The faster you press, the more you miss — the speed-accuracy tradeoff

Published 2026-09-08

Play Hard Brake long enough and you hit this wall. Concentrate on holding back for red and your green responses slow down. Push to hit green fast and your hand goes out on red. Every attempt to do both keeps failing.

That is not a skill gap. Speed and accuracy sit in a tradeoff, and once you see the shape of it, the way to raise your score changes.

One point on a curve

Fast and accurate are not two separate abilities. They are one curve where raising either lowers the other — a relationship studied for a long time in reaction-task research.

Where you stand on it is set by two different things:

The height of the curve itself — this is what people mean by ability: faster at the same accuracy, or more accurate at the same speed.
Your position along the curve — this is not ability. It is a choice you made: cautious, or bold.

When a score stalls, most people try to raise the curve. Far more often, the actual problem is that they picked the wrong position.

The scoring rule decides the best position

This is the core of it. Where it pays to stand is decided by how that game scores you.

Hard Brake scores like this: your average time on green, plus penalties in proportion to the reds you pressed and the greens you missed. But those two penalties do not weigh the same.

Pressing every red: 900ms of penalty
Missing every green: 7,200ms of penalty

Both are coefficients applied to a rate — press half the reds and you add 450ms. The miss coefficient is 8 times heavier. In other words, this game punishes excessive caution harder. If chasing a perfect hold on red makes you drop a few greens, you lose more than you saved.

Do not read those coefficients as the final score, though. There are two exceptions:

If you hit no greens at all, there is no average reaction time to add to, so a failure value of 3,000ms is used instead of the sum.
The final value is clamped to 150-3,000ms. So “missing every green adds 7,200” does not mean 7,200 is what you see.

Working memory runs on different rules. There a hit is +10 and a false tap is −5. So once it feels more than about 1/3 likely that this is a target, tapping is the better bet — at exactly 1/3 the +10 and the −5 cancel out.

Penalties accumulate for the whole run. A score that looks like zero mid-run does not mean your false taps have been forgiven — hit targets later and the earlier false taps show up then. (Only the final score is clamped to 0-100.)

That matters because “don’t press unless you’re sure” turns out to be much narrower advice than it sounds. At a coin-flip hunch, pressing wins. Only hold back when you have no read at all. The same person should be standing in different places in these two games.

So “good focus” is not a single value. Each game demands a different position, and your score comes from whether you can stand there.

Reading your own position

Break your result into its parts.

Many missed greens → you may be leaning cautious. The miss coefficient is 8 times heavier, so the bold end is worth trying.
Many pressed reds → you are too bold. Though in Hard Brake that penalty is light, so if your average is fast enough it may not be costing you.
Few of both, but a slow average → there is little left to win by shifting your lean. What remains is raising the curve itself.

In the first two cases, changing which way you lean is the cheap move to try before trying harder. You can shift your lean on the spot; raising the curve is not like that.

The same problem off-screen

This tradeoff works the same away from a screen. What changes from situation to situation is which kind of mistake is more expensive.

A message worth sending fast with a typo, versus a contract that must be right even if it is late
A lane change you need to make now, versus an intersection where stopping when unsure is better

More often than building the ability, the first step is knowing which mistake the current situation punishes harder.

One caution

These tests are for fun. Nothing here indicates impulsivity as a personality trait, driving ability, or any kind of assessment. Academic tasks (such as SART) explore similar structures, but what those studies validated were their own stimuli, trial counts and conditions — not FlowScore’s 30 color trials and additive penalties.

In short

Speed and accuracy trade off. Push both at once within a single run and one of them usually gives (that is a different question from both improving over time).
A score splits into the curve’s height (ability) and your position on it (choice).
Where it pays to stand is set by that game’s scoring rule — and it can be the opposite between games.
Reading misses, false taps and average separately shows your lean immediately.