Two kinds of task, one clock.
A racing task is the pure form: fixed turnpoints, everyone flies the same track, fastest around wins. Your decisions are about the air — when to stop for a climb, which energy line to follow — and the computer's job is mostly navigation and final glide.
An assigned-area task replaces each turnpoint with a circle or sector kilometres across and adds a minimum time. You choose how deep to fly into each area, drawing your own task inside the envelope the setter drew. Now the decisions are also about geometry and the clock — and the scoring formula, not the finish line, decides who won.
The AAT arithmetic.
The scored speed is your distance divided by max(actual time, minimum time). Finish over the minimum and you are scored on the time you actually used — a slightly longer, well-flown task costs only whatever the late-day air charged. Finish under the minimum and your distance is divided by minutes you never flew: the denominator stays at the minimum while the numerator stops growing.
That asymmetry is the entire strategy of AAT flying. Ten minutes over time, in air worth flying, costs almost nothing. Ten minutes under time is pure subtraction, and no brilliance elsewhere on course buys it back.
Turn now, or two more kilometres?
The red dot slides along the blue curve as the geometry changes: turning for home now, 35 km out with 30 minutes still owed, lands you at the marked point. The "push deeper" figure is the out-and-back distance into the last area that brings you home exactly on time. Ten minutes early at 100 km/h on a 2:30 task gives away 6.7 km/h; ten minutes late costs only what the evening air was already charging.
Sequencing the areas.
Depth is not spent equally. Fly deep into the areas that sit in the best air — under the working cloud streets, on the sunny side of the convergence — and clip the areas that sit in the blue or downwind of terrain. Take the into-wind legs early, while the day is strong enough to pay for them, and drift home with the wind late.
Keep the most flexible geometry for last: a final area that is large and roughly on track is your time buffer, absorbing whatever forecast error the day accumulated. Arriving at the last area with the decision already made is the common error — the whole point of that circle is to be decided as late as possible, with the freshest information.
The final ring.
In the last area the strategy collapses to one line: turn where remaining distance ÷ expected speed = remaining time. Every kilometre deeper is two kilometres of scored distance, as long as the clock and the air both cover it. The tool's "push deeper" figure is that equation solved continuously — and it is exactly what the task computer's over/under-time indication is telling you in flight.
The honest inputs matter more than the formula. The expected speed for the run home is the late-day speed, not the noon speed; the remaining time must include the final glide. Feed the equation optimism and it will cheerfully instruct you to arrive after the day has died.
In your cockpit.
The task page carries this whole article as three live numbers: time over or under the minimum if you turned for home now, distance achieved, and the arrival altitude for the run home. Set the task before launch, including the start line rules, and let the instrument do the geometry — your capacity belongs to the air. The pre-start gaggle deserves the rest of your attention; it is the most crowded airspace of the day.
Drift decides the sequencing more often than courage does — how the computer measures the wind you are planning around is the wind article.