YMori

Posted on Mar 1 • Edited on Mar 5 • Originally published at zenn.dev

Does Exit Velocity Come from the Front Foot? Exploring Driveline Motion Capture Data

#python #baseball #datascience #biomechanics

License: Motion capture data from Driveline OpenBiomechanics Project under CC BY-NC-SA 4.0 (non-commercial, share-alike).
Citation: Wasserberger KW, Brady AC, Besky DM, Jones BR, Boddy KJ. The OpenBiomechanics Project: The open source initiative for anonymized, elite-level athletic motion capture data. (2022).
Derivative works (graphs, GIFs) in this article follow the same license. Commercial use by professional sports organization employees is restricted.

Note: This is an exploratory analysis with n=40. I use "suggests" and "trend" rather than definitive claims.

The Question

Two hitters. Almost the same bat speed. But one hits the ball 20 mph harder. Why?

I tried to answer this using 40 hitters' worth of professional-grade motion capture data from Driveline OpenBiomechanics Project.

The short answer: bat speed alone explained almost nothing. Front leg mechanics did.

→ GitHub: https://github.com/yasumorishima/baseball-cv

The Data

Driveline OpenBiomechanics Project (OBP)

40 hitters, C3D motion capture format
45 body markers (shoulder, elbow, wrist, hip, knee, heel, etc.)
360 Hz sampling rate
Exit velocity range: ~70–110 mph

C3D is a binary format for storing 3D motion capture data. I used ezc3d to parse it in Python.

This is the hitting counterpart of my pitching analysis — same dataset, same method, applied to batters.

Bat Speed Barely Predicts Exit Velocity

The first surprise: wrist speed (a proxy for bat speed) explained only 9.7% of exit velocity variance in this dataset.

bat speed only         → R² = 0.097
+ height/weight        → R² = 0.183
+ stride length        → R² = 0.494  ← jump of +0.311

Adding stride length (how far the hitter steps forward) nearly tripled the model's explanatory power.

This doesn't mean bat speed doesn't matter — it clearly does — but among hitters with similar bat speed, body mechanics appear to matter more.

Defining a "Body Efficiency Score"

Same approach as my pitching analysis: calculate the residual after controlling for bat speed and body size.

lm = LinearRegression().fit(
    df[['peak_wrist_linear_speed', 'height_in', 'weight_lb']],
    df['exit_velocity_mph']
)
df['body_efficiency'] = df['exit_velocity_mph'] - lm.predict(...)

Positive = hits harder than expected given bat speed (efficient body use)
Negative = bat speed is there but exit velocity isn't (arm-dependent)

I split the 40 hitters into quintiles (Q1 = least efficient, Q5 = most efficient).

Q1 vs Q5: Slower Bat Speed, 20 mph More Exit Velocity

	Q1 (inefficient)	Q5 (efficient)
Bat speed (wrist)	9.42 m/s	9.04 m/s
Exit velocity	79.8 mph	98.4 mph
Difference	—	+18.6 mph

Q5 hitters have lower bat speed but hit the ball nearly 20 mph harder. Something in their body mechanics is doing the work.

Overview Charts

The Front Leg Wall

Correlation with exit velocity (n=27 after dropping NA)

Feature	r	Significance
Stride length	+0.548	p<0.01
Hip-ankle gap (how far hip drifts past ankle)	−0.459	p<0.05

The two strongest predictors: stride long, hip doesn't drift forward.

The physical mechanism (hypothesis)

Stride creates forward momentum (linear energy)
        ↓
Front foot plants — ankle becomes a fixed pivot
        ↓
Hip stops moving forward (hip-ankle gap stabilizes)
        ↓
Linear → rotational energy conversion
        ↓
Bat accelerates → higher exit velocity

This is the "lead leg block" (LLB) concept in baseball coaching — when the front foot plants, it creates a wall that converts forward momentum into rotation.

Visualizing the wall

Top: 3D skeleton animation (red = lead leg)
Bottom: hip-to-ankle gap over time in the forward direction

Q5 (blue): after foot strike, the gap stabilizes — the wall is working
Q1 (orange): after foot strike, the hip keeps drifting forward — no wall

When the Knee Stops

Digging deeper, the most striking difference is when the front knee stops moving forward after foot strike.

Front knee forward velocity (foot strike = 0 ms)

Time	Q1	Q5
0 ms (foot strike)	+0.628 m/s	+0.584 m/s
25 ms	+0.349 m/s	+0.389 m/s
50 ms	+0.203 m/s	−0.019 m/s (stopped!)
100 ms	−0.212 m/s	−0.623 m/s
150 ms	−0.358 m/s	−0.807 m/s

Q5's knee stops completely at 50 ms and then rapidly extends. Q1's knee keeps moving forward.

Brace quality

Metric	Q1	Q5	Difference
Time to peak extension	0.185 s	0.107 s	78 ms faster
Peak extension velocity	355 deg/s	468 deg/s	+32%
Knee forward decel	0.425 m/s²	0.602 m/s²	+42%

Q5 hitters "stop fast and extend fast" — the knee locks into a stable axis quickly, and then the hip can rotate around it efficiently.

Knee angle animation

Solid line = knee angle, dashed = extension velocity, dot = current frame.

Skeleton GIF: Same Bat Speed, 20 mph Gap

Left (Q1): bat speed 9.42 m/s → exit velocity 74.6 mph (stride 0.72 m)
Right (Q5): bat speed 7.80 m/s → exit velocity 97.2 mph (stride 0.99 m)

Red = lead leg. The Q5 hitter takes a longer stride and the front leg brakes sharply after landing.

Full Mechanism Summary

Long stride → forward momentum (stride r=+0.548)
        ↓
Knee stops at ~50 ms post foot strike
        ↓
Hip doesn't drift forward (hip-ankle offset r=−0.459)
        ↓
Linear → rotational energy conversion (pivot)
        ↓
Rapid knee extension (468 vs 355 deg/s, +32%)
        ↓
Bat accelerates → +18.6 mph exit velocity

Limitations

n=40 is small. One facility (Driveline), one population (amateur to minor league)
Correlation ≠ causation. "Better front leg = more exit velocity" is not established
Wrist speed ≠ true bat head speed
Representativeness to MLB/NPB is unknown

This is exploratory. The patterns are interesting but shouldn't be over-interpreted.

Summary

Bat speed (wrist speed) alone explains only ~10% of exit velocity variance in this dataset
Adding stride length jumps explanatory power to ~50%
Q5 hitters' front knee stops within 50 ms of foot strike; Q1's keeps drifting forward
"Lead leg block" — stride creates momentum, rapid knee extension converts it to rotation — may be a key differentiator

The data suggests there's a consistent front-leg pattern among efficient hitters. Whether training to replicate it actually improves exit velocity is a different question.