Early access

Match analysis that shows its working.

Put a phone on the balcony, upload the recording, and get court coverage, T control, recovery and work rate — measured during the rallies only, never across the walking about in between.

No app to install. Your phone footage is enough.

One player’s real movement in a 17-minute match, plotted only during rallies

Automatic court detectionMeasured during rallies onlyReports how much it could not read

What you get back

This is what lands in your inbox

One page per match and per game: where the players stood, how the rallies ran, and — every time — how much of the footage the analysis could actually read.

Thursday club night — Game 3Ready

Position heatmap — rallies only

Play time

80.4%

Court in frame

92% best

T control

61%

T at rally end

57%

1

The heatmap that means a place

Positions on a real court in metres, because the court was calibrated from its own painted lines first.

2

Impact before effort

Who held the T as rallies ended and how deep each player was pushed — the numbers that change how you train, with distance and work rate below them.

3

Every number, per game

A club evening is five or six games in one file. Each is found from the breaks and reported on its own.

4

An honesty line

Play time, detection rate and how much of the court was in frame, on every report.

One real 17-minute club match, analysed with no human input

55

rallies found, each one from its serve

64.9%

of the recording was play — the rest was not squash

98.8%

of samples were real detections, not filled in

879 m

covered during play, at 80.3 m per minute

Three things to get right when filming

You film it. Everything after that is automatic.

Full filming guide →
  1. 01

    Put the phone down and leave it

    Balcony or behind the back wall. A still camera is required — the court is found from frames across the whole clip.

  2. 02

    Get the whole floor in frame

    Both back corners if you can. The floor is what the court is found from, and what is out of frame cannot be measured.

  3. 03

    Walk off between games

    Both players leaving the court is how one game is told from the next. Stay on and a whole evening reads as a single game.

Why the numbers here mean something

Everyone publishes heatmaps.
Nobody publishes what they couldn’t see.

A squash video is not squash all the way through — roughly a third is the knock-up, the walk between points, and one player packing up while the other hits alone. Average across all of it and “work rate” describes neither playing nor strolling.

The usual

One warm blob over the whole floor. No way to tell an hour of squash from an hour of standing about.

Back to the T

17% of back court — out of frame

Measured rally by rally, with play time, detection rate, and any blind spot stated on the same page.

Built to be trusted with the answer

It says how much it saw

Play-time share, detection rate and court-in-frame on every analysis, not buried in a methods note.

It stops when unsure

If the court cannot be found, the analysis stops rather than guessing. A calibration that is quietly wrong puts every number in the wrong place while looking perfectly reasonable.

A club evening, split by game

Upload the whole evening in one file. The games inside it are found from the breaks, and every number is reported per game as well as overall.

No footage kept

The video is deleted once analysis finishes. What stays is where players moved, the rallies and the statistics — no images of anyone.

Who’s behind it

Built by one player, at home.

“Every squash tool I tried gave me a confident number and no way to check it. I wanted one that would rather tell me what it couldn’t see.”

V — founder, Back to the T. Filmed on two real courts in two buildings; every figure on this page is from those runs.

Try it on your next match

Back to the T is in early access while the analysis is proven on more courts. Request access and you get your own workspace — your recordings stay in it.

Request access