Can one splitter grip make two pitches by turning the seams?

Abstract

At low spin factors, the seam orientation of a baseball changes its Magnus lift. Lyu, Smith, Elliott and Kensrud (2022) measured lift coefficients of 0.05 and 0.07 for a two-seam spin axis against 0.15 and 0.16 for a four-seam axis at spin factors of 0.05 and 0.10, with the two converging by 0.15 [1]. We propose a "two-face splitter" in which a quarter turn of the ball in the split grip selects a floating four-seam face or a dead two-seam face with the same arm action. Our model gives about 8 inches between the two at 78 mph and 1,050 rpm [2]. The measurements were made near 66 mph and the effect appears to fade by about 80 mph, so the design is best suited to slower splitters, forkballs and split changes.

1. Motivation and gap

The splitter is the highest-whiff pitch family in our data, with a league baseline of 34.4% whiff and a .255 xwOBA [2], and Driveline's Travis Sawchik reports that MLB splitter usage has doubled since 2022 [3]. Low-spin versions already exist. Our analysis found 28 splitter and forkball pitcher-seasons under 1,000 rpm from 19 pitchers in 2025 and 2026, including Cade Smith's 2026 splitter at 578 rpm with -5.2 inches of IVB and 45.3% whiff [4]. Seams also visibly matter for splitters, since 52 of 170 splitter rows (31%) showed 30 degrees or more of deviation between spin axis and movement, the highest share of any pitch type [4].

Pitchers already make deliberate orientation choices with split grips. MLB.com reported that Roki Sasaki's 2026 splitter used a narrower grip in a two-seam orientation [5], and Spencer Medick told FanGraphs that finger depth and dexterity are what lower splitter spin [6]. What we did not find in the research we reviewed is any pitcher or coach describing the choice of spin-axis orientation as a way to select the amount of Magnus lift. That is the gap this proposal addresses.

2. Physical basis

2.1 Spin factor and the drag-crisis band

Magnus lift depends mainly on the spin factor, the ratio of surface speed to flight speed. In Nathan's units, S is about 8.53 × 10^-3 times spin in rpm divided by speed in mph, and the Reynolds number is about 2,180 times speed in mph [7].

S = Rω_⊥ / v ≈ 8.53 × 10^-3 · rpm / mph,   Re ≈ 2,180 · mphAt 78 mph and 1,050 rpm, S ≈ 0.10 to 0.11 and Re ≈ 1.7 × 10⁵. Most fastballs sit near S 0.2.

Smooth spheres undergo a drag crisis, a sudden drop in drag as the boundary layer becomes turbulent before separating, at a Reynolds number of 350,000 to 400,000, while a dimpled golf ball does so at 40,000 to 60,000 [8]. Roughness acts as a trip, as Anderson and Cadou's trip-wire sphere illustrates [9]. Mehta and Pallis describe a smooth "bald" tennis ball whose single equatorial ridge triggered transition near a Reynolds number of 140,000, far below a smooth sphere, after which its drag rose again above 200,000 [10]. A baseball, whose roughness is concentrated in the seams, sits in an intermediate band where the location of those seams relative to the separation points can matter a great deal.

2.2 The measured lift difference

Lyu, Smith, Elliott and Kensrud (2022) measured lift and drag in free flight at a Reynolds number of 1.44 × 10⁵, about 66 mph, using 104 balls with four throws each, with a standard deviation of 0.007 in lift coefficient [1].

The authors measured the asymmetry between upper and lower separation points directly. It averaged 1.0 and 0.7 degrees for the two-seam axis at S of 0.05 and 0.10, against 8.3 and 11.0 degrees for the four-seam axis. Their explanation is that on a two-seam axis, at the 0 and 180 degree orientations, the seams are absent from the top and bottom separation points, so the ball behaves like a smooth sphere near its drag crisis and shows a periodic reverse Magnus effect. The orientations converge at S near 0.15, where the boundary layer on the bottom surface reaches a relative Reynolds number of about 167,000 and becomes fully turbulent. Drag did not depend on orientation for spinning balls [1]. Magnus lift is itself a separation effect, since spin delays separation on one side and advances it on the other, deflecting the wake [11][12], so it is physically reasonable that seams placed at the separation points change it.

