2019 — Male vs Female Elite Fast Bowlers Compared

Source note: full accepted manuscript read, including both results tables. All numbers below are taken directly from it.

Research theme

Coaching manuals for female fast bowlers were written from male data, on the untested assumption that the optimal action is sex-independent (as it is for standing throws in baseball and handball). But cricket bowling has a run-up, like javelin — and elite male and female javelin throwers do sequence differently. This study tested the assumption directly: 55 kinematic parameters computed for 20 elite male and 20 elite female fast bowlers, all national-squad or professional-with-international-potential, measured on an 18-camera Vicon system at 300 Hz at the ECB National Cricket Performance Centre (indoor, full-length artificial pitch), 47 markers plus a ball marker, six maximal good-length deliveries each, best three averaged. Groups compared with independent-samples t-tests (α = 0.05), Cohen’s d for significant differences. This is a measured, descriptive between-groups comparison — it says what men and women do differently, not what either should do.

What they measured

• Ball release speed, and release height as a percentage of standing height
• How fast the whole body is travelling forward at back foot contact, front foot contact and release (horizontal centre-of-mass velocity), and how fast it is dropping (vertical COM velocity)
• How far in front of the body the front foot is planted (front plant angle) and the back foot (rear plant angle)
• Delivery stride length as a percentage of standing height
• How side-on or front-on the hips and shoulders are at each key instant (pelvis and shoulder orientation; 180° = fully side-on, 270° = fully front-on), and the twist between them (pelvis–shoulder separation)
• How bent the front and back knees are, and the front and back hips
• How much the upper trunk is arched back or flexed forward (upper trunk angle, normalised to each bowler’s own neutral spine)
• How far back the bowling arm is (bowling shoulder angle) — larger = more delayed circumduction
• The time taken between back foot contact → front foot contact, front foot contact → release, and back foot contact → release
• How much the front knee flexes and then re-extends during front foot contact; how much the trunk flexes; how far the pelvis and shoulders rotate

Between-trial repeatability was strong: between-trial SD was 1.1–13.6% (mean 4.9%) of between-bowler variation, ICC 0.94–0.99 (mean 0.98).

Findings

Physique and performance

1. Ball release speed: 34.9 ± 1.7 m/s (men) vs 27.9 ± 1.4 m/s (women), P < 0.01, effect size > 0.80. That is roughly 126 km/h vs 100 km/h — a ~20% gap.
2. The women were significantly shorter (1.67 ± 0.07 vs 1.88 ± 0.08 m) and lighter (64.4 ± 8.8 vs 81.5 ± 7.1 kg), both P < 0.01, ES > 0.80, at comparable age (19.9 ± 3.2 vs 20.1 ± 2.6 years, P > 0.05).
3. Release height was lower in women even after normalising for their own height: 109.7 ± 3.0% vs 112.8 ± 4.1% of standing height, P < 0.05, large ES. Women release relatively lower, not just absolutely lower.

The run-up and the arrival at the crease 4. Women arrive at back foot contact slower: horizontal COM velocity 5.31 ± 0.46 vs 5.76 ± 0.58 m/s, large ES. The gap persists to release (3.36 ± 0.52 vs 3.69 ± 0.40 m/s, medium ES). 5. At back foot contact the men had a higher front leg (front hip angle 105.6 ± 19.5° vs 123.0 ± 21.5°, i.e. men more flexed/higher knee lift) and a more extended, leaned-back upper trunk (185.2 ± 8.4° vs 177.2 ± 7.1°, where >180° = extension). Both large ES. The authors’ interpretation, borrowed from baseball pitching: women may lack the trunk strength to balance with a high front knee lift. 6. At front foot contact, women were dropping less: vertical COM velocity −0.94 ± 0.23 vs −1.25 ± 0.36 m/s, large ES; and planted the front foot further in front: front plant angle −37.4 ± 2.8° vs −33.8 ± 3.2°, large ES.

