Part III - Racing Tactics and Strategy
Build 27 · 20 September 2026 · not yet reviewed
9Starting at 49er Speeds
Sections in this chapter
- 9.1Line bias, geometry, and time-on-distance modelling
- 9.2Final-approach planning and line access
- 9.3Acceleration lanes and disturbed-air management
- 9.4Committee-boat-end and pin-end frameworks
- 9.5Risk-adjusted line position
- 9.6Bail-out decisions and late-start recovery
- 9.7Final-minute communication and task sequencing
The start is the one moment in a race when the fleet's geometry is fixed by the race committee rather than by the wind, and the 49er's speed turns the last minute into a problem of metres and seconds. This chapter treats the start as geometry, timing and constraint. Section 9.1 gives the geometry of line bias, the transits that locate the line and the time-on-distance model that converts a position and a speed profile into an arrival time, with the rule 26 sequence and the rule 30 flags as fixed constraints. Section 9.2 turns the model into a final approach that puts the boat on the line at speed with a hole to leeward. Section 9.3 covers the acceleration lane and the disturbed air that shortens it. Section 9.4 ties the choice of end to the first-beat plan through ladder rungs. Section 9.5 adds the series position, so that the same line is started differently depending on what the regatta needs. Section 9.6 defines the bail-out and the recovery after a late or premature start. Section 9.7 sets out who calls what in the final minute, with the crew's own protocol as the reference.
9.1Line bias, geometry, and time-on-distance modelling
The favoured end is the end on the higher ladder rung. Ladder rungs lie perpendicular to the wind, and upwind the boat that has climbed higher is ahead S53, so a line that is not square to the wind puts one end a rung above the other before any boat has moved. The bias is measured by comparing the bearing of the line with the head-to-wind bearing S65, and the rule of thumb is that "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. Behind the rule, is the line length multiplied by the sine of the bias angle, and the separation measured along a close-hauled course is that figure divided by the cosine of the close-hauled angle. The rule assumes a conventional angle; the 49er's is its logged value crew value, and the crew rescales the rule with it. as well as with the angle, so the comparison is always length times bias and never the bias alone; a modest bias on a long line can outweigh a large one on a short line, and a large bias on a short line can outweigh a modest one on a long line.
The bias is worthless unless the crew knows where the line is. A transit through both ends to a fixed object ashore locates it S51; Nick Craig, quoted by The Final Beat, keeps two or three transits at either end, one on the line, one over and one behind S66. Paul Goodison, on the same page, adds a safe transit taken from two boat lengths behind the committee boat, on which a boat at mid-line is only one boat length behind the line W5: a transit taken behind one end converges on the true line towards the other. The behind transit governs the approach, the on transit the gun, the over transit the recall question. Since directional run from about 3 minutes in heavy air to 10 minutes or more in light-medium air, with amplitudes from 5 degrees in light-medium air to 40 degrees in heavier air S50, a bias measured at the warning signal can be gone by the start; Dellenbaugh takes many wind checks and postpones the decision until the last possible moment W4.
Time-on-distance converts a distance to the line and a speed profile into an arrival time. Fletcher names time-and-distance judgement as the core skill and works the approach back from the outcome wanted S64. The 49er adds a term a displacement class can ignore: speed depends on the boat's own apparent wind, and as boat speed rises "the apparent wind speed increases and the apparent wind angle decreases, moving further forward" S26. A boat leaving a luffing hold builds its wind as it builds speed, with the sheets following the apparent angle forward, so the distance to the line must contain the whole acceleration run, and is counted at two speeds, the creep of the hold and the full speed of the run, with the transition set from logged runs crew value. A Vakaros Atlas 2 reports distance-to-line, time-to-line and time-to-burn S79, but permits tracking equipment only "where permitted by the ICA, Notice of Race and/or Sailing Instructions" S3, so the judgement, not the display, is the skill.
