2018 — Kinematic parameters contributing to spin in elite finger spin bowling

Research theme

Almost nothing was known biomechanically about what makes one elite finger spinner spin the ball harder than another — coaching of spin, as the paper says bluntly, had been “based on anecdotal evidence”. Twenty-three elite male finger spin bowlers (England senior/A/U19 squad members or first-class county players flagged by the ECB spin bowling national coach as potential internationals) each bowled ten maximal-spin good-length deliveries on a full-length indoor pitch with their full run-up. Motion was captured with an 18-camera Vicon system at 300 Hz and 56 markers; spin rate came from a Trackman Doppler radar. Thirty technique parameters were correlated against spin rate and fed into a forwards stepwise linear regression. This is a correlational, observational study — it shows what the fastest-spinning bowlers happen to do, not that changing a bowler’s pelvis angle will raise their spin rate.

What they measured

• How fast the bowler was travelling just before back foot landing (run-up velocity, mass-centre speed over the 0.060 s before BFC)
• Which way the back foot points when it lands (back foot orientation at BFC; 270° = parallel to the stumps, 360° = pointing straight down the wicket)
• Which way the front foot points at landing and at release (front foot orientation at FFC and BR), and how far it pivots afterwards (BR to follow-through)
• How open the hips are at four moments — back foot contact, front foot contact, ball release, follow-through (pelvis orientation; 180° = side-on, 270° = square-on/front-on, >270° = rotated past front-on)
• The most side-on the hips ever get during the stride (minimum pelvis orientation BFC→BR), and total hip rotation BFC→BR and BFC→FT
• The same four-moment set for the shoulders (shoulder orientation), plus minimum shoulder orientation and total shoulder rotation
• The wind-up between hips and shoulders at BFC, FFC and BR (pelvis-shoulder separation angle, i.e. the spin bowler’s “X-factor”)
• How far the bowler folds forward from front foot contact to release (trunk flexion FFC→BR)
• How much the bowling elbow straightens after the arm passes horizontal (elbow extension UAH→BR — the legality measure, 15° is the legal limit)
• Whether the front knee straightens or collapses (front knee extension FFC→BR)
• How much the front hip turns in as the pelvis swings round (front hip internal rotation FFC→BR)
• Trunk rotation, bowling-shoulder internal rotation, forearm pronation and wrist flexion, all FFC→BR

Findings

1. Spin rate range: 1432–2143 rpm (1685 ± 170 rpm); ball release speed 17.7–23.4 m/s (20.4 ± 1.3 m/s ≈ 45.6 mph). So the fastest spinner in this elite group spun it ~50% harder than the slowest.
2. Hip openness at front foot contact was the single biggest correlate of spin: r = 0.674, p < 0.001. Hip openness at ball release was essentially tied: r = 0.676, p < 0.001. Each on its own explained ~43% of the between-bowler variance in spin rate.
3. The bowlers who spun it hardest had their pelvis around “mid-way” (≈225°) at front foot contact — i.e. roughly 45° open from side-on — and then rotated it past front-on (>270°) by ball release. Group means: pelvis 204.6 ± 13.3° at FFC and 265.9 ± 14.2° at BR, so the high-spin group sat well above both.
4. Shoulders short of side-on at FFC, and past front-on at release, also went with more spin. Shoulder orientation at BR: r = 0.462, p = 0.027. Shoulder orientation at FFC: r = 0.405, p = 0.055 (near-significant). Group mean shoulder at FFC was 183.7 ± 9.5°, at BR 279.4 ± 29.6°.
5. Bigger hip-shoulder separation at front foot contact = more spin: r = 0.521, p = 0.011. Group mean 20.9 ± 9.8°. This is the classic stretch-shortening “wind-up”, and the authors read it as evidence that finger spin is a sequential proximal-to-distal action, not a push.
6. Both feet pointing further down the wicket at landing went with more spin. Back foot at BFC: r = 0.433, p = 0.039. Front foot at FFC: r = 0.416, p = 0.048. The authors read this as feet permitting the pelvis to be more open.
7. Forearm pronation from FFC to release was positively correlated at the relaxed threshold: r = 0.409, p = 0.053 (mean 11.0 ± 6.4°). The proposed mechanism: pronating late keeps the fingers on the ball surface longer, applying a bigger angular impulse.
8. Upper-arm internal rotation was flatly unrelated to spin: r = −0.031, p = 0.887. This is a genuinely surprising negative result — in throwing and racket sports long-axis upper-arm rotation is the classic final link. The authors argue finger spinners deliberately keep the shoulder externally rotated to keep the hand on the correct side of a straight arm.
9. Wrist flexion FFC→BR was also unrelated: r = −0.133, p = 0.544 (despite a huge spread, 0.2–52.2°).
10. Run-up speed did not matter: r = −0.155, p = 0.481. Mean run-up 2.53 ± 0.48 m/s — a jog.
11. Front knee extension did not matter: r = 0.156, p = 0.477. Note the mean was negative (−9.5 ± 17.0°), i.e. the average elite finger spinner’s front knee flexes through the delivery. There is no evidence here for the coaching mantra of a braced front leg.
12. Front hip internal rotation FFC→BR was positively correlated at the relaxed threshold: r = 0.371, p = 0.082 — the mechanism by which the pelvis gets round.
13. Elbow extension UAH→BR was unrelated to spin: r = −0.078, p = 0.722 (mean only 3.65 ± 3.24°, max 10.5°, i.e. this group was well inside the 15° legal limit).
14. Stepwise regression admitted only one variable. No second parameter reached p < 0.05.

