Fastball Paradox: Why Your Heater Hurts More Than It Helps

We have always been told “your fastball is your best pitch”, but is that entirely true? The four-seam fastball is the most used pitch in MLB (34% of all pitches). I have analyzed pitch-by-pitch data since the foreign substance ban in 2021. Using 4 seasons of data (2021-2024) I have tried to quantify the true run-preventing and contact‑disrupting value of each pitch type. Lots of different metrics were used, but the main ones were: raw ΔRE, xwOBA, wOBA, Whiff%, Strike%, and CSW%. Pitch selection and sequencing lie at the heart of modern pitching strategy. Traditional metrics like ERA and FIP aggregate season‑long outcomes, but conceal the individual contribution of each pitch type. Run Expectancy (RE) and Win Probability Added (WPA), especially when adjusted for context, reveal the real‑time value of every offering. This study leverages Retrosheet and Statcast data to:

• Isolate pure pitch‑type effects using context‑adjusted ΔRE;
  

• Augment with contact and miss metrics (xwOBA, whiff%, CSW%);
  

• Project season‑long impact over a typical 23 652‑pitch workload (averaging 146 pitches per game);
  

• Offer strategic recommendations for optimizing pitch mix and development.

Metric Breakdown

Raw ΔRunExpectancy (raw ΔRE) isolates a pitch’s contribution to run outcomes by subtracting the average run swing of its exact base-out state. Metrics like xwOBA and wOBA measures a player’s offensive value based on the result of each plate appearance. They weigh each outcome differently, where a home run is more valuable than a single, unlike regular on-base percentage where a home run has the same value as a single. wOBA constants are assigned each year based on run value on each outcome. While OPS takes into account slugging percentage, valuing a home run more than a single. OPS vastly undervalues OBP which is around 1.8x more valuable than slugging. xwOBA is used to estimate wOBA based on launch angle, exit velocity, and more. xwOBA is great because it takes out the “luck” factor of where defensive players are and only isolates true contact quality. Whiff % and Strike % are two complementary rates that show different dimensions of a pitcher’s effectiveness. Whiff % measures how often a batter misses the ball when swinging. A higher Whiff% is important for getting strikeouts and weak contact. Strike % measures how often a pitch is called a strike, which is important for controlling the count and staying ahead in the at‑bat. CSW% stands for Called‑Strikes plus Whiffs percentage. It’s a single, catch-all metric that combines called strikes (pitches in the zone that the batter doesn’t swing at) and whiffs (swinging strikes). By combining “getting the batter to take a strike” with “making the batter swing and miss”, CSW% captures a pitcher’s overall ability to control the zone and miss bats in one easy‐to‐interpret number. High CSW% pitches are called strikes and generate whiffs more often, an important ability for a pitcher suppressing contact and runs. Since 2021 there have been nearly 3 million pitches thrown at the MLB level with around 18 main pitches being used. I focused on all pitches that were thrown over 10,000 times in the last 4 years which are:

Where the four-seam fastball is used the most followed up by the slider and sinker. These are the pitches I will be examining to find the true run value to find the most effective pitch.

Data and Methods

I scraped baseball savant for every pitch recorded from Opening Day 2021 through the end of 2024 (2,845,847 pitches), filtered to the 10 pitch types thrown more than 10 000 times: Four‑Seam, Slider, Sinker, Changeup, Cutter, Curveball, Sweeper, Split-finger, Knuckle Curve, and Slurve.

• Context Adjustment: For each pitch, I computed ΔRE = (post‑pitch RE – pre‑pitch RE). Then grouped by the 24 base–out states to derive a baseline RE per state and subtracted it, yielding raw ΔRE.
  

• Season Projection: Multiplying each pitch type’s raw ΔRE by 23,652 pitches produced a “Season Value” in runs prevented (negative) or given up (positive). 23,652 pitches was chosen because that is the average number of pitches a team throws per 162 games.
  

• Complementary Metrics:

  • xwOBA vs wOBA to gauge expected vs actual contact quality;
    

  • Whiff Rate (% swinging‑miss), Strike Rate (% of Strike outcomes), and CSW%;

• Statistical Tests: ANOVA and Tukey HSD confirmed highly significant mean differences across pitch types.

Results

The slurve pitch, while rarely used, generates the most runs saved per season compared to other pitches at around 90 runs saved. While on the opposite end of the spectrum, the changeup, curveball, and four-seam all give up more runs, even though they are some of the most used pitches. A widely used pitch, the slider, saves around 50 runs per season, while being thrown 469,000 times in the past 4 years.

This graph illustrates perfectly how 3 of the top 6 pitches actually create a negative run value. The slurve and split-finger are miles ahead of the pack when comparing runs saved.

