfPR=ball%7Cblocked%5C.%5C.ball%7C&hfZ=&hfStadium=&hfBBL=&hfNewZones=&hfPull=&hfC=&hfSea=2023%7C&hfSit=&player_type=pitcher&hfOuts=&hfOpponent=&pitcher_throws=&batter_stands=&hfSA=&game_date_gt=&game_date_lt=&hfMo=&hfTeam=&home_road=&hfRO=&position=&hfInfield=&hfOutfield=&hfInn=&hfBBT=&hfFlag=&metric_1=&group_by=name&min_pitches=200&min_results=0&min_pas=0&sort_col=pitch_percent&player_event_sort=api_p_release_speed&sort_order=desc#results
Thank you both!