You did the hard part. The phone is leaned against a water bottle, you got through your working set knowing it was recording, and now there's a 40-second video of you squatting sitting in your camera roll.
You watch it once. It looks... like you squatting. Something feels slightly off around the third rep but you couldn't say what. You watch it again. Same. You put the phone away.
That's where form checks die — not at the filming, at the watching. And it's strange how little advice exists for this half. Search for how to film your squat and you'll find a dozen careful explanations of camera angles, tripod heights, and lighting. Search for what to actually look at once you have the footage and it thins out fast.
So here's that list. But first, the reason your last attempt didn't work.
"Is my form good?" is not a question
It feels like one. It has a question mark. But it isn't the kind of thing 40 seconds of video can answer, because it isn't asking about anything in particular — it's asking for a verdict, and verdicts need a judge.
What footage is genuinely excellent at is answering spatial questions with a yes or a no. Where was the bar. How low did the hip go. Which way did the knee travel. Those have answers you can point at with a finger, and they're answers you cannot get any other way, because you can't feel your own form — the corrections that matter happen in about 200 milliseconds and never reach the part of your brain that narrates the set.
So the fix is not to watch harder. It's to walk in with one specific question and let the video answer that one thing.
Decide the question before you press record
One per clip. Not a checklist you run down while the video plays — a single thing you're checking, chosen before the set, so you're looking for evidence instead of browsing for problems.
Rotate through them across sessions. Here's the order I'd suggest, roughly by how much each one changes when you find it.
1. Depth — and it's one frame, not a feeling
Scrub to the single frame where your hips are lowest. Not the general bottom of the rep; the one frame. Then look at where your hip crease sits relative to the top of your kneecap.
That's it. It's binary, it takes about four seconds, and it is the single most common gap between what a lifter believes and what the camera recorded. People miss depth by margins they'd swear were impossible — the sensation of "all the way down" arrives well before the position does, because your body's sense of where it is in space gets much worse under load.
Check the same frame on your last rep too. They're often not the same squat.
2. Bar path — watch the bar, not yourself
This one is hard for a specific reason: you will watch your own body. Everyone does. Force yourself to track only the bar and let the rest of the frame blur.
You're looking for whether the bar stays stacked over your midfoot on the way down and the way up. A bar that drifts forward out of the bottom is the most common single cause of a squat that turns into a good morning, and it is completely invisible from the inside — it feels like the weight got heavy, not like the bar moved.
3. Which rep stops looking like rep one
This is the one almost nobody checks, and it's the most useful thing on the list.
Play rep 1 and then the last rep back to back. Not the whole set — just those two. Most people's sets have a rep where the technique quietly changes: the chest drops a little, the depth comes up an inch, the tempo goes ragged. Finding which number that is tells you more about how to program your sets than any cue will. If your form falls apart on rep 4 of 5, the answer isn't a better cue. It's a different set.
There's a related version of this that's worth its own session: that grinding last rep is a measurement, not a method, and video is how you find out what it actually costs you.
4. Knee travel, honestly assessed
Knees drifting inward on the way up. Worth checking, worth fixing if it's dramatic — and worth being relaxed about if it's slight and symmetrical, because a small amount is normal and chasing it can make people squat like a robot. This is a real thing to look at, not a moral failing to catch yourself in.
The one rule: full speed first
If you take nothing else from this: watch it at normal speed before you ever slow it down.
At 0.25x, every rep looks broken. You can watch a knee travel a centimetre over three seconds of playback and become convinced you're collapsing. Slow motion is a magnifying glass, and everything under a magnifying glass looks alarming — including things that are completely fine.
So: full speed once, decide whether anything is actually wrong, and only then slow down the specific moment you want to examine. This order matters enough that it's worth writing on your hand.
The clip is not the point. The pair of clips is.
Here's the thing that turns this from a chore into an instrument.
A single video tells you where you are. It cannot tell you whether that's better or worse than last month, which is the question you actually care about. And the way almost everyone uses their footage — film it, watch it, delete it — guarantees you never find out. Every form check becomes a snapshot with nothing to compare against.
Keep one clip per lift per month, same angle, same rough weight, in one folder. That's twelve videos a year and it's the entire archive you need. When you put August next to May, the change is obvious in a way that no single video and no amount of remembering ever is. Progress in technique is slow and unremarkable frame to frame; it's dramatic across a season.
This is the same problem as not remembering what you lifted last week, pointed at your body instead of your numbers. The information exists. It just doesn't survive unless something keeps it.
What this looks like with less effort
Everything above works with a phone and a folder, and you should do it that way if that's what you'll actually do. The honest version of this advice is that most people won't — not because they're lazy but because the scrubbing, the frame-hunting, and the remembering-to-keep-it are exactly the parts that are tedious enough to skip on a Tuesday.
That's the part we built Flexion to take over. You record the set, and it goes through the footage frame by frame afterward — finding the bottom position, tracking the bar, flagging the rep where things changed — and keeps the clips side by side so the comparison is already made the next time you squat. Not a coach yelling cues in real time, which we think is mostly theater. Just the tedious part, done.
But the list is the list either way. Pick one question. Watch it at full speed. Keep the clip.
Related: there's no such thing as perfect form, and you've already watched enough squat tutorials — the gap between watching someone else lift and seeing yourself lift is most of the problem.