Timing — the single clearest structural difference 7. Men spend longer getting from back foot to front foot: 191.9 ± 31.3 vs 172.3 ± 27.6 ms (medium ES). 8. Men spend markedly less time from front foot contact to release: 103.3 ± 11.2 vs 127.7 ± 11.0 ms (large ES) — women take about 24% longer through the delivery. 9. Total back foot contact to release was not different (295.2 ± 33.4 vs 300.1 ± 29.8 ms, ns). The women redistribute the same total time differently. 10. Cross-sport context given by the authors: front-foot-contact-to-release is 103 ms (cricket, men) vs 115 ms (javelin) vs 141 ms (baseball); 128 ms (cricket, women) vs 137 ms (javelin) vs 163 ms (baseball).

At ball release 11. Six of eleven release parameters differed. Women were more front-on at release: pelvis orientation 298.2 ± 9.0° vs 288.0 ± 10.6°, shoulder orientation 322.8 ± 13.3° vs 310.4 ± 15.0° (both large ES; 270° = fully front-on). 12. Women had more upper trunk flexion at release (153.1 ± 7.7° vs 159.5 ± 7.8°, large ES) and a more delayed bowling arm (bowling shoulder angle 233.9 ± 18.1° vs 219.4 ± 15.3°, large ES). 13. Women rotate the pelvis further through the action: pelvis orientation rotation from back foot contact to release 78.6 ± 15.2° vs 68.7 ± 15.3° (medium ES). Shoulder rotation was directionally similar but not significant (89.3 ± 21.1 vs 78.9 ± 19.2).

What did NOT differ (worth knowing — these are the male cues that survive) 14. Delivery stride length as % of height: 76.3 ± 7.8 (women) vs 78.5 ± 7.2 (men) — not significant. 15. Front knee angle at front foot contact: 166.7 ± 7.0 vs 164.1 ± 6.1 — not significant. Front knee angle at release: 172.5 ± 24.9 vs 167.3 ± 18.8 — not significant. 16. Front knee flexion during front foot contact (12.7 ± 13.9 vs 9.2 ± 10.8) and re-extension (18.4 ± 13.3 vs 12.4 ± 8.6) — not significant, but note the very large SDs in the female group: front-leg behaviour is far more variable in women even though the group mean is similar. 17. Trunk flexion from front foot contact to release: 35.5 ± 7.3 vs 31.0 ± 8.3 — not significant. 18. Pelvis–shoulder separation (counter-rotation) at every instant, and minimum separation (45.1 ± 13.4 vs 39.6 ± 9.6) — not significant. 19. Back knee angles, back hip angles, rear plant angle, trunk orientation — not significant.

The authors’ synthesis 20. Men generate ball speed by building large linear momentum in the run-up, holding onto it through back-foot-to-front-foot, then converting it to angular momentum about the mass centre with the front leg braking the lower body. Done efficiently, that conversion is fast, hence the 103 ms. 21. Women arrive with less linear momentum, and lose more of it between back foot and front foot. The angular momentum available is therefore not enough on its own. So they spend longer (128 ms) and recruit the large trunk rotator muscles to drive the pelvis and torso round the long axis — a sequence “more akin to throwing.” That is the mechanistic reason for findings 11–13.

What a coach should look for on video

Cue 1 — The shape of the time budget: back foot to front foot vs front foot to release

• Camera view + frame: Side-on, full delivery stride in frame, slow-motion or frame-by-frame. Count frames from back foot landing → front foot landing, then front foot landing → ball leaving the hand.
• What “good” looks like: In elite men the second window is roughly half the first (≈103 ms vs ≈192 ms; at 240 fps that’s about 25 frames vs 46). In elite women it is ≈128 ms vs ≈172 ms (about 31 vs 41 frames at 240 fps).
• What the fault looks like: A delivery that “hangs” after the front foot lands — a long, drawn-out second window — means momentum is being bled off rather than converted.
• Why it matters: This is the most robust single discriminator in the paper (large effect size) and it is measurable on a phone with a frame counter. A shortening second window is direct evidence the front leg is braking effectively.

Cue 2 — Hips and shoulders at ball release: how front-on has the bowler got?