Two independent sources point the same way. Blazevich's textbook notes that four-seam pitches "produce more dramatic changes in trajectory although this is probably less pronounced at higher ball spin rates" [12]. And Kensrud and Smith (2018) found no appreciable two-seam versus four-seam lift difference at 40 m/s, about 89 mph [13], consistent with the effect disappearing at higher speeds.

2.3 The speed window

A 1,000 rpm splitter sits at S of roughly 0.11, inside the measured split. But the two orientations converge near a Reynolds number of 1.75 × 10⁵, about 80 mph [1]. The effect is therefore credible for slower splitters, forkballs and split changes, roughly 70 to 80 mph, and may vanish above that. Utah State notes that altitude lowers the Reynolds number, so seam effects behave as they would on a slower pitch [14], which suggests the window may extend slightly higher at altitude. That is our inference.

3. Simulated shape

We integrated each pitch with a fourth-order Runge-Kutta scheme using Nathan's pitched-ball coefficients, a fixed spin vector and an air density of 1.194 kg/m³, calibrated against 2025 Savant pitcher averages [15][2]. For the two faces we replaced the default lift curve with Lyu and colleagues' coefficients at S = 0.10, 0.16 for the four-seam axis and 0.07 for the two-seam axis. Inputs are a right-handed pitcher at a 40 degree arm angle, 78 mph, 1,050 rpm, 90% efficiency and a 1:30 tilt, paired with an 88 mph four-seamer.

The two faces finish about 8 inches apart along the line of the spin direction, both reaching the plate at 72.1 mph after 0.482 seconds. When each is aimed through the four-seamer's position at the Baseball Prospectus tunnel point, 23.8 feet from the plate [16], the float face finishes 6.4 inches below the four-seamer and the dead face 8.9 inches below and 2.0 inches to its glove side, both 53 milliseconds later. The difference that opens between the two faces after the tunnel point is therefore about 2.5 inches vertically and 2.4 horizontally [2].

The figure runs the same model in your browser and draws both faces. There is no seam-shifted wake slider, because the seam effect here acts through the lift coefficient rather than a side force. The model holds the lift coefficients constant, so it does not show the expected fading of the difference as speed rises toward 80 mph. With the tunnel box ticked, every pitch passes through the four-seamer's tunnel point, so the catcher's view shows the divergence after the swing decision. Untick it to aim each pitch at the middle of the zone, and use slow motion to replay at a quarter speed. Live values can differ from the table by about 0.1 inch, and approach angles change with the aiming mode. As in all such views, the divergence grows with the square of time, so the late "drop" a hitter perceives reflects prediction error rather than a force that starts late [11].

4. Expected effectiveness

Our gradient-boosting shape model (velocity, IVB, HB, arm angle, extension and differences from the primary fastball; five-fold cross-validation grouped by pitcher) explains 56% of out-of-sample variance in whiff rate (RMSE 6.6 points) and 24% in xwOBA (RMSE 0.058) [2]. It predicts 36.6% whiff and a .271 xwOBA for the float face and 34.7% and .267 for the dead face, against league baselines of 34.4% and .255 for splitters and 31.2% and .280 for changeups.

The float face sits in a lightly populated region (4 real pitcher-pitch rows within one distance unit, 20 within two). Similar shapes include Sean Sullivan's 2026 changeup (77.1 mph, 9.1 inches IVB, 10.1 arm side, 22.2% whiff, .281 xwOBA over 65 pitches), Brent Suter's 2026 changeup (76.6, 9.7, 10.1, 33.2%, .268 over 415) and Kyle Hendricks's 2025 changeup (79.5, 8.5, 10.8, 25.3%, .269 over 987). The dead face is more isolated (none within one unit, 3 within two), and its nearest neighbour is Fernando Cruz's 2026 splitter (80.0 mph, 1.4 inches IVB, 4.2 arm side, 52.0% whiff, .214 over 484 pitches) at 1.64 units [2][4]. These are similar shapes, not users of this design.

As with any pair, the model scores each face alone. The hypothesised added value comes from two pitches with the same release, arm speed and spin rate but about 8 inches of difference in movement, which the model has no feature for. Both faces are about 10 mph slower than the fastball in our arsenal. Whether hitters can distinguish the faces in flight is unknown.

