Tactics

The tactical wheel at foil

Foil's strategic logic is generated by two articles. t.83.2.d makes every bent-arm step or feint a preparation, and therefore attackable. t.84.3 makes continuous crossing steps forward a preparation over which any simple attack has priority. Together they make the attack in preparation structurally powerful and kill the running advance.

The canonical six-node cycle, each action beaten by the next:

1. Simple attack — beaten by the parry-riposte.
2. Parry-riposte — beaten by the compound attack, whose feint draws the parry. But if the defender finds the blade on the feint, they regain the riposte.
3. Compound attack — beaten by the attack in preparation, or by a stop hit that precedes the final movement by one period of fencing time.
4. Attack in preparation — beaten by counter-time: provoke it, parry it, riposte.
5. Counter-time — beaten by the feint in tempo: a false counter-attack that deceives the counter-time parry.
6. Feint in tempo — beaten by the simple attack, and the cycle closes.

Second intention layers on top: the first action is a deliberate loss designed to buy the opponent's response.

The modern long attack exists because of t.83.2.c — the arm need only be straightening before the end of the step forward, not before it begins. That legalises a long approach step with the arm extending late, converting into a lunge or flèche, which the referee reads as one attack. The counter is the attack on the step itself.

On the remise. Since the 2004–2005 apparatus change shortened the foil lockout to 300 ms, a remise made immediately after a parry lands inside the window far more often — so the box now shows two lights where it once showed one. But priority still annuls it if the riposte is immediate, simple, and made in one period of fencing time without withdrawing the arm. The remise's value at foil is therefore not that it scores, but that it forces the riposte to be immediate, simple and arm-forward — it removes the defender's compound and delayed options. Conversely, a defender who pauses after the parry forfeits the phrase to the remise. (The rule text is certain; this causal reading is coaching consensus, not a measured finding.)

One more foil-specific asset: an off-target hit stops the phrase and annuls whatever follows. For a fencer who is losing a phrase, that is a free reset.

Épée: the near-target game

Three structural facts drive everything:

• No priority. Both lights means a hit against each — so the double is a resource, not an error.
• The whole body is target, and the nearest part of the opponent is the hand, forearm or front foot. Épée is a near-target game before it is a body-hit game.
• The double window is 40–50 ms, roughly one seventh of foil's. A hand counter-attack landing 40 ms before a body hit yields a double; landing about 60 ms early it yields a single. The whole hand game is a fight over about 20 milliseconds.

The four distance situations — a usable decision tree, after Gary Copeland:

1. You advance, they stand still. Hit the near target — hand or arm. If they defend without retreating, close and take the deep target. Deceives, opposition thrusts, transports.
2. You advance, they retreat. Re-attack the near target rather than pursue the deep one. Beats and presses.
3. You retreat, they pursue. Attack the near target while opening distance — pulling, not being pushed. Counter-attacks, remises, invitations.
4. You attack the near target, they close. Stay balanced and score deep via parry-riposte, blade control or counter-riposte.

The general target rule: attack the target closest to your point before proceeding to a deeper one. And the specific version for a shorter fencer against a taller one: respect the danger zone — make simple attacks and counter-attacks to thumb, wrist and distal forearm without blade contact, and reach any deeper target only with a compound action or while controlling the blade with a bind, croisé or envelopment.

What the elite data actually shows. In 1,840 scoring events across 45 women's épée direct-elimination matches at three World Championships, the technique mix was attack 7.17%, counter-attack 5.04%, riposte 2.42%, other 2.04% of scoring actions. Single hits outnumbered doubles 10.90% to 5.76%. Crucially, winners and losers took identical numbers of doubles (5.76% each) — winners won by winning the single-light game (13.19% versus 8.61%), not the double game. And winners scored most in the middle area of their own half (21.59%), while losers spread evenly across all four central zones: winners lure the opponent onto their own side of the piste.

