This is a short game-like check of your sense of rhythm. You will tap along with clicks and a bit of music, keep tapping on your own, and listen for beeps that are a little off. At the end you get a score and some simple tips.
Timing has two sides. Doing it: can you tap in time with clicks, with music, and on your own? Hearing it: can you notice when a beep comes a little early, or when beeps do not fit the music? Both are checked, because people can be good at one and not the other.
For tapping, what matters most is how spread out your taps are around the beat, not whether you are a little early. Almost everyone taps slightly before a click without noticing (researchers call this the negative mean asynchrony, usually 20 to 80 ms). Your device also adds a constant delay. That is why the check starts with a short "tap with the clicks" step: it measures your personal offset on this device and subtracts it.
Each part of the check gives a number. We compare each number with published values from 108 adults, mostly non-musicians, who took the research version of these tasks (Kotz et al. 2024). A score of about 50 means "like a typical adult"; 70 and up means steadier or sharper than most; below 30 means there is clear room to grow. The overall score is a weighted average of the parts you completed: tapping with a beat (30%), keeping time on your own (30%), hearing timing (25%) and moving to music (15%). The quick check only has some parts, so it gives a "quick score". A constant device delay never lowers your score; only spread and drift count.
The verdict names the pattern in your results, using criteria from rhythm research. "Steady with help" (fine with clicks, drifting alone) is the most common pattern in people who feel they have no rhythm. The pattern researchers call beat deafness (not being able to find the beat in music, even though tapping with a plain click works) is rare, about 1 to 3 % of people by some estimates. If the check shows it, the most likely explanations are still sound delay, a misunderstanding of the task, or a bad day, so retake it on another day or device before you believe it. Even then, it is a screening result, not a diagnosis, and rhythm can be trained.
Studies find that people with ADHD are, on average, a little more variable at tapping tasks and less sharp at hearing timing changes (Noreika, Falter & Rubia 2013; Puyjarinet et al. 2017). Many adults with ADHD still score in the typical range. A rhythm score cannot tell anyone whether they have ADHD or anything else, and poor rhythm is common in everyone. What research does suggest: rhythm improves with practice, and an outside beat (a click, a song, counting out loud) is a good support while your inner beat is still forming.
| Part | Measures | Based on |
|---|---|---|
| Calibration | Median tap-to-click offset (ms) and its SD, 20 clicks at 600 ms IOI | Latency handling as in REPP (Anglada-Tort et al. 2022) |
| Tapping at your own pace (30 s) | Spontaneous motor tempo (mean ITI) and CV of ITI | BAASTA unpaced tapping; McAuley et al. 2006 |
| Tapping with clicks, 600 and 450 ms IOI, 40 clicks | Vector length R (logit), vector direction (% of IOI, calibrated), SD of asynchrony, CV of ITI, Rayleigh test | BAASTA paced tapping; Repp 2005; Repp & Su 2013 |
| Clicks stop, keep going (20 + 30) | CV of ITI in sync and continuation, mean continuation ITI, drift (regression slope of ITI, % of IOI over the phase) | BAASTA synchronization-continuation |
| Tapping to music (48 beats, 100 BPM synthesized drum and bass) | As paced tapping, plus Rayleigh p | BAASTA paced tapping to music; H-BAT Music Tapping Test |
| Early beep (5 tones, 600 ms, 4th tone early) | Anisochrony detection threshold, % of IOI, 2-down/1-up staircase (halve / x1.5, 8 reversals, mean of last 4), 3 catch trials | BAASTA (Dalla Bella et al. 2017; Kotz et al. 2024) |
| Beeps on the music (12 to 16 clips, 600 and 450 ms) | d' for aligned vs phase-shifted (33%) or period-shifted (10%) metronome, log-linear correction | Beat Alignment Test (Iversen & Patel 2008); BAASTA BAT; Gold-MSI beat perception (Müllensiefen et al. 2014) |
| Longer beep (full check only) | Duration discrimination threshold, % of 600 ms, same staircase, 2AFC | BAASTA duration discrimination |
From Table 3 (session 1 mean and SD) of Kotz, Sowiński, Dalla Bella et al. 2024, mobile BAASTA adult norms, n = 108, ages 18 to 87: spontaneous ITI 693.9 (251.8) ms, CV 0.056 (0.034); paced tapping 600 ms: CV 0.046 (0.012), logit R 2.95 (0.81), direction −6.4 (5.3) %; 450 ms: CV 0.052 (0.015), logit R 2.77 (0.85), direction −7.1 (6.0) %; music 1: CV 0.054 (0.026), logit R 2.86 (1.48), direction 0.6 (4.9) %; sync-continuation 600 ms: CV 0.044 (0.013), ITI 597.9 (26.1) ms; anisochrony with tones 11.7 (4.9) %; duration discrimination 28.2 (11.9) %; BAT d' 2.81 (1.01). H-BAT (Fujii & Schlaug 2013, 30 adults with a mean of 11.8 years of musical training) reported a music-tapping vector length of 0.93 (SD 0.03); shown for context only.
Not confirmed from the sources: a typical SD of asynchrony in ms and a typical ITI CV for non-musicians from Repp (2005) or Repp & Su (2013); the page uses no reference band for SD of asynchrony. The 20 to 80 ms negative mean asynchrony comes from secondary summaries of Repp 2005. The exact item count and tempi of the Gold-MSI beat perception subtest. Whether BAASTA's synchronization-continuation CV covers the continuation phase alone (assumed here). The exact vector-length cut-off in Sowiński & Dalla Bella 2013 (the secondary sources say "more than 2 SD from the control mean"). The drift reference (0 ± 8 % of the IOI over 30 taps) is an assumption, not a published norm.
Each measure is converted to a z-score against the reference mean and SD, signed so that positive is better (for two-sided measures such as held tempo, the absolute deviation is used). A sub-score is the mean z of its measures mapped through a logistic: score = 100 / (1 + e−z), so z = 0 gives 50, z = +1 gives 73, z = −2 gives 12. Sub-scores: keeping time with a beat (logit R and CV at 600 and 450 ms), keeping time on your own (continuation CV, held ITI vs 600, drift), hearing timing (anisochrony threshold, BAT d', duration threshold), moving to music (logit R and CV with music). Overall = weighted mean of completed sub-scores, weights 0.30 / 0.30 / 0.25 / 0.15 renormalised. Mean offset and direction are never scored.
Results (numbers only, never raw taps) are saved to this site's database under a random id, with the nickname you chose, if any. Other people who take the check can see your score and nickname in the "Everyone" view. The device type (phone or computer) is saved so comparisons are fair.