15Training and Performance Development

15.1Training blocks and performance objectives

A training block is a bounded period, defined by a calendar and a single dominant objective, whose success can be judged at its end by a measurement agreed at its start. The objective is written as a number the crew can log, for example the mean distance lost through a tack in a stated wind band, or the VMG deficit to the reference boat upwind in flat water, and the block is closed when the number reaches the target and holds across the conditions the block was defined for, not when it is reached once. An objective that cannot be measured produces sessions with no exit criterion.

The mechanism that makes a block work is the chain from objective to content to measurement. The objective selects the drills, the drills demand certain conditions, the conditions decide where and when the block is sailed, and the measurement decides whether the block has finished. Objectives are ranked by expected race-time gain, and published evidence gives the ranking its starting point. The regression of every tracked 49erFX race from August 2016 to August 2021 on 38 separate skills found that VMG on the first beat carried a coefficient of -2.0, "almost ten times that of any other skill", and that "having a good start and an efficient first beat is the most important combination to sailing a shorter race" S37. The GPS analysis of Olympic-class regattas concluded that training "should prioritise the upwind course and aim to enhance manoeuvring efficiency, more prominently in strong wind conditions" S30. A block calendar that does not weight upwind speed, starting and manoeuvres ahead of everything else argues against the only quantitative evidence available.

The calendar is fixed at the far end by the event. World Sailing has published LA28 sailing dates of 16 to 28 July 2028, with the 49er qualification series starting on 25 July and the final on 28 July S7; the format approved by the World Sailing Council on 10 December 2025 provides a three-day qualification series and a finals day of two single-point races for the top ten S8. Blocks are laid backwards from that date. The last block before a target regatta is a consolidation block in which nothing new is introduced and the measured skills are held; the earlier blocks carry the changes. A change to a setting, a technique or a decision rule is scheduled into a block with enough sessions after it to measure whether it transferred, because a change without a measured transfer is a risk carried into racing.

On the water each session inherits the block's objective, and the crew's log records the objective, the wind band, the sea state, the drill, the measurement and the value obtained. The testable standard for a block is a log in which the objective's value can be read at the start and at the end across every band sailed, and an entry value for the next block that is the exit value of this one. The block sheet carries crew value against each metric rather than a figure borrowed from another boat.

15.2Deliberate practice for speed, handling, and tactics

Deliberate practice is repetition with a defined target, feedback close enough to the attempt for the correction to be applied on the next attempt, and a difficulty set at the edge of current ability so that each repetition is an attempt at something not yet reliable. Volume without these three conditions consolidates the present technique, including its faults. The principle applies to speed, handling and tactics alike, but the three differ in how fast and how clean the feedback is, and that difference dictates how each is practised.

Speed feedback is fastest and cleanest when a reference boat is alongside, because the gain or loss over a minute is visible and the wind is common to both boats. The practice unit is a line-up in which one variable is changed and everything else held, followed by a second line-up with the variable restored. Gladstone's three sources of upwind power, angle of attack, depth and twist S20, give the variable list; the crew's base settings for each band are crew value, and the practice tests one departure from base at a time. Handling feedback comes from the loss measured across a manoeuvre, which a logger expresses as distance surrendered against the pre-manoeuvre VMG line. The Ovington manual treats tacking, gybing, two-sail technique, gennaker sets and drops and rounding onto the wind as separate skills S12, and deliberate practice treats each as separately measurable and separately drilled, with the manoeuvre divided into phases so that a loss is attributed to the entry, the crossing or the exit.

Tactical feedback is slowest and noisiest, because the outcome of a tactical decision depends on the wind and on other boats, and a correct decision can lose while a wrong one gains. Tactical practice therefore uses a process criterion, whether the decision matched the pre-agreed rule given the evidence available, rather than an outcome criterion. The drill content is the reading of the wind: calling a shift from the compass before the sails confirm it, calling the phase of the oscillation, calling which side the pressure lies on, then checking each call against what arrives. Bethwaite's measurements give the rhythm to be learned: wind speed rising and falling by about 7 per cent every 6 to 12 seconds, and directional oscillations with periods from about 3 minutes in heavy air to 10 minutes or more in light-medium air S50.

