What a Punch Tracker Can and Can't Tell You
There is an argument going round against punch trackers, and it goes like this. A sensor on your wrist counts punches. It cannot see your guard, your feet or your shape. A sloppy arm punch thrown with your chin in the air registers exactly the same as a clean cross from a proper stance. So the number it gives you measures how much you did, not how well you did it, and chasing that number teaches you to throw more bad punches, faster.
We make a punch tracker, so you would expect us to argue with that. We are not going to, because it is correct. What follows is a straight account of what a punch tracker genuinely cannot see, what the independent research says it can, and where that leaves you if you are deciding whether to buy one.
What a wrist sensor cannot see
Start with the full list, because the guides that sell these devices tend to skip it.
Your guard. A tracker reads the motion of the hand it is attached to. It has no idea where that hand is between punches. If your right drops to your hip after every jab, the tracker records a jab.
Your feet and your weight. A punch thrown off the back foot with your hips behind it and the same punch thrown flat-footed from the arm can produce similar wrist motion. The sensor measures the hand. Everything below the shoulder is invisible to it.
Defence. Slips, blocks, parries, head movement, whether you are getting hit. None of it. A sensor cannot tell a round where you took twenty clean shots from a round where you took none.
Whether the punch was any good. This is the real objection and it stands. A tracker can tell you a punch happened, roughly what type it was, how fast the hand was moving and how hard it decelerated. It cannot tell you whether it was technically sound, whether it landed where you meant it to, or whether you should have thrown it at all.
So if someone tells you a punch counter will fix your technique, they are wrong. The count is a measure of output. Output is not quality. On this, the sceptics and the people who build these things should be in complete agreement.
What the research says it can do
The next question is whether a tracker even measures the output reliably, and here you do not have to take a manufacturer's word for it. Two peer-reviewed studies from Charles University in Prague have tested commercial punch trackers against video and against laboratory equipment. Corner was in both. So were Hykso, Everlast and StrikeTec. Here is what they found, including the parts that are not flattering to us.
Counting punches. In the first study, published in the journal Sensors in 2021, twenty-one men, ten of them trained fighters and eleven with no experience, shadow-boxed set rounds of 54 punches while a camera recorded every one. Corner's count across all participants was within about three percent of the true number on average, and it was the only tracker in the study whose count was statistically equivalent to the video. The others undercounted the beginners badly. Hykso and Everlast each missed roughly one punch in seven thrown by an untrained puncher, and StrikeTec registered fewer than a quarter of the punches thrown. If you want to know how many punches you threw, a good tracker will tell you, and it will tell an experienced boxer more accurately than a beginner, because a beginner's punches are less consistent.
Telling punch types apart. This is harder, and the numbers say so. Corner correctly identified 83 percent of lead straights and 91 percent of rear straights. For hooks it fell to 65 percent on the lead side and 54 percent on the rear. Uppercuts sat between the two, at 74 and 78 percent. Every tracker in the study was better at straights than at hooks and uppercuts, because a straight punch is a straight line and a hook is a curve that different people throw differently. Corner was the strongest of the group on hooks and uppercuts, and Hykso was the strongest on straights. Both got worse the deeper a punch sat in a combination. The third punch of a three-punch combination was recognised less reliably than the first, and three was the longest combination the study tested.
Speed and power. The second study, in the Journal of Strength and Conditioning Research in 2023, had twenty men punch a wall-mounted force plate while 3D motion capture measured the true speed of the fist. No tracker's numbers correlated strongly with either force or velocity. Corner's power figure had a moderate relationship with peak hand speed. Corner's speed figure, in the version tested, did not track laboratory velocity at all, and the authors suggested it was probably closer to a measure of impact. The errors grew as punches got faster. Their conclusion was blunt: if you are not willing to accept any room for error, none of these devices should be used.
Two things to say about that. First, the app version the researchers tested was 1.3.1, collected in 2019. The hardware, firmware and algorithms have been revised since, but no independent lab has re-tested them, so we are not going to claim a newer number. Treat the published figures as what an outside party has verified, not as the ceiling. Second, the practical reading of both studies is the one the authors give themselves: a tracker's count is trustworthy, its punch types are trustworthy for straights and indicative for hooks, and its speed and power figures are best used to compare you against yourself over time, not against a laboratory or against a friend.
The half a coach cannot watch
Now the part the objection leaves out.
Think about what a coach can actually do with their eyes. A good one standing over you on the pads will see your guard drop, your weight sit back and your feet go flat. They will fix all three in a session, and no sensor will ever do that job. What they cannot do, and nobody can, is count. A coach watching twelve people on twelve bags cannot tell you that your output fell from around 80 punches in round one to around 40 in round four. Those are example numbers, not a real session, but the shape is one every coach knows and none can measure by eye. They cannot see that your right hand is throwing 60 percent of your punches when it should be closer to half, or that your speed in the last thirty seconds of every round is a step below the first thirty.
