Research theme
The question everybody in professional cricket wants answered: does getting stronger make you bowl faster, and does it change your technique?
The literature is a mess, and the paper says so. Studies have associated faster ball speed with greater lower-limb strength (Kiely et al., 2021; Letter et al., 2022) and shoulder strength (Ramachandran et al., 2021), but actual strength training intervention studies found no link with upper arm velocity (Hislen et al., 2023) or ball release speed (Callaghan et al., 2021), and no relationship between lower limb strength and front foot ground reaction force characteristics. Felton attributes the mess to heterogeneous samples, differing research designs and different strength-measurement methods.
The proposed solution: forward-dynamics simulation, where you can change strength and nothing else. Note carefully what this means — this is not a training study. Nobody lifted anything. The strength increase is a parameter change inside ten computer models.
Method, plainly: take the ten individual-specific 16-segment simulation models already built and optimised in Felton et al. (2023), raise the maximum isometric torque parameters at the ankle, knee, hip and shoulder flexors and extensors by 5% in every model, re-optimise each model (varying the joint torque activation parameters) to maximise ball release speed, and compare against the original optimised techniques. Paired t-tests on discrete parameters (SPSS), SPM1D on continuous parameters, alpha 0.05, Cohen’s d effect sizes.
What they measured
- How fast the ball leaves the hand (ball release speed, m/s).
- How long the phase lasts (front foot contact → ball release, ms).
- Six ground reaction force parameters (peak horizontal braking and vertical force, BW; horizontal braking and vertical loading rate, BW/s; horizontal braking and vertical impulse).
- Four joint angle time histories: front ankle, front hip, front shoulder, bowling shoulder.
- Four joint torque time histories: front knee, front hip, front shoulder, bowling shoulder.
- Two ground reaction force time histories: horizontal braking and vertical.
Findings
Causal within the model; ten bowlers.
A 5% strength increase bought 0.8% of ball speed. Optimised 40.7 ± 1.6 → 41.1 ± 1.6 m/s, p = 0.002, Cohen’s d = 0.22. Statistically significant, practically tiny. The paper’s own abstract calls it “<1%”.
Nothing else changed significantly. Phase time 102 ± 6.0 → 102 ± 6.0 ms (p = 0.411, d = 0.05). All six kinetic parameters non-significant: peak horizontal braking force 3.94 ± 0.4 → 3.83 ± 0.6 BW (p = 0.362); peak vertical force 5.68 ± 0.8 → 5.62 ± 1.0 BW (p = 0.630); horizontal braking loading rate 128 ± 45 → 132 ± 48 BW/s (p = 0.552); vertical loading rate 172 ± 70 → 195 ± 75 BW/s (p = 0.227, d = 0.31); horizontal braking impulse 0.16 ± 0.05 → 0.15 ± 0.05 (p = 0.566); vertical impulse 0.29 ± 0.06 → 0.29 ± 0.07 (p = 0.522).
The one technique change claimed: a delay in bowling arm circumduction. “The findings of this study suggest that increased isometric flexor-extensor strength facilitates a greater delay in the timing of bowling arm circumduction (Figure 2).” Greater arm delay has independently been identified as a key ball-speed characteristic (Worthington et al., 2013; Felton et al., 2023) — so the story here is that strength lets you hold the arm back longer.
The lower limb: no change at all. No significant differences in front knee joint torque history, and no significant differences in front knee or front hip kinematics. The authors’ interpretation is not that lower limb strength is unimportant, but that the lower limb strength of these elite bowlers is already sufficient for the torque demands of this phase.
The scale comparison — the important number. 0.8% from a 5% strength increase, versus 13.5% from optimising the initial body configuration and movement pattern (Felton et al., 2023). Technique is worth roughly seventeen times as much as this strength manipulation. The authors offer a second explanation alongside “technique matters more”: ball release speed in elite males may be limited by the time available to exploit extra strength during the front foot contact phase (~102 ms).
The conclusion and its caution. Increases in isometric strength are “associated with increased ball release speeds (although moderate in elite males compared to optimising technique)”, isometric strength “does not appear to limit technique” in elite bowlers, and interventions should be prescribed accordingly. Caution is urged over the individual nature of movement-pattern self-organisation and the role of strength in injury prevention and in other phases of the action — both of which are outside this model.
Explicit future work: investigate decreasing strength in elite bowlers, or increasing it in non-elite bowlers, to find out when isometric strength does become limiting. This is the honest boundary of the claim.
What a coach should look for on video
This paper’s main practical value is a negative result, and coaches should treat it as such. It supports one weak observational cue and one strong prescriptive message that is not a video cue at all.
Cue 1 — Bowling arm delay (weak support)
- The cue: How late the bowling arm starts to come over.
- Camera view + frame: Side-on, from front foot contact; find the first frame of arm circumduction.
- What “good” looks like: Arm still back at front foot contact and starting later.
- What the fault looks like: Arm leading the action.
- Why it matters: This paper claims added shoulder strength facilitates a longer delay, and arm delay is well established as a speed characteristic. But flag this as weak — see the contradiction note: the journal version of this same analysis found no significant joint angle differences at all, and the non-significant trend for the shoulder ran in the opposite direction.
