Towel drills have long been used as a bridge between rehab and return-to-throwing. Lightweight. Low-risk. An easy way to cue arm path without the load of a baseball. But what if these drills weren’t as stress-free as assumed?

A recent study by Ishigaki et al. (2024) investigated elbow valgus stress during towel drills and compared them to full-effort baseball throws. Seventeen healthy collegiate pitchers completed reps using three towel drill variations (face towel, short foam tube, long foam tube), followed by max-effort throws on flat ground. Using wearable sensors, researchers measured elbow torque, arm slot, and arm speed during each task.

The results? Eye-opening. Towel drills may reduce torque — but not by much.

What the Study Found

• Towel drills produced lower average elbow valgus torque than max-effort throws (p < 0.001).
  
  
• However, drills using a face towel still reached ~80% of the valgus torque seen during live throws.
  
  
• There was no significant relationship between elbow torque and arm slot in any condition.
  
  
• In towel drills, higher arm speed directly correlated with greater elbow stress (r = strong correlation, p < 0.01).
  
  
• This relationship was not observed in full-effort throws, where arm speed did not predict torque levels.
  
  
• The variability across individuals highlighted how self-selected effort levels in towel drills could mask high stress.
  
  

Why This Matters

This study challenges the perception that towel drills are a near-zero stress activity. When a face towel produces 80% of the torque of a max-effort throw, it forces us to rethink how and when these drills are used — especially in return-to-throwing protocols or low-load ramp-up days.

What’s most interesting is the disconnect in biomechanics. In full throws, arm speed didn’t predict torque — likely due to more refined sequencing, momentum transfer, and mechanical timing. But in towel drills, higher speed was tightly linked to higher stress. That suggests that even without a baseball, how fast you move still matters.

Subjective effort is not a reliable safeguard. A pitcher may believe they’re taking it easy, but if they accelerate their arm aggressively (especially without external load), stress can approach game-level demands.

This isn’t a critique of towel drills — they can still serve a purpose. But they must be dosed carefully. If used during recovery, ramp-ups, or mechanical work, the stress profile needs to be better understood and monitored.

How We Apply This at VeloU

At VeloU, we do not use towel drills in our throwing programs. While they’ve historically been used as low-stress mechanical tools or rehab bridges, we believe the nature of the drill introduces movement constraints that often alter mechanics to meet a preset target — rather than promote authentic movement patterns.

This study reinforces that concern. If a towel drill can produce 80% of the elbow torque of a full-effort throw, it’s not just “dry reps” — it’s a potentially high-load stimulus masquerading as a low-risk alternative. Combine that with altered mechanics, and the utility becomes questionable.

We prefer progressions that preserve arm patterning, reinforce authentic tempo, and allow for stress modulation through load management — not tool substitution. Towel drills, especially when effort is self-selected, offer limited feedback and may encourage compensatory strategies that increase risk rather than reduce it.

The findings from this study don’t rule out towel drills entirely — but they do add to the growing list of cautions when deciding how and when to deploy them.

This article is part of Applied Baseball Science by Dr. Nicholas Serio, where we break down the biomechanics, performance science, and injury research shaping the modern game. Powered by VeloU (Velo University) — where research meets real-world baseball.

Reference

Ishigaki, T., Yamada, K., Yoshida, R., Kawakami, J., & Uehara, K. (2024). Elbow valgus stress during towel drills in college baseball players. Journal of Shoulder and Elbow Surgery, 33(4), 765–772. https://doi.org/10.1016/j.jse.2023.09.010