• Camera view + frame: Behind-the-arm or high overhead, pause on the release frame.
• What “good” looks like: Men in this sample were at pelvis 288°, shoulders 310°; women at 298° and 323° — i.e. the women had rotated ~10–12° further towards fully front-on by release.
• What the fault looks like: A bowler who has already squared right up to the batter at release has spent their rotation early and has nothing left to give the ball.
• Why it matters: More front-on at release, combined with more pelvis rotation overall (78.6° vs 68.7°), is the signature of the rotational/throwing strategy. It is not a “fault” in a female bowler — it is how she is currently generating speed — but it tells you where her speed is coming from and therefore what to train.

Cue 3 — Front leg height and trunk position at back foot contact

• Camera view + frame: Side-on, pause on the frame the back foot lands.
• What “good” looks like: Men showed a higher front knee (front hip angle 105.6° vs 123.0°) and a trunk still leaning back past neutral (185.2° vs 177.2°).
• What the fault looks like: Front knee low and dragging through; trunk already upright or folding forward at back foot contact.
• Why it matters: Both differences are large-effect. But treat them as a strength diagnostic, not a technique instruction — the authors’ explanation is that the trunk strength to hold a high knee and an extended trunk simply wasn’t there. Coaching a taller position onto a bowler who cannot support it will just move the collapse somewhere else.

Cue 4 — Where the front foot lands relative to the body

• Camera view + frame: Side-on, pause on the front foot contact frame.
• What “good” looks like: This paper gives front plant angle −33.8 ± 3.2° (men) vs −37.4 ± 2.8° (women) — the women reach further out in front.
• What the fault looks like: An over-reached front foot planted a long way ahead of the mass centre.
• Why it matters: Large effect size difference, but note this paper does not relate plant angle to ball speed. Bull et al. (2026) in this cluster found larger plant angles associated with faster pre-delivery run-ups in men. Use this as an observation, not a target.

A cue this paper explicitly does NOT support: delivery stride length. It did not differ between the sexes and is not related to speed here. Do not coach stride length off this paper.

Caveats and limits

• n = 20 per group; elite/national-squad standard; both groups measured indoors on artificial turf (validated as representative by Alway et al. 2024, also in this cluster).
• Descriptive between-groups comparison, not a performance regression. It tells you men and women differ; it does not tell you which variables predict speed within the female group. For that, see Felton et al. (2015) and Lyons et al. (2023).
• Confounded by training history: the authors concede the male bowlers had “considerably more experience and time within a professional training environment,” given how recently the women’s game professionalised. Some of these “sex differences” may be exposure differences.
• Sample-size bias: the male professional population is far larger, so the male 20 may be more talent-selected than the female 20.
• No correction for multiple comparisons across 55 parameters, deliberately (to avoid Type 2 errors), which the authors flag as leaving elevated Type 1 risk. With 55 tests at α = 0.05 you would expect ~3 false positives by chance. The 11 large-effect findings are safer than the marginal ones.
• Strength and anthropometry were not measured, only inferred. Every mechanistic explanation in the discussion is speculative.

Relationship to other Felton work

• The full journal version of the female-bowling programme begun in Felton 2015 (5th World Congress) and almost certainly presented as Felton & King (2017) at ISB Brisbane.
• Built on the male reference model of Worthington, King & Ranson (2013), whose 55-parameter marker set and definitions it reuses wholesale, so male and female numbers are directly comparable.
• Directly cited and extended by Lyons 2023 in this cluster, which tests whether male-derived predictors hold inside a female sample.
• TENSION: front knee. The 2015 abstract found faster women had a more flexed front knee at release; here, front knee angle at release does not differ between the sexes at all (172.5° women vs 167.3° men, ns) — the women were, if anything, marginally straighter on average. What does differ is the variability (SD 24.9° vs 18.8°). The “female front knee collapses” story is about the spread of the female group, not its mean.
• TENSION: arm delay. Here the women had a more delayed bowling arm at release than the men (233.9° vs 219.4°). The 2015 abstract found that within women, the faster ones had an earlier arm. So women as a group are late, but the fast ones among them are relatively early — while in men and in Lyons (2023) later is faster. The arm-timing evidence in the female game does not hang together.