Across weapons the modal elite exchange is offensive action against counter-offensive action, not attack against parry-riposte. Parry-riposte is a minority scoring route at elite épée. The syllabus implication is direct: the counter-attack, and the counter-time that answers it, deserve more lesson time than the classical parry-riposte curriculum gives them.

Sabre: the four-metre zone

Priority applies, the target is large and includes the arms, and the weapon cuts — hits are easy to land and hard to avoid. Everything therefore rests on who has priority, which the referee assigns. Two rules bound the game: the flèche and any crossing step are forbidden, and the hit must arrive at the latest when the front foot lands. The attack is a bounded action with a hard deadline.

The fencers start on lines 2 m either side of centre, so they begin 4 m apart — and that gap is the entire battlefield. Sabre actions concentrate in the central four metres of the piste: 72% for men and 67% for women in elite competition analysis.

The reading problem. Because both fencers start 4 m apart and both must come forward, the referee's decision reduces to three ordered questions: who launched first; if simultaneous, who had the better distance; and whether both hit despite the simultaneity. Club analytics on the cost of getting this wrong: simultaneous attacks made up 49% of attempted actions but succeeded only 39% of the time, while fast preparations succeeded 57%, medium tempo 47% and slow 50% — it is better to go clearly fast or clearly slow than to sit in the middle. (Practitioner data, not peer-reviewed.)

The two macro-strategies: the march attack, a continuous forward preparation converting into a flunge, aiming to own priority from the start; and pulling distance, retreating so the march falls short and expires against the front-foot deadline, then taking the riposte or the attack on the recovery. The march is beaten by pulling distance; pulling distance is beaten by the chase — an advance-step then a second advance-lunge.

The four-metre wheel, pure rock-paper-scissors: the short attack (advance-lunge) beats the chase and loses to defence; the chase beats defence and loses to the short attack; defence — advance then retreat, making them fall short — beats the short attack and loses to the chase.

And the structural consequence of the lockout's width: at 170 ms the box shows more double lights, and therefore pushes more decisions onto the referee. Sabre's referee-dependence is a design feature, not an officiating failure. (The 170 ms figure is certain; the widely repeated "widened from 120 ms in 2016" history could not be confirmed against a rulebook of that period, so treat the date as unverified.)

Score state changes what wins

The strongest published finding in fencing notational analysis is that score state conditions technique choice. In elite women's épée, winners while leading used attack 19.70%, counter-attack 12.51%, riposte 6.46% and other 4.62% of the time; while trailing, the riposte became their least-used action. Losers, by contrast, showed no significant technique preference while leading or level, and only diverged when trailing — attacking and counter-attacking heavily.

There is also a round effect: the second of the three direct-elimination periods is the highest-scoring phase (6.86%, 9.43% and 8.72% of scoring actions across rounds 1, 2 and 3). Winners attacked in rounds 1 and 2 and counter-attacked in round 3. The one-minute rest between periods is the only guaranteed coaching window in an individual bout — and the break between periods 2 and 3 is precisely where that switch should be briefed.

Sabre gets a scheduling weapon out of this: because the first direct-elimination period ends at 3 minutes or 8 hits, a sabreur can reach the coaching break faster by scoring. A sabreur leading 8–3 gets a break they chose; one trailing 3–8 gets one they did not.

