Left / Right: Choice Reaction Time, Hick's Law and Simon
· 7 min read
Most real reactions aren't just “go”. They're “go which way”. A goalkeeper diving, a fighter slipping left or right, a driver swerving. Left / Rightis a two-choice reaction test, and it measures how much a decision adds to your reaction time. Here's how it works, what the research says, and what a good score looks like.
How the game works
After a random wait an arrow appears pointing left or right. Tap the half of the screen it points to, or press the matching arrow key on a keyboard. Your time runs from the frame the arrow appears to your touch. Picking the wrong side is an error: on Easy you redo the rep, on Medium the set resets.
Hard adds a trap. Half the time the arrow appears on the oppositehalf of the screen from the way it points, for example in the right half pointing left. Only the direction counts. Your instinct says “tap where it is”, the rule says “tap where it points”, and any wrong pick resets the set.
What a choice costs: Hick's law
In 1952 the British psychologist W. E. Hick showed that reaction time grows steadily as the number of possible choices goes up. It doesn't grow in proportion to the number of choices, though. It grows with the number of times the choices double: 1 to 2, 2 to 4, 4 to 8. Ray Hyman extended the idea in 1953 to choices that aren't equally likely. Together this is known as the Hick–Hyman law, one of the most robust findings in psychology. A 2018 review by Proctor and Schneider describes both how well it holds and where it breaks down.
How much each doubling costs depends heavily on how natural the mapping between signal and response is. When the signal and the response line up well, each doubling adds only around 30 ms. Sleimen-Malkoun and colleagues (2013) measured about 31 ms, in line with an average of 27 ms that Jensen (1987) found across nine studies of young adults. With an awkward mapping the cost more than doubled, to 68 ms. An arrow pointing to the side you should tap is about as natural as it gets, so on Easy and Medium, Left / Right should be only a few tens of milliseconds slower than a simple reaction.
Less natural mappings cost far more. In a large study by Der and Deary (2017), using the same equipment as the UK Health and Lifestyle Survey, 30-year-olds averaged 290 ms on a simple reaction and 539 ms on a four-choice task. On that task each digit mapped to a finger, which takes much more translating than an arrow does. The gap also widens with age. By 69, simple reactions averaged 354 ms and four-choice reactions 729 ms.
Why the Hard mode is hard: the Simon effect
In 1969 J. R. Simon found that people respond faster when a signal appears on the same side as the response, even when its position is irrelevant. We are wired to respond toward where things are. This builds on earlier work by Fitts and Seeger (1953), which showed that responses are fastest and most accurate when the layout of the signals matches the layout of the responses.
Our Hard mode is a version of this called spatial Stroop: an arrow whose direction disagrees with its position. Studies typically find conflicting trials about 30 msslower. Ponce and colleagues (2025) measured 493 ms when an arrow's direction and position matched and 524 ms when they conflicted, and Narganes-Pineda and colleagues (2023) found almost the same: 504 against 535 ms. Errors rise too, especially when people rush. Among the fastest responses in another study by Ponce and colleagues, accuracy fell from about 99 % on matching trials to 88.5 % on conflicting ones.
The cost grows when conflict is rare. In a study by Paap and colleagues (2020), only a quarter of trials conflicted, and the cost jumped to around 90 ms, because people stopped expecting a trap. In our Hard mode half the trials conflict, so you can stay on guard, but you'll still feel the pull.
What score to expect
| Your average | Grade | How it compares |
|---|---|---|
| Under 300 ms | Elite | Barely slower than a simple reaction: the choice is costing you almost nothing |
| 300–379 ms | Fast | A sharp two-choice reaction on a phone |
| 380–479 ms | Average | Typical, especially early on or when you're double-checking |
| 480 ms and over | Keep practising | Common on Hard: lab arrow tasks with conflicting trials average roughly 490–535 ms |
Expect Hard to cost you about 30 ms or more on average. The bands are the same on every difficulty, so comparing your Easy and Hard averages shows exactly how much the conflict slows you down.
How to get better
- Let the arrow, not its position, drive the tap.On Hard, focus on the arrowhead's shape and ignore where it sits.
- Keep both hands or thumbs ready. One finger hovering in the middle has further to travel than two ready thumbs.
- Don't pre-pick a side. Left and right are random. Guessing a side gets you a fast time half the time and a reset the other half.
Compare it with Go / No-Go, which tests the decision to act at all, and with a pure simple reaction like Lights Out. The gap between those scores and this one is your personal cost of choosing.
Sources
- Hick, W. E. (1952). On the rate of gain of information. Quarterly Journal of Experimental Psychology, 4(1), 11–26.
- Hyman, R. (1953). Stimulus information as a determinant of reaction time. Journal of Experimental Psychology, 45(3), 188–196.
- Proctor, R. W. & Schneider, D. W. (2018). Hick's law for choice reaction time: A review. Quarterly Journal of Experimental Psychology, 71(6), 1281–1299.
- Sleimen-Malkoun, R. et al. (2013). Age-related dedifferentiation of cognitive and motor slowing: insight from the comparison of Hick–Hyman and Fitts' laws. Frontiers in Aging Neuroscience, 5, 62.
- Der, G. & Deary, I. J. (2017). The relationship between intelligence and reaction time varies with age: Results from three representative narrow-age age cohorts at 30, 50 and 69 years. Intelligence, 64, 89–97.
- Simon, J. R. (1969). Reactions toward the source of stimulation. Journal of Experimental Psychology, 81(1), 174–176.
- Fitts, P. M. & Seeger, C. M. (1953). S-R compatibility: spatial characteristics of stimulus and response codes. Journal of Experimental Psychology, 46(3), 199–210.
- Ponce, R., Lupiáñez, J. & González-García, C. (2025). Investigating the gaze-driven reversed congruency effect in the spatial Stroop task: A distributional approach. British Journal of Psychology, 116(4), 1107–1127.
- Narganes-Pineda, C. et al. (2023). Explicit vs. implicit spatial processing in arrow vs. eye-gaze spatial congruency effects. Psychological Research, 87, 242–259.
- Ponce, R., Lupiáñez, J. & González-García, C. (2025). Exploring the spatial interference effects elicited by social and non-social targets: A conditional accuracy function approach. British Journal of Psychology, 116(1), 69–88.
- Paap, K. R., Anders-Jefferson, R. & Zimiga, B. (2020). Interference scores have inadequate concurrent and convergent validity: Should we stop using the flanker, Simon, and spatial Stroop tasks? Cognitive Research: Principles and Implications, 5, 7.