5. Complementary pitches and sequencing

The float face behaves like a firm, riding changeup that stays near the four-seamer's line, 6.4 inches under it after the tunnel point, while the dead face finishes lower and farther glove side. That gives a fastball-heavy pitcher three pitches on nearly one line, plus a curveball (73 mph, -11.7 inches IVB) in the opposite quadrant. The design suits pitchers with 84 to 90 mph fastballs, common in high school and college, or slower split-change and forkball pitchers, since their splitters naturally sit inside the speed window. For a pitcher whose splitter already exceeds about 85 mph, our analysis suggests the second face would add little.

6. Execution

Grip and arm action

Use one split grip in two orientations. For the float face, place the fingers across the seams so the ball spins on a four-seam axis. For the dead face, rotate the ball a quarter turn so it spins on a two-seam axis. Keep the same depth, pressure and arm speed for both, matching the fastball's arm speed. Film from behind at 1,000 frames per second or more to confirm which axis each face actually produces.

What anatomy permits and what it costs

The split is held by the intrinsic hand muscles. The dorsal interossei abduct the fingers and the long flexors press the pads, as Gray's anatomy describes [17]. Finger spread is limited by hand size and intrinsic strength, and a ball held deeper loses part of the late finger action that, in Matsuo and colleagues' (2024) data, makes the ball leave 8 to 12 m/s faster than the hand [18], which is one reason splitters lose speed. A practical risk follows, which is our inference. Fingertip shear in the last milliseconds creates the spin [19], and rotating the ball may change whether a fingertip sits on a seam, which plausibly changes how much shear it can apply before slipping. If the quarter turn also changes the spin rate, the spin factor changes too, and the comparison between faces is confounded.

No kinetic study of splitters or forkballs exists. Changeups carry the lowest elbow and shoulder loads of the pitch types studied (Fleisig et al. 2006; Dun et al. 2008) [20][21], but a splitter is not a changeup. Oshima and colleagues (2024) found that the forkball grip raised the stiffness of the flexor digitorum superficialis [22], a muscle that Park and Ahmad (2004) showed contributes to resisting valgus load at the elbow [23]. A deep split grip therefore works tissue that also protects the elbow, which raises a fatigue question rather than a known risk. Introduce the second face inside existing pitch counts and build tolerance gradually. For learning, Rob Gray's constraints-led approach suggests defining the target by its outcome (two spin axes, one arm action) and using ball-flight data as feedback rather than prescribing finger angles [24].

7. How to test it

• Use athletes or a pitching machine at 70 to 85 mph and 800 to 1,200 rpm, and film every throw from behind at 1,000 frames per second or more to confirm the spin axis orientation.
• Track spin and movement with Rapsodo, TrackMan or Hawk-Eye, and compare spin-based and observed axes to separate lift changes from any seam-shifted wake (Nathan 2023; Driveline 2020) [25][26].
• Throw at least 30 pitches per face at matched spin rate, because seam effects vary from shot to shot [27].
• Judge shape over dozens of repetitions and value over hundreds of pitches [28].

The prediction is that at matched spin rate and efficiency, the float face will show several inches more IVB and arm-side run than the dead face at about 70 to 75 mph, and that the gap will shrink toward zero as speed approaches and exceeds about 80 mph. Testing at three speeds within one session would check both halves of that prediction. No difference at 70 to 75 mph would falsify the design for that pitcher.

Compare both faces

See both versions beside your fastball

Import your TrackMan or Rapsodo CSV into Pitch AI's Pitch Lab to see your movement profile, with each version of your splitter beside your fastball, and use the grip library videos as a starting point.

8. Limitations and open questions

• Reynolds number extrapolation. The coefficients come from about 66 mph. At 78 mph (Re about 1.7 × 10⁵) the pitch is already close to the measured convergence near 1.75 × 10⁵, so our simulated 8 inches is probably an upper estimate.
• Constant coefficients. The simulation holds lift constant over the flight and across speeds, while the real effect depends on both.
• Aerodynamic torque. Borg and Morrissey (2014) measured seam-related aerodynamic torque on slowly rotating balls that can change their spin and orientation in flight [29]. We estimate this becomes small above about 1,000 rpm, but a 1,050 rpm splitter is near that boundary.
• Ball and seam variation. Lyu and colleagues' measurements were precise for their ball lot, but game balls vary in seam height and surface, which changes separation [27].
• Hitters and command. Whether hitters can see the difference, and whether the second face costs command, has not been measured.

