Updated October 2026
Before this tool — or any timegrapher — can tell you anything about your watch's accuracy, it needs one piece of information from you: your watch's beat rate, in beats per hour (bph). Get this wrong and every other number the tool reports will be meaningless, not just slightly off.
A mechanical watch's balance wheel swings back and forth at a fixed frequency, and each half-swing (one direction) produces one audible "tick" or "tock." Beat rate is how many of those half-swings happen in an hour, conventionally stated in bph (sometimes written as vibrations per hour, vph — same thing). A 28,800 bph movement, the most common rate in modern mechanical watches, produces a tick every 3600 ÷ 28800 = 0.125 seconds — 125 milliseconds — which is the theoretical interval this tool compares your watch's actual tick timing against.
The most reliable source is your watch's movement caliber specification — check the manufacturer's website, the watch's manual, or a caliber-lookup reference using the movement number (often visible through a display caseback, or in the watch's original documentation). Many watch brands also simply state the beat rate in their marketing material or spec sheet for a given model. If none of that is available, an experienced eye can sometimes estimate it by ear — a 36,000 bph movement's ticks are noticeably more rapid-fire than an 18,000 bph one — but this is rough; a documented spec is always better than a guess.
Nothing in the detection pipeline itself depends on bph — the highpass filter, envelope follower, and onset detection run identically regardless. Bph only enters at the final analysis step, where the tool compares your actual average tick interval to the theoretical interval for whatever bph you selected. Pick the wrong one, and that theoretical interval is simply wrong, so the computed rate will be wildly, obviously off — often hundreds of seconds per day — rather than subtly off. A reading like "+640 s/day" on a watch you know keeps reasonable time is almost always a wrong-bph selection, not an actual problem with the watch; double-check the dropdown before troubleshooting anything else. It can also distort the outlier-rejection step in how this works, since the 0.5×-1.8× acceptance window is calculated around the (wrong) theoretical interval, potentially discarding otherwise-valid ticks or admitting noise that would normally be filtered out.
If you're not fully sure of your watch's bph, try the two most common values first — 28,800 for anything modern, 21,600 for many older or specific in-house movements — and see which one produces a plausible rate (generally within roughly ±60 seconds/day for almost any mechanical watch, even a poorly regulated one). The wrong bph selection will usually produce a reading far outside any plausible range, which is itself a useful diagnostic.
Does a higher bph mean a more accurate watch? Not directly — beat rate affects how finely time can theoretically be subdivided and how smooth the seconds hand looks, but actual accuracy depends much more on the quality of regulation, the escapement design, and manufacturing tolerances than on beat rate alone.
My watch isn't one of the four listed rates — what do I do? These four cover the large majority of mechanical watches in circulation. A small number of unusual or specialty movements use other rates; this tool doesn't currently support arbitrary custom bph values.
Can two watches with different bph be compared on the same "seconds/day" scale? Yes — rate in seconds/day is already normalized for beat rate, so a +5 s/day reading means the same thing whether the watch beats at 21,600 or 28,800 bph.