Free tools Windows power users keep installed
One-click scans. No signup required.
In one analysis of 8,022 pitcher-and-pitch-type pairs from consecutive MLB seasons (2017–2025), whiff rate carried over from one year to the next much more strongly than run value did. Among pairs where the pitch was thrown at least 400 times in both seasons, the year-to-year correlation was 0.70 for whiff rate and 0.27 for run value. The figures come from a 2026 article by the author YMori, based on Baseball Savant pitch data, and they describe that sample, not a general law about any individual pitch.
What the study measured
The question is simple: if a pitch missed a lot of bats one year, does it still miss bats the next year? And if a pitch was hard to hit one year, is it still hard to hit the next? The author tests both by following the same pitcher and the same pitch type across two consecutive seasons. The pitch data cover 2017 through 2025. The ongoing 2026 season is excluded.
As an Amazon Associate I earn from qualifying purchases.
Eligibility and sample size
A pair qualifies only if the pitcher threw at least 100 pitches of that type in both seasons. That rule produces 8,022 pairs. Each pair is then placed in a band according to the smaller of its two season pitch counts. Sorting by the smaller count means a band can include a pair where one season had far more pitches than the other, so the bands are not equivalent in pitcher role or usage.
Removing the pitch-type baseline
Some pitch types generate more whiffs than others as a matter of course. To avoid rewarding a pitch merely for belonging to a whiff-heavy category, the author subtracts each season’s average for that pitch type before calculating correlations. The question becomes whether a pitcher’s version of a pitch sits above or below the typical result for that pitch type, season after season.
#1 Best Overall
How each metric is defined
- Whiff rate is the share of swings that miss.
- xwOBA allowed uses expected values derived from exit velocity and launch angle on batted balls. Strikeouts, walks, and other non-batted-ball outcomes count as they occurred.
- Run value is the sum of how each pitch outcome changed expected runs. In the article’s presentation, higher is better for the pitcher. The underlying pitch-level
delta_run_expfield is positive for the batter, so the sign is flipped when totals are summarised from the pitcher’s side.
Whiff rate persisted; run value persisted less
The gap between the two metrics is larger than the differences between sample-size bands. The table below shows the author’s year-to-year correlations by band.
| Smaller season pitch count | Pairs in band | Whiff rate correlation | Run value correlation |
|---|---|---|---|
| 100–199 | not stated | 0.53 | 0.09 |
| 200–399 | not stated | 0.63 | 0.17 |
| 400–799 | not stated | 0.70 | 0.25 |
| 800+ | 401 | 0.74 | 0.35 |
Two comparisons stand out. Whiff rate in the smallest band (0.53) is higher than run value in the largest band (0.35). And the 400+ pitch subset, which the author analyses separately with 1,962 pairs, gives 0.70 for whiff rate and 0.27 for run value.
Rank #2
Who stays near the top?
The author also asks a more practical question: of the pitches that ranked in the top 20% in one season, how many were still there the next year? For pairs with 400+ pitches in both seasons:
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware match| Metric | Stayed in top 20% next season | Fell to bottom half |
|---|---|---|
| Whiff rate | 58% | 12% |
| Run value | about one in three | 36% |
| Expected if seasons were unrelated (author’s figure) | about 20% | about 50% |
Across all qualifying pairs, the author reports that 52% of top whiff-rate pitches and 29% of top run-value pitches stayed in the top 20%. The comparison is made against the pitch-type-adjusted baseline described above, and the 400+ figures are the stricter version of the same test.
Fixed-count estimates from a signal-and-noise model
Raw correlations depend on how many pitches each season contains. To separate that effect, the author fits a signal-and-noise model and estimates correlations at fixed pitch counts. These are model outputs, not raw measurements, and the intervals come from 200 resamples of pitchers.
Correlation at 500 pitches
- Whiff rate: 0.68 (95% interval 0.65–0.70)
- xwOBA allowed: 0.41 (0.38–0.44)
- Run value: 0.20 (0.18–0.23)
The 0.20 run-value figure at exactly 500 pitches is a model estimate. It should not be read as the same calculation as the 0.27 observed among 400+ pitch pairs.
Rank #4
Run value at a starter’s main-pitch volume
At roughly 1,100 pitches, a figure the author describes as a starter’s main pitch over a season, the modelled run-value correlation is about 0.35–0.36. The main text gives an interval of 0.27–0.39.
How many pitches before a season means something?
- Whiff rate: about 100 pitches is the point where half of the season’s whiff-rate number is signal (interval 76–138).
- xwOBA allowed: about 300 pitches (interval 176–498).
- Run value: about 1,800 pitches, but the interval is very wide (455–2,212). The author says the model parameters are hard to tell apart within the pitch counts available, so this number should not be treated as precise.
Why run value carries over less
The author splits each pitch’s run value into components: pitches in the middle of a plate appearance, strikeouts, walks and hit-by-pitch outcomes, batted-ball quality, batted-ball luck, and base-out situation. Among pairs with 400+ pitches in both seasons, the components behave differently.
| Component | Share of one season’s run-value variation | Year-to-year correlation |
|---|---|---|
| Batted-ball luck | 27% | 0.06 |
| Base-out situation | 4% | −0.03 |
| Strikeouts, walks, hit-by-pitch | not stated individually | 0.57 |
| Mid-plate-appearance pitches | not stated individually | 0.60 |
The strikeout/walk/hit-by-pitch and mid-plate-appearance components together account for about 30% of one season’s variation, and about 70% of the link between a high season and a high following season. Luck and situation carry much of the noise but little of the persistence.
A simple forecasting test
As an exploratory check, the author adds the three components that persist (excluding luck and situation) to a prediction of next season’s run value. The correlation rises from 0.29 to 0.35. The model was trained on pairs whose second season was 2021 or earlier and tested on pairs whose second season was 2022 or later. The 95% resampling interval for the improvement is 0.03–0.10. This is one author’s test, not a forecasting method that has been validated for clubs or players.
Two examples
The article illustrates the decomposition with Adam Wainwright’s sinker and Corbin Burnes’s cutter. The author notes that two hand-picked pitches show no general pattern. Their purpose is to make the components concrete.
Recommended Free Tools
How to read these numbers
- Correlation mixes noise with real change. The author writes: “The correlations mix random noise and real change in the pitch (a new grip, lost velocity and so on).” A correlation cannot tell you which of the two is responsible for a given pitcher’s movement.
- Selection may bias the sample. Pitchers who fall below 100 pitches in a season, including some who stop pitching after a poor year, drop out of the pair sample. The author says this is not corrected.
- The 2020 season is included. Pairs touching the shortened 2020 season remain in the analysis.
- The largest band is small. The 800+ group contains only 401 pairs.
- The decomposition involves judgement. The split of run value into quality and luck depends partly on modelling choices the author made.
- Validation is self-reported. The author says a run-value table was checked with an alternate computation, and that correlations and counts were recomputed from raw Savant data with pandas. No independent replication has been published.
Source
All sample sizes, estimates, and conclusions above come from YMori’s article on DEV Community, posted September 29, 2026 and edited September 30, 2026: Does a pitch’s performance carry over to next season? Whiff rate vs run value on 8,022 MLB pairs. The author links a code repository, which is the place to check the calculations.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