The crew uses the method by counting. Each drill has a repetition count and a loss or gain recorded against every repetition, and the drill stops when quality falls for consecutive repetitions, because repetitions performed with a degraded movement pattern train the degraded pattern. The testable standard is a log in which the trend across sessions in a given band is visible for each skill, and in which a skill with a flat trend is either re-drilled with a different cue or moved into simulation to find out whether the plateau is real.

15.3Isolated drills versus race simulation

An isolated drill removes variables so that the cause of a loss can be attributed; a race simulation restores them so that the crew learns whether the skill survives sequence, load and the presence of other boats. The two are stages rather than alternatives, and a skill moves between them by rule. Early in a block the ratio favours drills, because a change is being installed and its effect must be seen cleanly; late in a block it favours simulation, because the question has become whether the change holds when the crew's attention is on the fleet.

The mechanism behind the distinction is attribution. In a tacking drill the crew chooses when to tack, the water is uniform and the loss can be assigned to the manoeuvre alone. In a race the same tack is forced by a layline, the wave set is not chosen, a boat to leeward limits the exit angle and the crew's attention is divided between the tack and the next decision. The loss in the race is the sum of several causes, and without data it cannot be decomposed. Skills also interfere with one another: a gybe timed to a chosen moment in a drill uses different cues from a gybe forced by the approach of the gate, and the timing of the crew's hands has to be re-learned against the new cue. Simulation exposes this; drills do not.

Simulation content is chosen from the parts of the race that data identify as decisive. Starting is built with a committee boat and a pin, short lines and many sequences, with the line set at a known bias so that the crew practises the bias calculation, for which "for every 5 degrees of upwind bias to one end, that end will be upwind by a distance equal to about 1/8th of the length of the line" S65. Tracking data from the 2015 49er Europeans, which showed a minimum gap to the line of 10 m and a starboard-tack advantage of under one second in time to the line after the gun measured on speed over ground S16, give the scale of the gap and the acceleration rather than a standard to be matched. The 49erFX data study found extra manoeuvres on the first and last legs associated with shorter race times S37, which indicates that the crowded legs, where manoeuvres are forced by other boats, must be simulated with enough boats to force them; a simulation with two boats does not reproduce a first-beat crossing situation.

The decision rule for moving a skill is measurable. A skill enters simulation when its isolated loss metric has stabilised across sessions in the relevant band; it returns to drills when simulation shows the loss rising under load, and the drill is then modified to include the cue that broke it, for example a tack executed on a called mark rather than at the crew's choice. The testable standard is a training log in which each skill's current location, drill or simulation, is stated for the block, with the metric that justified it.

15.4Wind-range and sea-state-specific training

The boat changes character across the wind range: in light air it is a displacement hull steered largely by weight, in the middle of the range it planes with both crew on the wire, and at the top the problem becomes control. A skill demonstrated in one band has not been demonstrated in the others, so the training plan is organised by band and by sea state, and every logged measurement carries the band it was obtained in. Comparisons across bands are not made.

The set-up boundaries in the class tuning guide mark where the boat's requirements change. The jib sheet sits in the middle hole and the track pins in the innermost holes for winds up to about 12 knots and flat water, and cap shroud tension holds its baseline up to about 11 to 12 knots and is "progressively tighter to flatten the mainsail for progressively stronger winds" S13. Around that boundary the crew's task changes from generating power to shedding it, and the technique for each side of it is trained separately. At the bottom of the range, World Sailing's rule 42 guide for the class notes that breaches "do occur in the 4-8 knot wind range" S18, so light-air training carries a body-movement discipline as an explicit objective. At the top, Bithell describes the risk of pitchpoling when the boat ploughs into the back of a wave at full speed and the response of oversheeting the gennaker to slow it S29, a technique that cannot be practised in any other conditions. The GPS study of the class at one Olympic-class regatta covered 5.5 to 11.6 knots S30 and concluded that manoeuvring efficiency matters more prominently in strong wind S30; the crew's own losses by band crew value decide how the strong-wind hours are spent.

