Physical preparation & prevention

What the sport actually demands

Three findings coexist and are usually misreported as contradictory. They answer different questions:

• Individual actions are alactic and maximal. Fencing is a high-intensity anaerobic sport relying on alactic energy sources — though some bouts do evoke blood lactate at or above 4 mmol/L.
• The match is aerobically financed. In simulated three-period épée direct eliminations, the oxidative contribution is 80–90%, and it is significantly higher than the phosphagen and glycolytic contributions in all three periods.
• The competition day is an endurance and thermoregulatory event. An international tournament lasts 9–11 hours; bouts are only about 18% of total competition time; effective fighting time is 17–48 minutes.

The correct coaching statement is all three at once: actions are alactic and maximal, the match is aerobically financed, the competition day is an aerobic-recovery and heat-tolerance event. Encoding only "fencing is anaerobic" produces exactly the training error documented among coaches below.

V̇O₂max is modest, and that is the finding. Elite fencers sit around 42–50 mL/kg/min — measured at 42.5 ± 5.6 in men and 34.4 ± 3.2 in women during actual foil fighting. Aerobic capacity in fencing is a tolerance and recovery qualifier, not a performance discriminator.

The direct-elimination step-up is real. Compared with pools: heart rate 170 ± 10 versus 163 ± 13 bpm; distance covered 459.9 ± 117.7 m versus 162.6 ± 74.2 m; perceived exertion markedly higher. Across a simulated full competition day, average heart rate sat at 88–91% of age-predicted maximum in every DE bout while distance per fight fell from 886 m to 598 m — a fatigue signature within the day.

And the interval structure differs sharply by weapon: mean work time per exchange 17.9 s at épée, 5.8 s at foil, 1.7 s at sabre, with work-to-rest around 1:0.9, 1:2.6 and 1:9.2.

Which qualities matter

Ranked by the strength of the evidence behind them:

• Reactive strength (RSI) — well evidenced. Elite Polish national-squad fencers 1.50 m/s versus 0.85 for non-elite.
• Horizontal lower-body power (standing broad jump) — the only causal S&C evidence in the sport. A training group improved Repeat Lunge Ability from 15.80 to 14.90 s with concurrent gains in broad jump and change-of-direction speed. The authors' recommendation: jump training and plyometrics designed to enhance horizontal propulsion.
• Change-of-direction speed — well evidenced. Correlates with Repeat Lunge Ability at r = 0.70.
• Rear-leg knee-extensor power — moderate. Elite lungers show higher rear-knee peak power, moment and range.
• Aerobic capacity — well evidenced as a tolerance qualifier. The Fencing Endurance Test is the single strongest discriminator of elite status found (15.0 versus 12.4 min).
• Maximal strength — thin. Trainable, but transfer to fencing performance is unproven.
• Trunk anti-rotation — no evidence. No fencing study was found.

A notable negative: the 5 m sprint does not discriminate elite from non-elite fencers (1.10 versus 1.09 s). Short linear sprint is a poor fencing test — which matters, because it is a common one.

Coach belief diverges from the evidence, and it is documented. A Delphi study of 26 international épée coaches with Olympic-level experience ranked agility (mean rank 2.9) and reaction time (3.3) highest, and strength lowest (7.6). The authors' own conclusion is that the undervaluation of foundational qualities such as strength and aerobic fitness shows a partial disconnect with the sport-science evidence. The same group did reach consensus on two footwork points: teach stepping first and introduce bouncing later, and start the guard static and comfortable before adding bounce.

Is conditioning weapon-specific?

The largest test says no. Across 79 national-standard fencers, on lower-body power, reactive strength, change-of-direction speed and repeat lunge ability, there was no significant main effect for weapon — but a significant effect for sex. The authors' conclusion is blunt: épée, foil and sabre fencers do not require a weapon-specific approach to strength and conditioning, and each fencer should target their own weakest area rather than what they feel best represents their weapon.

One dissenting dataset deserves to be shown rather than harmonised away: in 45 elite female fencers, épée advance-lunge velocity was higher than foil and sabre, and sabre showed better early rate of force development than épée, leading those authors to argue for differentiated programming. Smaller sample, one sex.

The workable reconciliation: the gym menu is common across weapons; the conditioning interval structure is not. Run a shared strength and power battery, and a weapon-specific work-to-rest template.