Endgames and the priority minute

• Épée pool at 4–4. The fencers fence for a deciding hit up to the time limit and any double hit is not counted — the fencers stay where they are. The double is free, so a fencer with a superior double game has nothing to lose by throwing simultaneous actions repeatedly. This is the classic pool endgame.
• The priority minute. Level at time, in pools, direct elimination or team, the fencers fence one extra minute for a deciding hit, and the referee draws lots beforehand to decide the winner if it is still level. The fencer with priority wins by not being hit: maximum distance, high defensive readiness, no initiative. The fencer without it must score, must take single-light risk, and at épée must specifically avoid the double, which does not resolve the minute in their favour.
• Non-combativity does not apply in the extra minute. The rule says so explicitly — so the fencer with priority can legally stall the full 60 seconds. There is no passivity penalty to save you.
• How often this matters: in 45 elite women's épée direct-elimination matches from the table of 16 to the final, 11 went to sudden death — 24.4%. Roughly one elite épée DE in four is decided in or after the priority minute. That is a very high return on rehearsing the coin-toss endgame.
• At 14–14 (and 44–44 in teams) no P-card may be awarded — the passivity penalty is switched off for both fencers, so that exchange is played with no clock pressure other than the period clock.

The exploitable asymmetry in the passivity rule. P-cards fall on the fencer or team with the lower score when scores are unequal, so a leading fencer is rule-sanctioned to be passive — the penalty lands on the opponent. It is the single most exploitable feature of the rule, and it is exactly why the trailing fencer must generate action. Note that this regime changes on 1 September 2026: the P-yellow disappears and the ladder becomes P-red on the first minute, P-black on the second.

Breaks. The five-minute medical break is frequently used as a momentum-breaker, and the rules criminalise that: any fencer who improperly causes or prolongs interruptions is penalised, and an unjustified demand for a medical break is a second-group offence. The legitimate reset is an equipment change — weapon, body wire or lamé. Coach the equipment check; do not coach the false injury.

Video appeals are a scarce resource with an asymmetric payoff: one appeal per bout in pools, two in direct elimination, and an upheld appeal is retained rather than consumed. The referee may consult the monitor at any time on his own initiative, and must use video before deciding the decisive hit at equal scores. The rational policy is to appeal early and aggressively on any genuinely ambiguous priority call, and to hold the last appeal for a game-state-critical touch. Note that no published study quantifies appeal success rates in fencing.

Team relay

Nine bouts to 45, cumulative in steps of five, maximum three minutes per bout. Three fencers plus one reserve; each of the three meets each opponent exactly once. The key structural fact is what happens when a leg runs out of time: the next pair takes up the score where it was left off and fences to their own cap. An unfinished leg does not roll a deficit forward as extra touches — it rolls forward as extra time pressure.

Three consequences follow directly:

• You cannot win back more than five touches in a leg from an even start. A nine-touch deficit mathematically requires at least two dominant legs.
• When losing a leg, let the clock run into the cap. Holding the opponent under five in that leg is worth more than chasing it.
• A trailing team cannot defensively grind, because the passivity cards fall on the team with the lower score. The rule forces the trailing team to attack.

Line-up roles (a practitioner framework, not data): the closer fences bouts 1, 5 and 9 — your strongest fencer, who opens and closes the match; the brick wall fences 1, 4 and 8 and meets the opponent's closer, so needs defence and composure; the third fencer covers the middle of the match and adapts. The structural argument for the anchor is solid: bout 9 is the only leg with no subsequent leg to correct it, and the only one reachable by the tie-breaking priority minute — so the anchor must be your best endgame fencer, which is not necessarily your highest-ranked one.

The reserve is a single-use resource that doubles as insurance. One substitution per match, for tactical or medical reasons — and after a P-black the reserve may only enter if it has not already been used. Because of that, the expected cost of substituting early is higher than the raw touch differential suggests.

Left-handers

Two independent findings converge. Left-handers are over-represented in fencing relative to the general population baseline of about 10.6% — foil 22–27%, épée 17–21%, sabre 12–13% across 15,099 elite athletes — and the over-representation increases with rank. Historical finalist counts are far higher still (41–50% of World Championship finalists in some years), which is the rank gradient rather than a contradiction.

The weapon ordering is mechanistically sensible: foil's small trunk target and priority conventions make the mirrored line structure maximally disruptive, while sabre's very short exchanges and large target reduce the value of the asymmetry.