Sea state is a second axis, not a consequence of the first. Leftover waves can run at an angle to a new breeze so that one tack is into the waves and the other across them, and each tack then needs its own mode S77; wind against current raises wave height and sea disturbance above normal, and wind with current lowers them S76. A wind band in flat water and the same band in a short sea are two entries in the log, with different targets for heel, trim and the timing of manoeuvres.

The crew uses the two axes as a coverage map. Hours are tallied by band and sea state, and a cell with few hours before an event is a known weakness whatever the crew's impression of it. When the forecast does not deliver the planned band, the session objective is replaced by the objective belonging to the band that arrives, rather than running the planned drill in the wrong conditions and logging a result that does not belong to it. The testable standard is a log from which the hours and the metric values can be read for every cell, and a pre-event review that names the cells with the thinnest coverage.

15.5Training-partner benchmarking

Boat speed has no absolute measure on the water, because the wind that produces it is never known exactly; it is measured as gain or loss against a reference boat sailing in the same wind at the same moment. Training-partner benchmarking is the procedure that makes that measurement repeatable, and its design is governed by the sources of error it must exclude.

The line-up places two boats on the same tack, bows level, separated laterally by enough to keep each out of the other's disturbed air and by little enough that they sail in the same wind. Dirty air extends to leeward and aft in the direction opposite the leading boat's apparent wind S54, and upwind a boat on or to weather of another's bow disturbs her within 6 to 10 boat lengths, the greater distance in lighter conditions S56, so the separation is set above that and adjusted with the wind. The line-up is held for a fixed time or until a clear gain has appeared, the gain is recorded in lengths, and the boats then swap sides and repeat, because any lateral separation exposes the boats to different wind and wave, and a result that reverses on the swap was a difference in the wind rather than in the boats. Both boats read the compass at the start and end of each line-up: a 5-degree shift changes the relative position of two boats by 12 per cent of their lateral separation S52, so an unlogged shift can produce a false gain larger than the effect being tested.

The reference boat holds its settings fixed while the other changes one variable, and the roles alternate so that neither crew's technique drifts unmeasured. The result is reported in two components, height and speed, because the two modes are not interchangeablewhy: the boat that has climbed higher is ahead on the ladder rung S53 only if it has not paid for the height in speed, and a gain recorded simply as "ahead" cannot be turned into a trim decision. Downwind line-ups are sailed at the same wind angle, the gain is expressed as a VMG difference, and the high-road and low-road choice Bithell describes S29 is tested by having one boat take each and comparing the gate arrival.

The crew uses benchmarking as the primary speed instrument and the logger as the secondary one, because the reference boat cancels the wind that the logger cannot. A session produces a table of variable tested, band, sea state, side, tack, gain in lengths per minute crew value and shift logged; a result is accepted when it repeats on both sides and both tacks. The testable standard is that any base-setting change the crew adopts can be traced to a line-up result that met that repetition condition.

15.6Video review and shared crew analysis

Video shows the crew what it cannot observe while sailing: the sail shapes from outside the boat, the heel and fore-and-aft trim, the position and timing of each body through a manoeuvre, and the interval between a call and the response. Class rule C.5.2 permits camera and sound recording equipment aboard S3, and a coach boat can carry a camera and a microphone, so the record includes both the picture and the crew's words. The value of the footage lies in the question asked of it, and the question is set by the data before the review begins.

Review starts from a loss the logger or the line-up has already identified and finds the corresponding footage, rather than watching a session from the beginning and waiting for something to appear. A tack with a known loss is stepped through phase by phase, with the timing of the helm's turn, the crew's release and the sheet-in read against the crew's own reference timings crew value. Heel is checked against the skiff standard "Flat is FLAT, +/- 5 degrees is acceptable, +/- 10 degrees is only just tolerable" S36. Downwind the Ovington manual's description, that the boat is sailed heeled to windward with the windward chine in the water and that the gennaker is not under-trimmed because the leech will rotate up S12, gives two things the coach-boat angle shows directly and the crew's own view does not: the chine in the water and the leech position. Sail shapes from the coach boat are compared with the settings the crew logged for that moment, and a shape that does not match the intended setting exposes either a control that has slipped or a setting that does not produce the shape it was meant to.