TENSION: the paper’s text says pelvis orientation at ball release was the best individual predictor explaining 43.1%, with pelvis at FFC explaining 42.9% — but Table 3 labels model (a) “pelvis orientation at FFC … 43.1%” and model (b) “pelvis orientation at BR … 42.9%”, i.e. swapped relative to the text. The correlations (BR r = 0.676 vs FFC r = 0.674) favour the text. Sanders’ own thesis abstract then uses the table ordering (FFC = 43.1%) while the thesis’ research-question summary uses the text ordering (BR = 43.1%). The two numbers are 0.2 percentage points apart, so nothing coaching-relevant hangs on it — but do not quote “43.1% for FFC” or “43.1% for BR” as if the source were unambiguous.

What a coach should look for on video

This paper is unusually friendly to phone-camera coaching, because every variable that predicted spin is a whole-body orientation angle, not a hand action. The pelvis and shoulder lines are exactly what a behind-the-arm or high overhead camera resolves best.

Cue 1 — Hip line at front foot contact (“are the hips already opening?”)

• Camera + frame: High overhead is ideal; behind-the-arm from a raised position works. Scrub to the frame the front foot first touches down.
• What “good” looks like: Hips roughly half-open — about 45° between fully side-on and square to the batter (≈225° in the paper’s convention). Draw a line across the two hip points and compare it to the crease.
• What the fault looks like: Hips still fully side-on (hip line parallel to the pitch) at front foot landing — the classic coached position. In this elite sample that went with less spin.
• Why it matters: Strongest single correlate of spin rate found (r = 0.674). Also, the thesis notes bowlers who counter-rotated the pelvis past side-on could never reach the open ball-release position afterwards.

Cue 2 — Shoulder line at front foot contact (“shoulders not fully closed”)

• Camera + frame: Same view, same frame.
• What “good” looks like: Shoulders slightly open of side-on, and critically more open than the hips are — a visible gap between the hip line and the shoulder line of roughly 20° (group mean 20.9°; more was better).
• What the fault looks like: Hip line and shoulder line stacked on top of each other — no wind-up — or shoulders opened past the hips.
• Why it matters: Separation at FFC r = 0.521, p = 0.011. This is the stretch-shortening wind-up. Note the direction: hips open first, shoulders lag.

Cue 3 — Hip line at ball release (“do the hips finish past square?”)

• Camera + frame: High overhead or behind-the-arm; the frame the ball leaves the hand.
• What “good” looks like: Hips rotated beyond square to the batter (>270°). Only 23% of this elite group achieved that, and they were the big spinners.
• What the fault looks like: Hips stalling square-on or short of it at release.
• Why it matters: r = 0.676, p < 0.001.