This graph shows wOBA and xwOBA given up when comparing each pitch type. Much to be expected: the offspeed pitches have a lower wOBA while the fastballs have higher wOBAs. This was expected because as the pitcher's velocity increases so does the exit velocity of the batter, resulting in harder and farther hits and more bases.

This graph illustrates the difference between the Whiff%, Strike%, and CSW%. When looking at the graph the best pitches are going to be higher up, farther to the left, and have a lighter and larger circle. The four-seam fastball has a great strike rate at almost 50%, which is expected as it is the go-to pitch for most pitchers and they have the most control over it. The changeup and split-finger are great at generating high whiff rates, but pitchers do not have a lot of control of them, which results in a low strike rate.

Pitch Groups:

Takeaways

The Four‑Seam Paradox

The Four‑Seam Fastball is the most used pitch in MLB, with nearly 1 million throws across four years and excels at getting called strikes (49.6%). Yet its raw ΔRE (+0.0010) and high xwOBA (0.345) reveal it yields the hardest contact and contributes to around 24 runs per season. Its value is in count leverage and tunneling, not pure run suppression. I think the four-seam would be much more valuable if it was used as a secondary pitch. It could be used in many cases such as:

• A two strike count where the batter is most likely sitting off speed

• A 3-0 count where the pitch needs to get a strike

• A batter who underperforms against four-seams.

• Setting up off speed/breaking balls

All of these instances are where a pitcher can catch a batter off guard or where a four-seam is favored.

Leveraging High‑Value Pitches

Slurve and Split‑finger deliver the greatest run savings (–90, –79 runs/season), but have lower strike calls (CSW ~30–32%). To maximize their value:

• Use them after fastballs to exploit arm‑speed deception.
  

• Elevate usage in mid‑to‑low leverage counts (1–1, 0–2, etc.) where getting a pitch called a ball will not change the run expectancy much.
  

• Develop tunneling between Slurve and fastballs to hide release points.
  

Sweepers and Sliders offer a middle ground: strong run suppression (–50 runs) with above‑average strike rates (~44–46%) and whiffs (~13–16%).

Situational Value of Changeup & Curveball

Although “expensive” in aggregate (+41, +26 runs), these pitches excel in specific matchups (opposite‑hand hitters) and two‑strike counts. They serve as timing disruptors, increasing fastball deception. Coaches should use them selectively by decreasing their usage, but not eliminating them.

Slider Group is Dominant

The slider group (slurve, sweeper, slider) have become very popular especially since late 2021. Pitchers are finding ways to increase spin rate and movement on these pitches while keeping a high velo. The slider group is continuously at the top of highest performing pitches: wOBA, xwOBA, raw ΔRE, Whiff%, Strike%, and CSW%. They are far and away the best pitches in baseball, even at their usage rate (21%).

Recommendations for the Future

Pitch‑Mix Optimization

• Mixing up pitches is still one of the most important things as a pitcher. Keeping a batter guessing on what pitch is going to come is crucial when trying to win an at-bat. This is what a four-seam is mainly used for, but I don’t think we should keep a four-seam dominant arsenal.

  • The fastball group should mainly be used as a get back in the count pitch, a strikeout pitch. I understand the whiff rate is low; however when a batter has only seen mid 80s and low 90s in a plate appearance and then sees a mid to high 90s fastball, the batter usually has a hard time catching up to the fastball.

  • All fastballs are not created equal, sinkers (two-seams) and cutters offer a negative raw ΔRE, saving 14 and 20 runs a year, while having relatively high wOBA and xwOBAs. These are great to set up off speed pitches, while saving runs.

  • Because of the extremely high wOBA, ΔRE differences, and less movement: fastballs are inherently inferior to off speeds because of the difference in exit velos.

  • The fastball isn’t all bad though, it sets up the offspeed to catch batters off guard. Without a fastball the offspeed pitches would not be as effective, and in turn we would see an increase in run expectancy.

• Am I saying we should get rid of four-seam fastballs and other fastballs because of the high wOBA and contact rate: absolutly not. The fastball is a staple at getting a strike and setting a tone for a pitcher. A batter needs to keep it in the back of their mind that any time they can see a 97 mph four seam and that they shouldn’t sit on a mid 80s slider every pitch. Four seams are overused in my opinion and should be dialed back to a lower usage rate, allowing other pitches to be used and keep hitters guessing. Pitchers should try to transition from the four-seam to a two-seam or cutter, both of which have a negative run expectancy, while keeping a high velocity and similar CSW%.