Shared analysis means both crew watch the same footage and each first identifies their own error before commenting on the other's, so that the review does not become a negotiation. Disagreements about what happened are resolved by the footage and the data, not by who steers. Each manoeuvre reviewed yields one agreed change, since changing several elements at once makes the next session's result unattributable, and the change is written with the cue that triggers it and the check that will show it worked.

The crew uses the review to close the loop between measurement and drill: the review produces the next session's drill content, and the next session's logger data show whether the change reduced the loss. Audio is reviewed as strictly as picture, against the communication protocol in section 15.11, for calls made late, calls not acknowledged and words outside the agreed vocabulary. The testable standard is a review log in which every entry has the loss that prompted it, the footage reference, the agreed change and the date on which it was checked.

15.7GPS, compass, heading, and speed-data analysis

Logger data replace impression with record, and the record is only as good as the question defined before the session and the check of data quality performed after it. The class limits what the crew may carry: rule C.5.2 permits timing devices, up to two compasses without functions beyond heading, heading memory and timing, camera and sound recording equipment, and tracking equipment "where permitted by the ICA, Notice of Race and/or Sailing Instructions" S3. The Paris 2024 Notice of Race prohibited voice, radio or data transmissions from the first warning signal to the end of the last race of the day, except in an emergency or with organiser-provided equipment S4, so in racing the device is a recorder and a display, and the analysis happens ashore.

The current loggers publish their capabilities. The Vakaros Atlas 2 uses a 25 Hz dual-band GNSS receiver, logs at 2, 5 or 10 Hz and reports distance-to-line, time-to-line, time-to-burn, shift tracking and VMG S79; the Sailmon MAX records boat speed, heading, heel, pitch, position, augmented true wind and line bias S80. Njord builds polar targets and gain-and-loss comparisons from log files S31. The sample rate matters for manoeuvre analysis: a tack lasts seconds, and the loss through it is resolved only if the logger samples several times a second, so the rate is set before the session and confirmed in the file.

Heading and speed data give the crew more than a track. The difference between the headings on the two tacks is the tacking anglewhy, and its bisector is an estimate of the true wind direction at that moment; the heading held on one tack over a leg is a record of the shifts; boat speed and heading combined with the wind estimate give VMG; and the compass on the boat, restricted to heading and heading memory S3, gives the crew the same shift record in real time. The data do not give a measurement independent of the water: speed and course over ground include the current, so VMG over ground in a tidal venue is not VMG through the water, and the crew records the current before treating a gain as a trim effect. The GPS study of an Olympic-class regatta reported that mean velocity, VMG and distance travelled increased with wind velocity on both legs S30, which is why speed values are compared across sessions within a wind band or not at all.

No published polars for the class were found, so the crew's own logged values are the reference, held as crew value per wind band and updated after each session. The analysis order is fixed: data quality first, meaning sample rate, dropouts and time alignment with the video; then assignment to a band and sea state; then comparison within the band against the crew's table; then attribution of any difference to a logged setting or technique change. The testable standard is a polar table the crew can produce for any band it has sailed, with the date of the last update and the sessions that fed it.

15.8Manoeuvre, leg, and race performance metrics

A metric earns its place if a change in it can be attributed to a decision or a skill; a number that only ranks the boat is a score and does not tell the crew what to train. The metrics are organised at three levels, manoeuvre, leg and race, and each level is read against the level above it, so that a manoeuvre loss is interpreted in the light of the leg that required the manoeuvre.

Manoeuvre metrics describe a single tack, gybe, set, drop or rounding: the distance surrendered against the pre-manoeuvre VMG line over a fixed window, the time to regain target speed, and the number of manoeuvres on each leg. Leg metrics come from the tracking systems used at class events. SAP Sailing Analytics publishes per-boat gap to leader, average speed over ground, distance travelled, number of manoeuvres, leg VMG and cross-track error S78, the last defined by the class as "the extra distance sailed upwind from course middle line on either side" S17. Cross-track error is the metric most easily misread: a large value is the record of a side chosen, and it is right or wrong only in relation to the strategy that chose it, so it is read next to the pre-race plan and the shift log, never alone.