Asymmetry, without the myths

Morphological asymmetry in fencers is real and consistent. In male Olympians, femoral cortical thickness was 5.4% and 13.8% greater on the lead side at two shaft levels, trabecular bone density 54% greater, and thigh muscle area 12.2% greater. Fencers' dominant forearm cross-sectional area is larger than controls' with higher maximal isometric force.

Two qualifiers change the coaching implication entirely. There is no difference in the stress ratio — force divided by cross-sectional area — so the asymmetry is hypertrophy, not altered muscle quality. And forward-leg muscle mass is higher on CT while fibre composition is not different between legs. It is size, not tissue type.

Functional asymmetry flips to null. Isokinetic knee peak torque shows no significant difference between dominant and non-dominant legs. A study of 38 elite young fencers across anthropometry, unilateral vertical and horizontal jumping, isokinetic strength, lunge and step-lunge found no significant bilateral asymmetries at all. Elite inter-limb jump asymmetry of 9.3 ± 8% sits inside the normal range for field and court sports.

Does asymmetry cause injury? Unproven. The load exposure certainly is asymmetric — of 1,044 lower-extremity injuries at INSEP, 60.6% were front limb, and of 269 upper-extremity injuries, 96.3% were the dominant arm. But the study that reports those numbers says explicitly that it was not designed to test such a correlation, and no prospective study has tested measured asymmetry as an injury risk factor in fencers. A 60/40 injury split reflects which limb absorbs the load. Contralateral and compensatory training is recommended universally in coaching texts and has never been tested in fencing — ship it as practitioner consensus, labelled.

One federation does publish a usable threshold: the DFB's anterior reach test flags a left-right difference of more than 4 cm as an elevated lower-limb injury risk.

Injury: two denominators

Both numbers below are true, and any knowledge base that gives one without the other misleads.

Competition time-loss risk is exceptionally low. Across five prospective years and 637,776 athlete-exposures at FIE events, the rate was 0.28 injuries per 1,000 athlete-exposures; a parallel five-year study of US national events across ages 8 to 70+ found 0.3. Sprains 40.8% and strains 20.1%; lower extremity 72.4%; ankle sprain is the single most common injury at 25.3%; median time loss four weeks. Men carried a relative risk of 1.42. By weapon, épée was safest — relative risk 0.52 against foil and 0.47 against sabre.

Total injury burden is high. Seven retrospective years at INSEP, 117 elite French fencers training about 24 h a week: 1,470 injuries, 2.55 per fencer per year, 2.41 per 1,000 hours. Lower extremity 71%, upper extremity 18.4%, trunk 10.4%. Thigh 20.3%, knee 18.8%, ankle 11.2%, hand and wrist 10.5%. Women showed more pelvis and hip injury (10% versus 5.3%). A separate prospective year using IOC forms found 3.22 per 1,000 training hours, with overuse outnumbering acute injury about four to one. A survey of adult competitive fencers found 89.1% reporting at least one injury.

In young fencers the picture is sharper still: mean age at injury 14.6 years, 73% lower extremity (knee 49%, ankle 16%, hip 11%), 77% of injuries in athletes aged 13 or over, and extensor-mechanism dysfunction — essentially patellofemoral pain — as the single most common diagnosis. 80% were managed with physiotherapy and only 5% required surgery.

Fencing is a low-catastrophe, high-attrition sport. And the dose-response finding is counter-intuitive: cumulative years of fencing shows a significant non-linear association with injury, while age at starting does not. That argues for managing career-total asymmetric load rather than merely delaying the start age.

What prevention evidence exists

What has to be transferred, and how well it fits the fencing target — overuse-dominant, lower-limb dominant, front-leg dominant, concentrated at knee, thigh and ankle, plus a dominant-arm upper-limb cluster:

• Ankle proprioception and balance — the strongest case. Ankle is a top-three fencing injury site everywhere and the largest single time-loss category. Across seven moderate-to-high-quality trials and 3,726 participants, balance training gave a relative risk of 0.65; an eight-week unsupervised home programme in a 522-athlete trial reduced recurrences from 33% to 22% — a number needed to treat of 9.
• Hip and adductor strengthening — plausible, untested in fencing. Female fencers show elevated pelvis and hip injury. The Copenhagen adduction exercise, used as a single exercise in a cluster trial across 35 teams, reduced seasonal groin-problem prevalence from 21.3% to 13.5% — a 41% risk reduction.
• Hip-and-knee exercise for the front knee — transfer as management, not prevention. The consensus recommending combined hip and knee exercise is a statement on treating patellofemoral pain, and reports no effect sizes.
• The Nordic hamstring — the weakest case despite being the most cited. Restricted to genuinely randomised trials, the protective effect is inconclusive; and the mechanism mismatch is severe, since that evidence targets acute high-speed-running strains and fencing is overuse-dominant with no comparable sprinting exposure.

Four honest limits on any transfer: the mechanism mismatch; the asymmetry problem — football programmes are bilateral by design, while fencing loads the front leg and the dominant arm; the fact that 18.4% of elite fencing injuries are upper extremity, dominated by dominant-arm overuse, and no established programme contains anything for it; and compliance, which is poor even in the sports where the programmes were developed.

Load monitoring

Ship session RPE. It is validated in fencing: across a full season of 67 training sessions and three competitions with elite foilists, session-RPE showed a coefficient of variation of 6.0% and correlated with heart-rate-based load at r = 0.84 to 0.98 within individual fencers. The formula is trivial — load = duration in minutes × session RPE — and it is the most directly implementable finding in the whole evidence base.

Youth: what federations actually say

The Swiss J+S material for 5–10s contains no loaded-strength content at all. The entire physical content is running games, throwing and catching, jumping over objects, rope skipping, ball coordination and fencing footwork — and that absence is the position for the age. Note carefully: no J+S document was found prohibiting strength training; do not attribute a ban to them.

The clearest federation statement on the resistance-training question comes from the DFB, and it is criterion-based, not age-based: before the squat is performed with resistance in athletic training, it must be ensured that clean movement technique is present, or the joints, tendons and muscles involved risk structural damage. The DFB names no age. It also insists on not anticipating adult content in young athletes, and on taking biological maturity into account alongside chronological age and training experience.

The international consensus on youth resistance training says the same in different words: irrespective of chronological age, a child engaging in resistance training should be emotionally mature enough to accept and follow directions and possess competent balance and postural control. It is also explicit that fears that resistance training would injure the growth plates are not supported by scientific reports or clinical observation. Dosage: 1–2 sets at 60% of one-rep-max or less for the untrained, progressing to 2–4 sets of 6–12 repetitions at 80% or less, on 2–3 non-consecutive days a week.

The monitoring rule that matters in practice: children training more hours per week than their age, or more than 16 hours per week in intense training, and who are specialised, should be closely monitored for burnout, overuse injury and performance decrement. Its empirical basis found that athletes training weekly hours in excess of their age, or twice as many hours as their free play, were significantly more likely to be injured.

Given that 77% of paediatric fencing injuries occur at 13 or over and patellofemoral pain is the dominant diagnosis, front-knee capacity and landing mechanics are the defensible youth priority. And one measurement caution: maturity offset estimated from the standard equations is acceptable for describing a group and unsafe for classifying an individual — the prediction runs the wrong way, placing early maturers later and late maturers earlier than they observably are.

The one published youth mobility exercise found across three federations is worth naming because it targets exactly the right tissue: from a squatting lateral lunge, transfer the trunk from one supporting leg to the other, with the feet staying in place. The FFE prints its purpose — stretch the adductors and hamstrings and mobilise the hip joint.

Heat, kit and hydration

This is where the sport's own equipment becomes the problem, and it is measured. In well-trained épéeists fencing pools then direct eliminations in a July hall with no air conditioning:

• Post-fight gastrointestinal temperature rose across the pool from 37.9 to 38.7 °C.
• Pre-fight temperature exceeded 38.2 °C from the second DE onward and did not return toward baseline between fights.
• Post-fight temperature stayed above 38.6 °C from the second DE, with some fencers exceeding 39.0 °C.
• Mask temperature rose about 1.1 °C across the direct eliminations — the largest effect in the entire study.
• Thermal sensation stayed at "very hot" throughout and did not fall even as skin temperature fell.