Not a video cue, but the strongest message in the paper — a prescribing message: For an elite male bowler already inside a professional S&C programme, adding 5% isometric strength at the ankle, knee, hip and shoulder is worth under 1% of ball speed in this model, while fixing the landing position and movement pattern is worth 13.5%. If speed is the goal, spend the coaching hours on the front foot contact position and sequence, not on adding strength. Strength remains essential for injury prevention and for the earlier phases of the action — the paper says so explicitly — but it is not the speed lever at this level.
Explicitly not supported:
- Any claim that getting stronger straightens the front leg. No significant front knee or hip kinematic differences were found, and no knee torque differences.
- Any claim that added strength changes front-foot ground reaction forces. All six kinetic parameters were non-significant.
- Any transfer of these findings to junior, club or female bowlers. The authors flag this as an open question and specifically call for the study to be done on non-elite bowlers.
Caveats and limits
- This is a simulation, not a training intervention. The “5% strength increase” is a parameter change in ten computer models. No bowler in this study trained, got stronger, or bowled faster in the real world. If someone quotes this paper as evidence that lifting makes you bowl faster, that is a misreading.
- n = 10 male ECB elite fast bowling group members (age 20.7 ± 2.4 y; height 1.91 ± 0.08 m; mass 86.9 ± 8.5 kg). Statistical power is limited.
- Elite males only, already well trained. The finding is essentially “these particular athletes are already strong enough for this particular 0.1 s”. It says nothing about a 16-year-old club bowler.
- Isometric strength only, uniformly scaled by 5% at the flexors and extensors of four joints. No power, rate of force development, elastic qualities, or whole-body strength qualities.
- Front foot contact phase only. Strength’s role in the run-up, back foot contact, delivery stride and follow-through is not modelled — and this is exactly where strength is most likely to matter for a bowler’s ability to arrive in the optimal landing position.
- Centre-of-mass velocity at front foot contact was held constant. In reality, a stronger bowler might run in faster; that pathway is excluded by design.
- Planar (2D) model, single-parameter-set optima, robustness untested.
- Conference paper. The journal version (2025) supersedes it and disagrees with it.
Relationship to other Felton work
- Builds directly on Felton 2023 — optimal initial position and technique for the front foot contact phase: the same ten models, the same optimised techniques (40.7 ± 1.6 m/s) used as the baseline.
- Uses the model from Felton 2020 — optimising the front foot contact phase.
- The equivalent question was asked in the PhD (Felton 2015 — PhD thesis: factors limiting fast bowling, +1.3% for one bowler) and the 2015/2017 conference papers (+1%).
- This paper is superseded by Felton 2025 — effect of increased strength on ball release speed, the peer-reviewed journal version of the same analysis — which reports different numbers and a materially different technique conclusion.
CONTRADICTION: (conference vs journal — the technique finding reverses) This conference paper’s central technique claim is that increased isometric strength “facilitates a greater delay in the timing of bowling arm circumduction”. The 2025 Journal of Sports Sciences paper, using the same ten models and the same 5% manipulation, reports that “No joint angle time histories were observed to significantly differ”, and that the (non-significant) trends were that the increased-strength techniques showed less knee extension, reduced trunk flexion and greater shoulder extension — which it describes as “contrary to expectations” and “contrary to previous research”. Where the conference paper says strength reinforces the known optimal technique, the journal paper says strength trends away from it and warns coaches against using strength interventions to alter front foot contact technique. The journal version is the one to trust.
CONTRADICTION: (the numbers changed) This paper reports raising isometric strength 5% at ankle, knee, hip and shoulder and gaining 0.8% (40.7 → 41.1 m/s). The 2025 journal paper reports the nominally identical manipulation (“lower body + shoulder”: ankle, knee, hip and shoulder, 5%) and gains 2.0% (40.7 → 41.5 m/s) — two and a half times as much. The journal paper splits the manipulation into lower-body-only (+1.5%, 41.3 m/s) and lower-body-plus-shoulder (+2.0%, 41.5 m/s), and neither matches 41.1 m/s. The analysis was evidently reworked between conference and journal; the discrepancy is not acknowledged in the journal paper.
TENSION: (with the earlier single-bowler work) the thesis and the 2015/2017 conference papers reported that a 5% strength increase let the bowler keep the front leg straighter and produce more front arm extension — strength reinforcing the optimal technique. This paper found no significant front knee or hip kinematic differences at all, and the 2025 journal version found the trend running the other way.
TENSION: (a strength result that has never been tested in the field) across the whole cluster, no Felton paper reports a strength training intervention on real bowlers. Every strength finding here (thesis +1.3%, conference +1%, this paper +0.8%, 2025 +1.5%/+2.0%) is a simulated parameter change. The real-world intervention literature the paper itself cites is contradictory: Callaghan et al. (2021) and Hislen et al. (2023) found no ball-speed gain from real strength training; Taliep & Maker (2021) and Feros et al. (2020) found gains (with Feros also finding reduced accuracy). The simulation’s answer — “a small but real gain, easily lost in the noise of an underpowered field study” — is a plausible reconciliation, but it has not been tested.