Race metrics weight the levels below them. The 49erFX study of every tracked race from August 2016 to August 2021 broke a race into 38 skills and regressed total race time on them; VMG on the first beat carried a coefficient of -2.0, distance behind the line at the start carried 0.060, and the reading given was that "having a good start and an efficient first beat is the most important combination to sailing a shorter race" S37. The same study found extra manoeuvres on the first and last legs associated with shorter race times S37. The second finding is read with care: a coefficient in a regression on race time is an association across the fleet, and the boats that manoeuvre more on the last leg may be leaders covering rather than boats made faster by manoeuvring. The crew takes from it that the manoeuvres forced on the crowded legs must be cheap, not that manoeuvres should be added. The Journal of Navigation conclusion that training should prioritise the upwind course and manoeuvring efficiency, more prominently in strong wind S30, is consistent with the weighting.

The crew uses the three levels as a diagnostic chain after every race. Places lost are listed, each is assigned to a leg, the leg metric that moved is identified, and the manoeuvre or speed loss beneath it is named; a lost place that cannot be traced to a metric is logged as unexplained rather than given a story. The crew's target values for each metric are crew value by band. The testable standard is a post-race sheet in which every place lost has one attributed cause and one training consequence, and in which the causes accumulate across a regatta into the objectives of the next block.

15.9Physical preparation for power, endurance, and resilience

The boat sets the physical demand, and the demand is defined here so that the crew's physical programme, designed with its own specialists, is built against the right specification. The programme itself is not stated; the specification is, in three parts: the trapeze, the manoeuvre and the series.

The first demand is the trapeze. Righting moment is crew weight applied over the longest available lever S25, and the mast is built to support a combined crew weight of up to 165 kg from its dual trapeze S9. On the wire both crew hold a sustained position in which the load passes from the harness through the trunk to the feet on the gunwale, and the position is never static: the wind speed rises and falls by about 7 per cent every 6 to 12 seconds S50, and the crew answer each change by moving in and out along the lever so that the righting moment tracks the heeling moment and the boat stays flat. Low trapezing dampens roll and keeps the sail plan upright S25, which lengthens the lever and adds to the load. The demand is a long-duration hold with continuous small adjustments, repeated on every upwind leg and, with the gennaker set, on every downwind leg in planing conditions.

The second demand is the manoeuvre. Every tack and gybe is a transition from wire to cockpit to wire, performed against the clock, in which the crew also handles the gennaker or the jib and the helm crosses with the tiller extension and the mainsheet. The Ovington manual lists tacking, gybing, gennaker sets and drops and rounding onto the wind as separate skills S12, and each has its own movement pattern that must be executed at full speed when the boat is at its least stable. The 49erFX data study associated extra manoeuvres on the first and last legs with shorter race times S37, so the tempo of the crowded legs is high and the demand is repeated explosive movement between periods of loaded holding.

The third demand is the series. Paris 2024 scored the Men's Skiff over an opening series of 12 races plus a Medal Race S6, and the format the World Sailing Council approved for LA28 on 10 December 2025 provides a three-day qualification series and a finals day of two single-point races for the top ten S8, with no World Sailing confirmation found after the 31 May 2026 determination date S8. The demand is several races a day on consecutive days, and the last race counts as much as the first, or more where final-day points are not discardable. Endurance in this sense is the preservation of the movement pattern and the decision quality through the final race, and resilience is the same preservation after a capsize, a cold day or a poor result.

The crew converts the specification into a testable standard rather than a training figure: the loss through a manoeuvre in the last race of a day, compared with the first in the same band and sea state, is the measure of whether the physical preparation matches the demand, and the target difference is crew value. Where that difference grows across a regatta, the specification has not been met, and the programme is revised with the specialists who own it.

15.10Psychological readiness and decision-making under pressure

A decision made under time pressure is only as good as the rule prepared for it, because the crew has no capacity in the moment to construct an argument, only to recognise a condition. Readiness is not a state of mind but the completeness of the rule set and the crew's agreement on it before the warning signal. It covers every recurrent decision: tacking on a header, gybing on a lift, holding or bailing from a start position, choosing a gate mark, covering or splitting on the last beat.