The club implication is concrete. Right-handers at a club with roughly 10% left-handers get about a fifth of the left-handed practice they will meet at national level. Deliberate sparring allocation, and a left-handed lesson in the individual-lesson rotation, are cheap fixes for a real gap.

Scouting a club can sustain

Elite programmes chart up to 38 fields per touch and budget about 1.5 hours of analysis per bout, aiming for a one-page report readable in two minutes — with roughly 80 touches needed to understand an opponent's general approach and 150+ for real familiarity. No club can sustain that.

A nine-field chart, taking about ten seconds a touch and chartable live on a phone, gets most of the value:

• Touch number and score before
• Period — 1, 2, 3 or priority minute
• Who scored — us, them, or double
• Initiative — we pushed, they pushed, neither
• Scoring action — attack, attack in preparation, counter-attack, parry-riposte, counter-riposte, remise, point in line, other
• Zone — our end, our middle, centre, their middle, their end
• Target (épée) — hand and arm, foot and leg, body, mask
• Seconds since the last touch

What falls out of those nine fields: action mix compared against the elite baselines above; the score-state split, which is where the actionable asymmetries live; push/pull ratio and initiative conversion; double rate when leading versus trailing, the single most diagnostic épée metric given that winners and losers take identical numbers of doubles; seconds per touch against a roughly 17–20 s elite baseline, which doubles as a passivity-risk proxy; and a zone heat map — a fencer whose scoring is centred on the opponent's half is fencing the losers' pattern.

One reliability caveat worth designing around: in the published research, agreement between expert coders was near-perfect on the objective fields but lower on piste location, and both coders had 15–20 years of épée experience plus national refereeing certification. Charting reliability degrades fastest on spatial and action-type fields. If parents or junior athletes chart, restrict them to who scored, the score before and the clock; reserve action-type coding for coaches.

What to teach when

• Around 8–9: guard, balance, distance as a game, the salute, safety, "hit and do not get hit". No named action vocabulary, no priority theory, no weapon specialisation.
• Around 10–11: the simple attack, the simple parry-riposte, retreat to avoid. Chained attack-and-defence sequences. Begin naming actions. Not yet: second intention, counter-time, compound attacks beyond one feint, or deliberate double-touch strategy.
• Around 12–13: compound attack with one feint, attack on preparation, choice reactions, bout management, pool administration, refereeing. Not yet: full second-intention chains or clock manipulation.
• Around 14–15: second intention, counter-time, distance-pulling as a strategy, weapon-specific score-state basics — the épée 4–4 rule, the sabre 8-hit period break. Explicitly not priority-minute stalling, medical-break gamesmanship, or anything that trains the athlete to exploit the passivity asymmetry.
• 16 and up: the full wheel per weapon, opponent-specific game plans, the priority minute with and without priority, non-combativity as a rule to be managed, video-appeal policy.

The reasoning for the "not yet" column is worth stating, because the counter-attack is the tempting exception. It works immediately against children who march with a bent arm — which is exactly the trap. It is rewarded by immature opposition and punished by mature opposition through counter-time, and it prevents the athlete developing the attack that the counter-time game presupposes. The defensible ordering is attack → parry-riposte → attack in preparation → counter-attack → counter-time: teach each action only after the action it is designed to beat.

Sources

FIE Technical Rules, 18 December 2025 (t.38–t.45, t.59–t.62, t.83–t.89, t.101–t.106, t.124) · Hsu & Chen (2026), J Hum Kinet 100:151–164 — 1,840 scoring events · Aquili et al. (2013), J Strength Cond Res 27(3) · Tarragó, Iglesias & Ruiz-Sanchís (2018) · Simon, Loffing & Frasnelli (2025), R Soc Open Sci 12(9) — 15,099 elite athletes · Sydney Sabre tactical notes (practitioner) · Coaches Compendium four-situations framework, after Gary Copeland · Lehfeldt Evaluation Framework (practitioner). Verified July 2026.