• The slider group is the most effective pitch at creating deception and generating strikes. The sliders have a great strike % while maintaining a high whiff %. The slider group could be used a little more as they are the most effective pitches at preventing runs and getting outs. Right now the entire group sits at around 21%, we could bump this up by using it as a two strike pitch more and substituting it instead of the changeup and curveball.

• Limiting the changeup usage is important for an elite pitcher. The changeup is by far the worst pitch when it comes to run expectancy, with the curveball coming in second. The changeup is set up by the fastball, utilizing tunneling to deceive hitters. However, even with the large velocity difference batters are able to adapt well and have a near .300 wOBA and xwOBA, coming in 4th behind all fastballs as the worst wOBA. Even when going up against opposite handed batters, the changeup has around 33.5 raw ΔRE, which would still make it the worst pitch in baseball. Limiting a changeup in favor of a slider or another off-speed pitch like a split-finger or even a curveball/knuckle curve would set up a pitcher for more success. Abandoning the changeup isn’t the best idea, but ideally a pitcher would not use it more than a couple times in a game.

• The curveballs are usually subpar compared to other breaking pitches, but not always. When going up against opposite handed batters, pitchers almost break even when it comes to run expectancy (slightly favoring positive raw ΔRE). This is a great opportunity for pitchers to use it, especially as it has a sub .278 wOBA and being in the bottom half of xwOBA pitches. Curveballs on their own against same sided batters tend to be very detrimental to a pitcher and his run expectancy; however against opposite handed batters pitchers can excel if they set it up correctly. If a pitcher cuts out their changeup and focuses mainly on a fastball, slider, and curveball combo, a pitcher can use his curveball against opposite handed batters catching them off guard with two different breaking pitches. Left handed pitchers would excel with this as around 75% of batters are right handed, making them have the upper hand in most situations.
  

Player Development & Scouting

• Prioritize high‑spin, late‑break training for Slurve and Split‑finger specialists.

  • I found that as spin rate increases with all pitches, wOBA, xwOBA, and raw ΔRE decreases.

  • Slurve and split-finger pitches were not used often, but when they were they were the most effective pitch in baseball, find pitchers that use them and invest in them.

• Invest in spin‑axis and release‑tunneling analytics to replicate elite off‑speed profiles.

  • Tunneling is one of the most important skills for pitchers. If a pitcher cannot hide his off-speed to go along with his fastballs he will get crushed. Hiding off-speed pitches is essential to being an elite pitcher.

• Identify prospects with raw “stuff” that maps to top raw‑value pitches.

  • Pitchers with high stuff+ will succeed in the long run, especially if they are using high value pitches. Having high stuff+ and CSW% is essential for a pitcher to succeed to the next level.

• Lefties are dominant with a good slider and curveball.

  • Lefties are a hot commodity. A lefty with a good curveball and slider are usually a good pitcher because most batters are righties and when a breaking pitch is coming in on your hands it is so much harder to hit. As the pitch breaks into the hands of the hitter, there is a lot less surface area for a batter to make contact with so that in turn results in a high whiff rate and lower exit velocities which creates lower wOBA and negative ΔRE.

Statistical Significance

My findings are statistically significant by any conventional criterion (α = 0.05), both my overall ANOVA and many of the Tukey pairwise contrasts show p < .05 (in fact, p ≪ .001 in most cases).

• Overall effect:
  ANOVA gives F(16, 2 843 657) = 18.73, p < 2 × 10⁻¹⁶

A one‐way ANOVA on 2.84 million pitch–by–pitch Δrun_exp values revealed a highly significant effect of pitch type on run‐expectancy change, F(16, 2 843 657) = 18.73, p < 2 × 2 × 10⁻¹⁶, indicating that not all pitch types produce the same average shift in run expectancy. Tukey’s HSD post‐hoc tests (95% family‐wise CI) confirmed several pairwise differences after controlling for multiple comparisons; for example, Eephus pitches produced a mean ΔRE 0.0373 runs higher than Changeups (95% CI [0.0186, 0.0560], p_adj < 0.001), whereas Split‐finger fastballs reduced run expectancy by 0.0051 runs compared to Changeups (95% CI [–0.0086, –0.0017], p_adj < 0.001). While the large sample makes these differences highly “significant” on paper, the actual run expectancy on pitches are very small (just 0.001–0.04 runs per pitch), so it’s essential to weigh real‑world impact, not just p‑values.

Conclusion

No single metric fully captures a pitch’s value. Raw ΔRE, xwOBA, whiff%, and CSW% provide a good profile: breaking and off‑speed pitches suppress runs most effectively, while fastballs serve as the indispensable “anchor”. Future pitch designs and usage strategies should embrace a balanced arsenal with less fastball use for better run value, but still in use for control and deception. By integrating advanced statistical modeling with player development, teams can unlock the next frontier in pitching performance.