The model runs against fixed constraints. Rule 26 flies the class flag at five minutes and the preparatory flag at four, gives one long sound at one minute and removes the class flag at zero, and "Times shall be taken from the visual signals; the absence of a sound signal shall be disregarded" S1; the watch starts on the flag. The preparatory flag prices the course side: under flag I a boat on the course side in the last minute must sail round an end, under Z she takes a 20 percent scoring penalty, under U she is disqualified unless the race is restarted, and under black the disqualification survives a restart S1, the last three applying to a boat identified in the triangle formed by the line's ends and the first mark during the last minute W1. Under those flags the run cannot touch the line early, so the plan holds a margin behind the line at one minute crew value that neither drift nor a leeward boat's luff can consume.
9.2Final-approach planning and line access
The final approach is planned backwards from the state wanted at the gun: hull on the line, at full speed, on starboard, with a gap to leeward large enough to bear away into. Fletcher works the approach back from the outcome wanted S64, and the evidence for what that state is worth is direct. Tracking of the 2015 49er Europeans found a of 10 m at the start S16, about two hull lengths on a 4.99 m boat S9, and attributed the larger gaps to port-and-starboard interactions W3. In the 49erFX of race time on 38 skills, distance behind the line at the start carried a coefficient of 0.060 and lengthened the race, while first-beat VMG carried -2.0 S37; the coefficients carry different units and are not comparable as bare numbers, but the start is the cheaper of the two to secure.
Line access is the problem of reaching a point from which that state is attainable, against boats that want the same point. Starboard is the default: "Unless you have a very good reason, always cross the line on starboard tack" W4, and in the 2015 data starboard boats held under one second of advantage in time to the line after the gun S16 and more control over start quality when they found clear lanes W3. The access is governed by the same-tack rules. A windward boat keeps clear of a leeward boat W1; a boat that becomes overlapped from clear astern within two hull lengths to leeward may not sail above her proper course, but "A boat has no proper course before her starting signal" W1. Before the gun the boat arriving from astern and to leeward may therefore luff the boat above her, giving room to keep clear as she changes course W1; after the gun she is bound to her proper course. The hole is taken from leeward, defended by luffing before the gun and cashed at the gun by bearing away into it.
The 49er cannot hold a point. A luffing skiff makes leeway quickly because "When the centerboard and rudder are moving through the water slightly, they are much more effective at generating lift" S35, so the hold is a slow creep, and Dellenbaugh's warning that head to wind is "the quickest way to lose steerage" W4 applies with force. The approach stays within half the line's length of the line, especially in light air W4. Fletcher's light-air example, holding the starboard end from potentially nine minutes before the gun in six knots W2, shows that in light air access is bought with time, because a boat arriving late has no speed to buy it with. The final run is a close reach for a few seconds, then a luff to close-hauled as the gun goes W4, with the sheets coming in as the apparent wind moves forward S26; the Atlas 2 S79 is the machine form of the same count.
On the water the plan is a sequence of positions at signals: a spot at four minutes, a transit and a gap at one minute crew value, a creep speed and a run-up point. The log records the gap and speed at the gun, the tack and the state of the lane one minute later. The 2015 analysis found between start quality and position at the first windward mark because of the tack choices made afterwards W3, which is why the log pairs each start with the first-beat plan of 9.4 rather than scoring it alone.
9.3Acceleration lanes and disturbed-air management
The resource contested in the seconds after the gun is a lane of undisturbed air long enough for the boat to reach full speed. Dirty air extends to leeward and aft in the direction opposite the leading boat's apparent wind S54; upwind a boat tacks away from one on or to weather of her bow "within 6 to 10 boatlengths (the greater distance is in lighter conditions)" S56, and wind shadows extend farther and slow more in light air S55. Because a 49er's apparent wind is well forward at speed, the shadow of a boat ahead and to windward lies further aft than a displacement sailor expects, and a bow-out boat casts hers onto the boat behind and to leeward. The lane is defined by the boat to windward and ahead, not by the one abeam.