Frequently asked questions

Start from your own data

Build your next pitch in Pitch Lab

Pitch AI's Pitch Lab has grip library videos, maps your movement profile, estimates your arm angle and gives pitch design recommendations to help you build your next pitch.

References

1. Lyu B, Smith L, Elliott J, Kensrud J (2022). The dependence of baseball lift and drag on spin. Proceedings of the Institution of Mechanical Engineers, Part P: Journal of Sports Engineering and Technology (online August 1, 2022). https://doi.org/10.1177/17543371221113914
2. Osnacz A (2026). Pitch AI new-pitch trajectory and shape models: RK4 integration of Nathan's pitched-ball coefficients calibrated to 2025 Baseball Savant pitcher averages, and a gradient-boosting shape model of whiff% and xwOBA (5-fold cross-validation grouped by pitcher; whiff R² 0.56, RMSE 6.6 points; xwOBA R² 0.24, RMSE 0.058). Unpublished analysis, October 2026. Method summary in New pitch types from first principles.
3. Sawchik T (2025). The search for missing pitches: Has pitch design given the game all the shapes it needs, or are we at the end of the beginning? Driveline Baseball, September 22, 2025. https://www.drivelinebaseball.com/2025/09/the-search-for-missing-pitches-has-pitch-design-given-the-game-all-the-shapes-it-needs-or-are-we-at-the-end-of-the-beginning/
4. Baseball Savant, MLB Advanced Media. Statcast search exports grouped by pitcher and pitch type, plus the spin-direction and active-spin leaderboards, 2025 and 2026 regular seasons (2026 season to date). Retrieved October 2026 and analysed by the author (5,102 pitcher-pitch-seasons with at least 50 pitches). https://baseballsavant.mlb.com/statcast_search
5. MLB.com (2026). Does Roki Sasaki throw a forkball or a splitter? Why not both? MLB.com, May 2026. https://www.mlb.com/news/roki-sasaki-s-forkball-and-splitter-explained
6. Laurila D (2026). Spencer Medick Addresses Splitter Spin. FanGraphs, May 7, 2026. https://blogs.fangraphs.com/spencer-medick-addresses-splitter-spin/
7. Nathan AM (2008). The effect of spin on the flight of a baseball. American Journal of Physics 76(2), 119-124. https://doi.org/10.1119/1.2805242
8. Smits AJ, Ogg S (2004). Aerodynamics of the golf ball. In Hung GK, Pallis JM (eds.), Biomedical Engineering Principles in Sports. Kluwer/Springer, chapter 1, p. 9 (Figure 1.5).
9. Anderson JD, Cadou CP (2023). Fundamentals of Aerodynamics, 7th edition. McGraw Hill. Section 6.6, The flow over a sphere, the real case, pp. 513-515 (Figures 6.9 and 6.10).
10. Mehta RD, Pallis JM (2004). The aerodynamics of a tennis ball. In Hung GK, Pallis JM (eds.), Biomedical Engineering Principles in Sports. Kluwer/Springer, chapter 5, p. 105 (Figure 5.4).
11. Bahill AT, Baldwin DG (2004). The rising fastball and other perceptual illusions of batters. In Hung GK, Pallis JM (eds.), Biomedical Engineering Principles in Sports. Kluwer/Springer, chapter 10, pp. 275-276.
12. Blazevich AJ (2017). Sports Biomechanics: The Basics. Optimising Human Performance, 3rd edition. Bloomsbury. Chapter 16, The Magnus Effect (sports examples box on baseball and cricket).
13. Kensrud JR, Smith LV (2018). Drag and lift measurements of solid sports balls in still air. Proceedings of the Institution of Mechanical Engineers, Part P: Journal of Sports Engineering and Technology 232(3), 255-263. https://doi.org/10.1177/1754337117740749
14. Smith BL (2022). Effect of velocity or altitude on seam shifted wake, part 3 of 4 (Post 74). Baseball Aerodynamics, Utah State University, April 28, 2022. https://www.baseballaero.com/2022/04/28/effect-of-velocity-or-altitude-on-seam-shifted-wake-part-3-of-4-post-74/
15. Nathan AM (2026). Trajectory Calculator, new 3D version (updated June 26, 2026). University of Illinois. https://baseball.physics.illinois.edu/trajectory-calculator-new3D.html