The mechanism is the ensemble itself — jacket, plastron, breeches, long socks, glove, chest guard and mask, with a metallic lamé on top at foil and sabre. This protective clothing limits all heat-loss mechanisms available to the fencer, and fencers anecdotally do not remove it between fights. The policy gap the authors name is real: fencing is an indoor sport, is therefore not covered by the IOC heat consensus, and the FIE has no heat policy. Scope limit worth stating: that study was seven male épéeists, and foil and sabre lamé are untested and by the same logic likely worse.

Fencers arrive already dehydrated. In a national team measured in winter, 66.7% were dehydrated before training and 37% seriously so. Sweat rates were 938 ± 251 g/h in men versus 506 ± 92 in women, and 1,136 ± 156 g/h at foil versus 796 ± 207 at épée. There is a direct contradiction in the literature here — one review states that sweat rates in fencing are modest and that fencers have ample chance to rehydrate. That is inconsistent with the measurements above, and its supporting citations predate them. Do not propagate "sweat rates are modest."

Transferable guidance: avoid hypohydration beyond 2% of body mass; monitor weight, urine and thirst, keeping body-mass change under 1–2%. For sports where protective equipment reduces sweat evaporation, permit a progressive heat-adaptation period with minimal equipment coverage over several days before the full ensemble is worn. And the medical rule for a heat casualty: cool first, transport second.

One more caution about caffeine, because fencing has an unusually clear dose-response: reaction time improved significantly at 1.5, 3.0 and 6.0 mg/kg but the combined reaction-and-accuracy effect deteriorated at 7.5 mg/kg. Meanwhile 94.1% of surveyed fencers habitually consume caffeine, only 25.7% had received any education about it, and 35% reported symptoms of caffeine toxicity. That gap is a coaching-education opportunity, not a supplementation recommendation.

Senior conditioning: read the label

The published fencing-specific conditioning drill is a 2–4–2 m shuttle run performed with fencing footwork, facing forward throughout: both feet return behind the start line after each shuttle, and part of the lead foot must cross the 2 m and 4 m lines. Programmed as high-intensity intervals, the work-to-rest prescriptions are 1:1 for men's épée (15 s work), 1:3 for foil (5 s) and 1:5 for sabre (3 s), in three sets with a minute between them.

Read that table carefully. The authors flag the sabre row and both female foil and sabre rows as hypothesised, not measured. The épée and men's foil rows rest on data; the rest is inference by the same authors, and should be presented as such.

Two further points from the same source are worth carrying. The asymmetry progression: completing the same drills in the non-dominant stance is reported anecdotally to help offset the visible muscular imbalance between legs. And an explicit prohibition: fencers should be critical of traditional long slow distance running, because low-intensity high-volume work negatively alters force-velocity relationships.

Finally, the one periodisation finding in fencing is a warning rather than a model: in international épéeists tested mid-off-season and mid-in-season, V̇O₂max was significantly lower in-season, and isokinetic knee-extensor peak torque was significantly lower in-season in both legs at all velocities. Bilateral, so it is a dose problem rather than an asymmetry problem. The study is old and tiny and has never been replicated — but it is the only seasonal data fencing has, and no periodisation model has ever been validated in the sport. Any template you use is expert opinion; label it that way.

Sources

Holmes & Bottoms (2026), PLoS One 21(6) — scoping review, 445 studies · Yang et al. (2022) · Roi & Bianchedi (2008), Sports Med 38(6) · Bottoms et al. (2023), PLoS One 18(2) · Turner et al. (2016), J Strength Cond Res 30(8) and 30(11); (2017), J Strength Cond Res 31(6) and IJSPP 12(1); (2013), Strength Cond J 35(1) · Hekiert et al. (2025), Life 15(10) · Cree et al. (2026) Delphi · Harmer (2008, 2019) · Gondouin et al. (2025), Front Sports Act Living 7 · Cross et al. (2024) · Schiftan et al. (2015) · Harøy et al. (2019) · Impellizzeri et al. (2020, 2021) · Dalen-Lorentsen et al. (2021) · Lloyd et al. (2014), BJSM 48(7) · Oates, Price & Bottoms (2024), Temperature 11(4) · Eda et al. (2022) · DFB/IAT Athletiktest manual v2 (2023) · FFE, Escrime à l'école primaire and Recueil pédagogique · J+S-Kids Escrime (BASPO). Verified July 2026.