Acceleration is the mechanism that fills the lane. From a creeping hold the apparent wind is close to the true wind, wide and weak; depth is one source of acceleration, "More depth creates more power and better acceleration" S20, and angle of attack the other. Dellenbaugh trims the jib first to pull the bow off towards close-hauled, then the main W4. As speed builds the apparent wind moves forward S26 and the trim follows it in; sheeting to the final angle at once stalls the flow. Heel is held to the skiff standard, "Flat is FLAT, +/- 5 degrees is acceptable" S36, and the foils recover their lift with the first flow S35, which cuts the leeway of the hold. The hole to leeward is the bear-away room, "big enough to give you room to accelerate, but small enough so you don't tempt another boat" W4; its size in the 49er is set by the acceleration run logged in 9.1 crew value.
Disturbed air is managed by position relative to the nearest boats at the gun. A safe-leeward position gives "a lift and more pressure as the wind bends around the front of their sailplan" S54; the boat that is bow-out and to leeward of its nearest windward neighbour gains this and the neighbour loses it. Once a boat to leeward and ahead is lifting away, the boat behind is in the air deflected around the front of that boat's sailplan rather than inside the 6 to 10 length shadow zone S56, and the only remedy is to tack. The 49erFX data associated extra manoeuvres on the first leg with shorter race times S37, which is consistent with a tack out of a dying lane costing less than the leg sailed in it. The crew tacks when the lane is lost by a criterion set in advance, such as the leeward boat's bow passing a reference on the deck crew value, not when the lane feels lost.
Holding the lane before the gun is bounded by rule 42. Only wind and water may propel the boat S1; the class rules allow sculling only "if the sculling does not propel the boat in any direction" S3, and class breaches "do occur in the 4-8 knot wind range" S18, where the hold is longest. Under Appendix P an observer penalises a boat "by, as soon as reasonably possible, making a sound signal, pointing a yellow flag at her and hailing her sail number" S2, and at Paris 2024 and the 2025 World Championship the penalty was one turn S4S5. Backing a sail is legal but costly in rights: a boat moving astern, or sideways to windward, by backing a sail keeps clear of one that is not W1. The hold is therefore a creep with the foils working, and the double tack is the legal way up into the hole of the boat to windward W6.
9.4Committee-boat-end and pin-end frameworks
Position on the line follows the first-beat plan, pin end for the left side and committee-boat end for the right, and line bias is overridden when one side of the course is strongly favoured S51. The reason is the relative scale of the two effects. Line bias is a fixed fraction of the line, about an eighth for 5 degrees S65; a 5-degree shift on the beat changes the relative position of two boats by 12 percent of their lateral separation S52, and a fleet spread across the whole line is at its maximum lateral separation the moment the gun goes. The first shift therefore repays or refunds the choice of end. Walker's sorting applies: an oscillating wind, one side obviously advantaged, and everything else, where the boat starts for the middle and consolidates S62.
The pin-end framework is a bet on the left. The pin boat sits on the highest rung of a left-biased line and, once clear, holds starboard tack towards the left with right of way over everything crossing on port S1. Dellenbaugh lists its advantages, the line is easy to judge and "you won't have any boats to worry about on your lee bow", and its costs, "This end is usually crowded, especially if it's favored" and "If you get a bad start, it will be very difficult to find clear air" W4. The costs follow from the geometry: the pin boat cannot tack without ducking the whole fleet, so her plan is a long starboard tack and her only exit is speed. The mark gives no help: Section C rules do not apply at a starting mark surrounded by navigable water W1, so there is no mark-room at the pin, and the leeward-most boat holds rule 11 rights against everything to windward W1.
The committee-boat framework is a bet on the right or on freedom. Dellenbaugh's advantages are the easy line judgement and "If you have a bad start, it's easy to bail out and get clear air on the other tack" W4. The mechanism is the shadow geometry of 9.3: a boat that tacks onto port at the boat end crosses behind the starboard fleet, whose dirty air extends to leeward and aft S54, away from her, and she buys clean air with a duck or two under rule 10 S1. The same absence of mark-room applies at the committee boat W1: a boat reaching in to windward of a leeward boat has no room to the mark, and rule 11 W1 puts her out. Fletcher's light-air ownership of the starboard end from nine minutes out W2 is the price of this framework when the fleet wants the same corner.
