Lights Out at Sepang: The Random Wait That Decides an F1 Start

For the first time since 2017, Formula 1 has raced at Sepang. The weekend of 2 to 4 October 2026 is officially the Bahrain Grand Prix, moved from its April date because of the conflict in the Middle East and staged in Malaysia with Bahrain funding it. The name on the trophy is unusual. The 5.543-kilometre circuit, the 56 laps and the afternoon rain clouds over Selangor are the Sepang that hosted the Malaysian Grand Prix every year from 1999 to 2017.

Every one of those races, and every race this weekend, begins with the same short test of the human nervous system. Five red lights come on above the grid, one at a time. Then, after a pause, they all go out together, and twenty-two drivers try to be the first to respond.

The pause is the clever part. According to Formula 1’s own explanation of the procedure, the lights go out after “a randomly-chosen and computer-controlled wait of between 0.2 and 3 seconds.” The wait is picked by a computer, not a person, so nobody on the grid can know how long it will be — and that is exactly the point.

We wrote about why randomness matters in any reaction test at cerdikcerdik.com/blog/average-human-reaction-time. If the wait were always the same, drivers would stop reacting to the lights and start counting, and the start would become a contest of rhythm instead of reflexes. A random delay forces a genuine response. It also makes guessing expensive: a car that moves before the lights go out has jumped the start, and that means a penalty.

Formula 1 says drivers launch their cars about 0.2 seconds after the lights go out, on average. That is fast, but it is not superhuman. It sits at the quick end of the 200 to 250 milliseconds that a rested, attentive adult manages on a simple visual test. What separates a racing driver is not a faster nervous system so much as everything around it — total focus at exactly the right moment, and a launch they have rehearsed thousands of times, feeding the clutch out with a paddle while bringing in the throttle without bogging down or spinning the wheels too much.

But there is something subtler happening during that random wait, and it is one of the more elegant findings in the psychology of reaction time. The wait is unpredictable, but it is not uninformative. It has a maximum. Once the fifth light is on and two and a half seconds have passed, the lights have to go out very soon, and every tenth of a second of waiting makes the next tenth more likely to be the one.

Put numbers on it. Suppose the delay is spread evenly across its range. Just after the fifth light comes on, the chance that the lights go out in the next half second is about 18 percent. One second in, it is 25 percent. Two seconds in, 50 percent. At two and a half seconds, it is certain. The longer nothing has happened, the more certain it becomes that something is about to.

Brains pick up on this without being told. In laboratory reaction-time studies with a variable waiting period, people reliably respond faster when the signal arrives late in the window than when it arrives early — not because they guessed, but because their readiness quietly climbs as the probability does. Psychologists call it the foreperiod effect. It is why the most testing start is not the long wait but the short one: lights that go out a fifth of a second after the fifth light catch drivers at their least prepared moment.

The 2026 regulations have added another step to all this. The new power units no longer have the MGU-H, the part that used to keep the turbocharger spinning, so the turbo now needs time to spool up before the car can deliver full power off the line. To give drivers that time, the grid panels now flash blue for five seconds before the red light sequence begins. It makes the procedure longer and the pre-start more predictable, but the part that matters for reactions has not changed. The final wait is still random, and the race is still decided by who responds best to lights they could not time.

Milliseconds sound abstract until you turn them into distance. At 300 kilometres per hour, a car covers about 83 metres every second. That means 0.2 seconds of reaction is almost 17 metres of track, a little more than two of the 8-metre gaps between grid slots, and a 50-millisecond difference between two drivers is more than 4 metres. At the start itself the cars are only just moving, so the gap in metres is far smaller, but a car that gets away a few hundredths of a second sooner is the one that reaches the first corner on the inside.

Reaction time matters far less once the race is under way, and that is worth knowing too. Drivers do not react to most corners. They know exactly where the braking point is from practice and qualifying, and they anticipate it. Pure reaction comes back when something unexpected happens — a spinning car ahead, debris on the track, a sudden downpour of the kind Sepang is famous for. Those are the moments when the 200 milliseconds matter again, because there is nothing to anticipate.

You can test the same reflex at home with our reaction time test. The idea is the same: a red screen, a random wait, then a signal to respond to as fast as you can. Our wait is longer than F1’s, randomised between two and five seconds, and the same foreperiod maths applies. Four seconds in, the chance of green in the next half second is already 50 percent. Watch whether your fastest rounds are the ones where the green came late. If they are, you are not cheating. Your brain is doing exactly what a driver’s does on the grid.

The score is the average of five rounds, and that average will almost certainly be slower than 0.2 seconds. Some of the gap is your phone or mouse, which adds delay that F1’s timing system does not have, and some of it is that you are not a professional driver who has done nothing but prepare for this moment. Try it at cerdikcerdik.com/games/reaction-time, and see how close you can get to the grid at Sepang.