Each rule has three parts, the condition, the evidence threshold and the call, and the threshold protects the crew from reacting to noise. The wind's directional oscillations have periods from about 3 minutes in heavy air to 10 minutes or more in light-medium air S50, and a single compass reading can reflect a gust-driven wobble rather than the shift, so the tacking rule reads as a header beyond crew value degrees held for crew value seconds, not the first reading below the mean. The threshold is set by the cost of the two errors: tacking on a false header costs the manoeuvre and the distance to the wrong side, while holding through a real one costs 12 per cent of the lateral separation to the fleet for every 5 degrees of shift S52. Where the two costs are unequal, the threshold moves toward the cheaper error.

The rule set also carries a series-state input, because the value of risk changes with the standing. Dellenbaugh's principle is to "Go for many little gains rather than one big gain" and to treat a first-beat flyer as a lottery ticket S51; Fletcher states that "some starting approaches are riskier than others" and that the start plan weighs risk against reward for the result the series needs S64. A series score excludes the worst race unless the notice of race says otherwise S1, so where a discard exists it is a resource a rule may spend once and must not spend twice; Medal Race points at Paris 2024 were doubled and not discardable S6, and the two single-point races of the approved LA28 finals day S8 concentrate the series into a small number of decisions. Walker's sorting of legs into an oscillating wind, an obviously advantaged side and everything else, where the boat sails conservatively toward the middle and consolidates S62, is the default rule for the third category, and the crew's rule sheet states the exceptions.

Pressure narrows what the crew takes in, and the countermeasure is that the call is a recognition rather than a debate: the caller, the words and the response are all fixed by the protocol in section 15.11. After the race each decision is graded on process, whether the rule was applied with the evidence available, not on outcome, because outcome contains the wind's noise. A rule that fails repeatedly on process review is revised between events, never on the water. The testable standard is a written rule sheet with a threshold for every recurrent decision, and a post-race grading that marks each decision as rule followed, rule broken or no rule, the third category feeding the next revision.

15.11Crew communication and cohesion systems

The communication system is the mechanism by which two people with different views of the boat and the course act as one, and it is designed rather than left to develop. World Sailing describes the division of labour: "The helm generally makes the tactical decisions and steers the boat, while the crew undertakes more of the physical work and controls the sails" S9. The Sail Canada progression at CANSail 6 trains the crew to initiate tactical decisions upwind and the helm downwind S81, which follows from where each pair of eyes is: upwind the helm's attention is on the jib telltales and the waves and the crew, forward on the wire, has the wider view of pressure and the fleet; downwind the crew's attention is on the gennaker luff and the helm has the view. The class advises new teams to build "teamwork and communication" before racing S15, and the reason is that the system must exist before the load tests it.

The crew's own protocol, in the season plan and the Precision Communication document held in the campaign folder, is the reference for this crew: it defines the channels, the vocabulary, the priority order and the silence discipline, and nothing here replaces it. The theory beneath it is short. A call carries one piece of information and demands one response; a call is acknowledged, so that the caller knows it was received; a fixed vocabulary removes the interpretation step; a priority order decides which of two simultaneous needs is spoken first, with boat-handling timing above tactics and safety above both; and silence is defined, so that its meaning is agreed rather than guessed. The Paris 2024 Notice of Race prohibited voice, radio or data transmissions from the first warning signal to the end of the last race of the day S4, so the system is closed on the water and every input the crew acts on during a race originates aboard.

Cohesion is the state in which both would make the same call from the same evidencewhy, and it is built off the water: the shared video review in section 15.6, the shared data in section 15.7 and the rule sheet in section 15.10 are the instruments, because each produces an agreed reading of a situation that both crew carry into the next race. Disagreement is resolved after the race with the data, and a disagreement that recurs is treated as a missing rule rather than a difference of temperament.

The system is tested the way any other skill is. Audio from the coach boat is reviewed for calls outside the vocabulary, calls not acknowledged and calls made after the moment they were needed, and each is logged against the manoeuvre or decision it affected. The testable standard is that an observer listening to the boat in training hears the same vocabulary and the same acknowledgement pattern as in racing, hears the crew initiate upwind and the helm downwind as the protocol prescribes, and can match every call to a response within the interval the protocol sets crew value.