Between the ends the middle offers freedom to tack either way and the worst line judgement; Hunt notes that the longer the line, the greater the illusion of being over when the boat is in fact behind, and that the effect is strongest in the middle W5, and Dellenbaugh's mid-line rule is to hold with the others in a sag and go only when there is just enough time to reach the line at the gun W4. The port-tack start is a framework of its own: in the 2015 data starboard held under one second of advantage S16, and a port-tack start in an extremely clear lane won a race at that regatta W3. The end is chosen by the plan, the spot within the end by the lane, and the tack by the rights.
9.5Risk-adjusted line position
Fletcher's statement that "some starting approaches are riskier than others" and that the plan weighs for the result the series needs S64 turns the frameworks of 9.4 into distributions rather than choices. Each spot on the line has a . The favoured pin has the highest mean and the heaviest bad tail, crowded and without clear air after a bad start W4; the boat end has a lower mean and a thin tail, because the bail-out onto port is cheap W4; the middle has a moderate mean and a sag risk W4. Risk-adjusted position means choosing the spot whose distribution, not whose mean, suits the series.
The scoring system sets the . The Low Point System scores each race by finishing place, and the series score excludes the worst score unless the notice of race says otherwise S1; at Paris 2024 the Men's Skiff sailed 12 races plus a Medal Race with the worst score discarded, the top ten qualifying and Medal Race points doubled and not discardable S4S6, and the LA28 format approved on 10 December 2025 provides a three-day qualification series and a finals day of two single-point races for the top ten S8. A discard in hand gives a start an : the bad tail is thrown away and the good tail kept. A discard already used, a medal race or a finals-day race removes the option, and variance becomes a pure cost. Fletcher's SailGP arithmetic, accepting fifth at the first mark when that average reaches the season goal so that "we don't need to take massive risks on the start line" W2, is the same calculation.
The preparatory flag reshapes the tail. Under P the cost of being over is a return; under I a return round an end; under Z a 20 percent scoring penalty, under U disqualification unless the race is restarted, and under black a disqualification that survives a restart S1. Each step moves probability from the bad tail towards the catastrophic end, so the margin behind the line at one minute crew value and the willingness to push the on transit both change with the flag. Dellenbaugh's percentage principle, "Go for many little gains rather than one big gain", with the first-beat flyer treated as a lottery ticket S51, is the general form: the start is one of the little gains and is not bought with a ticket.
The distribution is measured, not felt. The crew logs for each start the spot chosen, the flag, the gap and speed at the gun, the lane state at one minute and the position at the first mark; the 2015 finding that start quality did not with weather-mark position because of the tack choices that followed W3 is why the last field is included. Over a season the log gives each framework its own mean and spread at each venue, and the risk-adjusted decision becomes a lookup in the crew's own table rather than a judgement made in the last minute.
9.6Bail-out decisions and late-start recovery
A bail-out is a planned reset with a deadline, not a reaction. Fletcher's rule is to "Bail out with at least a minute to go because that'll give you more time to reset" W2, and the deadline coincides with the moment the rule 30 flags take effect S1W1. The triggers are conditions the model of 9.1 can test: the boat cannot reach the target hole at speed by the gun; the hole has closed from leeward and there is no bear-away room; a shift has moved the bias so that the chosen end no longer serves the plan; or, under flag U or black, the margin behind the line at one minute is threatened. Any one is sufficient. After one minute the reset costs more than the bad start it avoids.