Those things are where boxing breaks down first when you are tired. Work rate that decays across rounds. Speed that fades late. A hand you stop using. A combination you cut short. They are also exactly what an accelerometer strapped to each wrist is good at, because they are questions about how much and how fast, not about how well.
A tracker worn on both hands, through a full session, gives you a round-by-round record of punches thrown, punches per minute, average and top speed, and the split between left and right. Session after session, that record is something no coach could have written down, and it shows you the two things you most want to know as an amateur: whether you are doing more work than you were three months ago, and whether you are still doing it in round six. The guard and the footwork are the coach's department. Output across time is the tracker's.
Put it another way. A running watch cannot see your form either. It measures pace, distance and heart rate, and runners use it to see whether they are getting faster, not to fix their stride. Nobody thinks that makes it useless. A punch tracker is the same kind of instrument, and the same rule applies: it measures your work, and someone else has to judge your technique.
What about MMA?
A quick, honest note, because Corner describes itself as boxing and MMA technology and you deserve to know what that means.
The trackers read the hands. In an MMA session that means your punches on the pads, on the bag and in boxing rounds, with the same numbers as above. It does not mean kicks, knees or elbows. Nothing on your wrist can count a kick, and we do not claim it can. If your striking training has a boxing component, a tracker measures that component. If you want your whole striking output counted, no wrist sensor does that today, ours included.
How to read the numbers well
If you do use one, the research points to a few habits that make the data useful rather than misleading.
Trust the count most. It is the figure with the strongest independent backing. Straights are counted and classified most reliably, so if you want to test a tracker, throw a round of jabs and crosses and check it against your own count.
Read speed and power as trends, not measurements. Your top speed in a tracker's units is a number to beat next week. It is not a laboratory figure, and it will be least reliable on your very fastest punches.
Compare like with like. Same drill, same bag, same round length. Shadow boxing, bag work and pads produce different numbers for the same effort, so a fall in your count that coincides with a change of drill tells you nothing.
Look at rounds, not totals. The total for a session hides the pattern inside it. The drop from round one to round four is the useful figure.
Do not chase the count. This is the sceptics' point and it is worth keeping. If your tracker tells you that you threw 600 punches and your coach tells you that half of them were arm punches, your coach is right and the number was a warning, not an achievement. Use the count to see your work rate fall. Use your coach to see why.
Can and cannot, side by side
| What you want to know | Can a punch tracker tell you? |
|---|---|
| How many punches I threw, per hand, per round | Yes. Counting is its most reliable job, and the one with the strongest independent evidence |
| How my work rate changed from round one to round six | Yes. This is what it exists for |
| Whether I am faster or throwing more than I was three months ago | Yes, as a trend in its own units |
| Whether my left and right hands are doing equal work | Yes |
| What types of punch I threw | Straights reliably. Hooks and uppercuts less so, and later punches in a long combination least of all |
| How fast my fist was really moving, in laboratory terms | No. Use the figure to compare yourself with yourself |
| How hard I hit | Not accurately. No tracker tested tracked impact force well |
| Whether my guard came back | No |
| Whether my feet and weight were behind the punch | No |
| Whether I slipped, blocked, or got hit | No |
| Whether the punch was technically good | No. That is your coach's job |
| Kicks, knees, elbows | No |
Where Corner fits
Since you may be here to decide on one: Corner's trackers are a pair of sensors, one for each wrist, worn under your wraps or on the supplied wristbands. They record every punch through a session with a timestamp, type, hand, speed and power, and the app shows you the round-by-round graphs, the punches per minute, the left and right split, the punch-type breakdown and the longest combination, along with your history over weeks and months. There is no subscription for tracking your own training. The pair starts at $159.
We are one of two trackers the researchers described as likely to give the most valid punch data, and we have published the parts of their findings that are unflattering to us because that is the only version of this article worth reading. The numbers above are also why our own product page lists punch count, work rate, average speed and average power, and not "technique". If you already own a smartwatch and are wondering how the two overlap, that question has its own article. If you want to see the trackers, they are here.
The bottom line
The people who say a wrist sensor cannot see your technique are right, and you should not buy one expecting it to. What it can do, and what independent research has confirmed it does well, is count your work, hand by hand and round by round, and show you how that work holds up as you tire and as the months pass. That is half of getting better at boxing. Your coach has the other half. A tracker does not replace them. It gives them, and you, the numbers neither of you could have counted.
Sources: Omcirk D et al., "Punch Trackers: Correct Recognition Depends on Punch Type and Training Experience", Sensors, 2021; Omcirk D et al., "Validity of Commercially Available Punch Trackers", Journal of Strength and Conditioning Research, 2023. Both studies tested Corner app version 1.3.1. Both were funded by Charles University, not by Corner or any tracker manufacturer.
