Dip progression

The dip ladder starts with a top support hold on the bars, moves through bench dips, and builds to sets of bar dips.

A full-range dip goes from locked arms down until the elbows reach 90° or below, then back to locked arms. A rep that stays shallow, or does not lock out at the top, is named and not counted.

The steps

Each skill is a ladder of steps. Every step has a target — clean seconds for a hold, full-range reps for a set — and you move on when you hit it in one set.

1
Top support hold, 30 sTarget: 30 clean seconds in one set.Arms locked straight on the bars, shoulders down.
Check this step on camera
2
Bench dips, 10Target: 10 full reps in one set.From locked arms down to elbows at 90° or below, and back to locked.
Check this step on camera
3
Bar dips, 5Target: 5 full reps in one set.From locked arms down to elbows at 90° or below, and back to locked.
Check this step on camera
4
Bar dips, 12Target: 12 full reps in one set.From locked arms down to elbows at 90° or below, and back to locked.
Check this step on camera

The targets are common coaching numbers, not a standard. The thresholds for “clean” are starting values, and they are written out on each page.

What counts as clean

Elbow angle (shoulder–elbow–wrist) from locked out to 90° or deeper.

Top support hold

Elbows at least 165° straight (shoulder–elbow–wrist), hands below the shoulders.
The line score starts at 100 and loses 1.25 points per degree the body is off the ideal shape; a second counts when the score is 75 or more — about 20° off in total.

Faults the camera names: “elbows bent”

Dip reps

Every rep starts with the elbow (shoulder–elbow–wrist) at 155° or straighter.
It reaches 95° or less at the bottom.
And it returns to 155° or straighter before the next rep begins.

Faults the camera names: “not deep enough — elbows above 90°”, “arms not locked out at the top”

A fault is named on screen once it costs 8° or more. Reps that fall short are named and not counted.

Setting up the phone

Stand the phone on its side two or three metres away, looking at you from the side, whole body in frame. The pose model runs in the browser; the video never leaves the phone.

What it cannot see

One camera sees one side. It cannot see a handstand drifting towards or away from it, or one arm doing less than the other.
It needs the whole body in frame and reasonable light. When it cannot see a joint, it stops counting rather than guess.
The thresholds — how straight is straight — are starting values from the geometry of each position, not yet tuned against many athletes. Expect them to change.

Check step 1 on camera All progressions

Join the waitlist

We'll email you when your place is ready. Nothing else — no newsletter.