The reset mechanics differ by end. At the committee boat the bail-out is a tack onto port and a duck W4, which works because the fleet's shadow lies to leeward and aft of it S54. At the pin it is a bear-away and gybe down the line to the next usable hole, because a tack at the pin means ducking the fleet. Both must end with a run-up: the 49er needs distance to build its apparent wind S26, so a reset into a hole with no run-up is worse than a reset to a poorer spot with one. Moving within a hole has its own rules: a boat moving astern or sideways to windward by backing a sail keeps clear of one that is not W1, so the slide is used only with space, and the double tack moves a boat up into the hole of the boat to windward or out from a boat tight to leeward W6.
Late-start recovery is a different problem: the boat is behind at the gun and in someone's shadow. Fletcher's rule against the panic tack is that "it's nearly always better to wait 10 seconds, take stock of your situation, and choose your moment" W2; a tack made in the first seconds lands the boat in the next shadow, which extends 6 to 10 lengths to leeward and aft S56. The recovery target is the beat, not the start. In the 49erFX first-beat VMG carried a coefficient of -2.0, "almost ten times that of any other skill", while distance behind the line carried 0.060 S37; the coefficients carry different units, and a two-length deficit sailed in clear air at the boat's VMG costs less than a clean start followed by a beat in bad air. Extra manoeuvres on the first leg were S37, which is consistent with the tack that buys the lane.
An over-early boat recovers under a fixed procedure. The race committee signals an individual recall with flag X S1; a boat starts only when her hull has been entirely on the pre-start side at or after her starting signal W1; a boat sailing towards the pre-start side to start, or to comply with rule 30.1, keeps clear of one that is not W1, and under flag I she must cross an extension of the line S1W1. Craig's over transit S66 answers the question before the flag does. A boat that sees the transit over returns at once, because at 49er speeds every second on the course side is distance that must be sailed twice more, and the return is made towards the nearer end so that the re-cross has a run-up.
9.7Final-minute communication and task sequencing
The final minute is a scheduling problem with a deadline that cannot slip: a fixed set of observations, time, distance to line, transit, wind, lane and neighbouring boats, must be turned into three decisions, hold, bail or go, by two people whose hands are also on the sheets and the tiller. The published baseline for the division is World Sailing's: "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; Sail Canada's progression trains the crew to initiate tactical decisions upwind and the helm downwind S81. The crew's own protocol, Temporada 26-27 and the Precision Communication document, defines the channels, vocabulary, and silence discipline and is the reference; this section gives the theory it rests on.
Ownership follows the sensor. Each datum is owned by the person positioned to read it and is spoken by that person only; the other does not duplicate it, because a duplicated call and creates a second version of the truth. The person facing the shore object owns the transit; the person with the watch owns time and starts it on the visual signal, since "Times shall be taken from the visual signals" S1; the person with the tiller owns the boats to leeward and the hole; the person forward owns the boats to windward and astern. Data are given as numbers and states rather than opinions, and the closed loop is a short acknowledgement, so that the decision-maker knows the datum was received without a conversation. Devices may supply distance-to-line, and time-to-burn S79 where and the event documents permit S3; the Paris 2024 Notice of Race prohibited voice, radio or data transmissions from the first warning signal S4, so nothing arrives from outside the boat.
Sequencing follows deadlines and reversibility. The bias decision is re-taken as late as possible W4 but before one minute, because after that the reset is too expensive; the bail-out decision is taken at one minute W2; hole defence, the creep and the luff of a windward boat fill the interval to the run-up point; the go call is made by the owner of time-to-burn at the point set from logged runs crew value, and from that call the only remaining communication is trim and speed. Each decision has a single decision-maker and a single call, and each call has a default: silence at the deadline means the plan continues. The tasks, the bail and the go, come last, and the reversible ones, wind and lane, are updated continuously until then.
The standard is observable from a coach boat. In the final minute every datum is heard once, from one voice, in the order the plan fixes, and no call is repeated or questioned; the go call is followed by trim and by nothing else until the boat is at speed. A crew that meets the standard can be interrupted by a shift, a boat closing the hole or a missing sound signal, and the observer sees the sequence resume from the same point rather than restart. Whether the calls themselves match the crew's protocol is checked against that document, not against this chapter.