Cue 4 — Foot alignment at both contacts

• Camera + frame: High overhead. Two frames: back foot landing, then front foot landing.
• What “good” looks like: Both feet pointing further down the wicket than the traditional “back foot parallel to the crease” model. Group means were ~298° back foot (i.e. already ~28° open of parallel) and ~337° front foot.
• What the fault looks like: Back foot dead parallel to the popping crease, front foot heavily closed off.
• Why it matters: r = 0.433 and r = 0.416. The feet are the enabler for the pelvis positions in cues 1 and 3 — the paper treats them as consequences rather than causes, so treat them as a check on hip position rather than a target in themselves.

Cue 5 — Sequencing (order, not magnitude)

• Camera + frame: Side-on at the highest frame rate the phone offers (240 fps if available). Step frame-by-frame from front foot contact to release.
• What “good” looks like: Hips start turning first, shoulders follow, arm and finally forearm pronation last. The paper’s summary sentence: “the segments then rotated sequentially, starting with the pelvis and finishing with the pronation of the forearm.”
• What the fault looks like: Hips and shoulders turning together as one block (a “push” action).
• Why it matters: Supported indirectly (via the separation correlation), not measured as a timing variable in this study. Treat as a plausible cue, not a proven number.

Cues this paper does NOT support — do not coach these off this study:

• Bracing the front knee. No correlation (r = 0.156, p = 0.477), and the average bowler’s knee flexed.
• Running in harder. No correlation (r = −0.155).
• Snapping the wrist. No correlation (r = −0.133).
• “Bowl with a big internal rotation of the upper arm.” No correlation (r = −0.031) — and the authors argue the opposite, that elite finger spinners stay externally rotated.
• Finger action. Not measured at all in this study (see Caveats).

Caveats and limits

• n = 23, all elite males. Large for this population, tiny for regression. Only one variable survived the stepwise model.
• Correlational. Nothing here is causal. It shows what high-spin bowlers do, not what would happen if you changed a bowler.
• Circularity risk the authors flag themselves: all 23 have been coached under the existing side-on orthodoxy, so their techniques are not an unbiased sample of what is possible.
• Alpha inflated to 0.10 for several of the “interesting” variables (forearm pronation, front hip internal rotation, shoulder orientation at FFC, pelvis at BFC). Treat those as hypotheses.
• Finger motion was not assessed. The authors name this as a limitation. Sanders’ thesis reveals why: four finger/knuckle markers were placed for this exact purpose but had to be discarded because the markers were too close together to track reliably even with 18 cameras.
• Maximal-spin deliveries in an indoor facility, not match deliveries. The bowler was told to spin it as hard as possible; there was no batter, no match consequence, and no accuracy requirement.
• Ball release speeds here (mean ≈45.6 mph) are ~14% slower than elite finger spinners bowl in test matches (~52 mph) — see the thesis file. Lab spin rates may likewise not be match spin rates.

Relationship to other Felton work

• Restated almost verbatim as the 2019 World Congress conference paper, with Felton as first author. Same 23 bowlers, same numbers.
• Forms Chapter 5 of Sanders’ 2019 PhD thesis.
• Directly motivates the ROM paper: if open hips at FFC drive spin, what physical capacity permits open hips? Answer: front-hip total rotation arc.
• CONTRADICTION (with prior literature, stated by the authors): contradicts Woolmer & Noakes’ coaching model of side-on hips and shoulders at FFC rotating to front-on at BR. The elite high-spin technique is more open earlier than that model.
• CONTRADICTION (with Beach et al. 2017): Beach, using club-level bowlers, described finger spin as a push-like movement and found maximum internal shoulder rotation velocity correlated with spin rate. This paper finds a sequential action and no relationship with upper-arm internal rotation. Population differs (club vs elite) and Beach measured velocities where this study measured displacements, so the two may not be strictly comparable — but the coaching implications point opposite ways.
• CONTRADICTION (with the wrist-spin study in the same thesis): for wrist spinners, less shoulder rotation and a shoulder orientation short of front-on at release predicted more spin (r = −0.837, r = −0.875) — the reverse direction to finger spin. Do not transfer these cues across spin types.
• TENSION (with Spratford et al. 2018, raised in Sanders’ thesis §9.6.1): Spratford recommends bowlers seeking to reduce elbow extension should become more side-on at BFC, which is exactly the position this paper associates with lower spin. Sanders explicitly calls this “a mixed message to the coaching community”.