16. Long J, Judge J, Pavlidis H (2017). Introducing Pitch Tunnels. Baseball Prospectus, January 24, 2017. https://www.baseballprospectus.com/news/article/31030/prospectus-feature-introducing-pitch-tunnels/
17. Drake RL, Vogl AW, Mitchell AWM (2022). Gray's Basic Anatomy, 2nd edition. Elsevier. Chapter 7, Upper Limb, pronation and supination pp. 412-413; intrinsic hand muscles p. 430.
18. Matsuo T, et al. (2024). Factors of individual differences between the speeds of the hand and ball during baseball pitching. Journal of Sports Sciences 42(23), 2216-2228.
19. Kinoshita H, Obata S, Nasu D, Kadota K, Matsuo T, Fleisig GS (2017). Finger forces in fastball baseball pitching. Human Movement Science 54, 172-181. https://doi.org/10.1016/j.humov.2017.04.007
20. Fleisig GS, et al. (2006). Kinetic comparison among the fastball, curveball, change-up, and slider in collegiate baseball pitchers. American Journal of Sports Medicine 34(3), 423-430. https://doi.org/10.1177/0363546505280431
21. Dun S, Loftice J, Fleisig GS, Kingsley D, Andrews JR (2008). A biomechanical comparison of youth baseball pitches: is the curveball potentially harmful? American Journal of Sports Medicine 36(4), 686-692.
22. Oshima R, et al. (2024). [Forkball grip and flexor digitorum superficialis elasticity, shear-wave measurement.] Journal of Sports Medicine and Physical Fitness 64(11), 1127-1132.
23. Park MC, Ahmad CS (2004). Dynamic contributions of the flexor-pronator mass to elbow valgus stability. Journal of Bone and Joint Surgery (American) 86(10), 2268-2274.
24. Gray R (2021). How We Learn to Move: A Revolution in the Way We Coach & Practice Sports Skills. Perception Action Consulting & Education. Chapter 7 (constraints-led approach and transition feedback), chapter 8 (differential learning), chapter 14 (representative practice).
25. Nathan AM (2023). What Is The Hawkeye Spin Data Teaching Us? III: Searching for the Seam-Shifted Wake. University of Illinois, March 1, 2023. https://baseball.physics.illinois.edu/Hawkeye-4S-2S.pdf
26. Driveline Baseball (Aucoin D, Caravan A, eds.) (2020). More than what it seams: An introduction to seam-shifted wakes and their effect on sinkers. Driveline Baseball, November 1, 2020. https://www.drivelinebaseball.com/2020/11/more-than-what-it-seams-an-introduction-to-seam-shifted-wakes-and-their-effect-on-sinkers/
27. Smith BL, Garrett J, Dufour P, Francis E (2024). How seams alter boundary layer separation points on baseballs. Experiments in Fluids 65, 27. https://doi.org/10.1007/s00348-023-03751-3
28. Premier Pitching Performance (2026). The Pitch Design Phase in Baseball: Developmental Fit, Workload-Informed Timing, Biomechanical Determinants of Effectiveness, and Evidence-Based Evaluation Thresholds. Premier Pitching Performance blog, February 2, 2026. https://premierpitching.com/blogs/premier-pitching-chronicles/the-pitch-design-phase-in-baseball-developmental-fit-workload-informed-timing-biomechanical-determinants-of-effectiveness-and-evidence-based-evaluation-thresholds
29. Borg JP, Morrissey MP (2014). Aerodynamics of the knuckleball pitch: Experimental measurements on slowly rotating baseballs. American Journal of Physics 82(10), 921-927. https://doi.org/10.1119/1.4885341

Related reading

• New pitch types from first principlesThe hub. Methods, the gap map and all six proposals.
• The pole-seam riserA low-slot four-seam that rides like a higher slot.
• The float cutterA riding cutter in the mid 80s.
• Mirror sweepersOne spin, two breaks, set by seam orientation.
• The reverse hammerAn arm-side 12-6 curveball for the empty corner of the plot.
• The bullet fastballA gyro pitch at fastball speed.
• Splitter or changeup?How to choose your off-speed pitch.
• How to throw a changeupGrips, including split changes.