CHESS · POKER

One thing. Until it's fixed.

One Move reads your own games, finds the one gap that's costing you most against players a Band above you, and gives you a number to beat. A Review every Monday, until you beat it twice.

No account. Your last 30 games. About 3 minutes.

Free report · 5 USD/month for the weekly loop

Play poker? Upload PokerStars hand histories instead

EXHIBIT A · WEEK 01MON 2 NOV

CHESS · RAPID · 1400 · CHRISGOODKNIGHTS

Conversion from +2

45%71%Target
30 gamesBelow target
5 drillsNext Review Mon 9 Nov
Exhibit A. The maker's own report. Target: median for the 1600–1800 Band, Lichess open database, Sep 2026.

The Constraints One Move finds

CHESS
  • Conversion from +2
  • Blunders on fast moves when winning
  • Blunders with under 60 s left
  • Blunders in moves 1-10
  • Endgame blunders
POKER
  • Opening too wide from early seats
  • VPIP–PFR gap
  • Continuing after betting and being raised
  • Blind defence

THE WEEK CARD

Game review is per game. Improvement is per player.

Game review tells me what went wrong in one game. It can't tell me what keeps going wrong in all of them. So I built the thing that does.

— THE MAKER, 1400 RAPID

  1. [01]Evidence. Your last 30 rated games.
  2. [02]Constraint. The one gap furthest below the Band above you.
  3. [03]Target. That Band's median. A real number to beat.
  4. [04]Review. Every Monday, 7am, by email.
  5. [05]Drill. Five positions from your own games. Predict first.
  6. [06]Broken. Hit the Target twice in a row. We pick the next one.
THREE VERDICTS, ONE CARDBelow targetAt target · 1 of 2Broken
CONVERSION FROM +2 · MEDIAN BY LICHESS RAPID BAND
BandMedianNote
600–80060%
800–100060%
1000–120070%
1200–140067%
1400–160067%EXAMPLE BAND · MAKER: 45%
1600–180071%← TARGET FOR 1400
1800–200067%
2000–220071%
2200–240075%
2400–260073%
2600–280078%
Share of games that reached +2.00 and were won. About 3,000 players' rated games, Lichess open database, September 2026. These are Lichess ratings; chess.com runs a few hundred points lower at club level.

03 · TARGET

The Target is a median, not a vibe.

We don't invent a target. We take the number the median player one Band above you actually hits. At 1400 the maker converts 45% of his winning positions; at 1600–1800 the median player converts 71%. That's the gap.

05 · DRILL

Your mistakes, not someone else's puzzles.

You predict how you'll do. Then you find out. The gap between the two is the part that transfers.

BEFORE · PREDICTION

How sure are you that you'll find a good move? 60%

AFTER · GRADED BY CODE

Not quite.

You played Rf2. Best was Qe3, offering the queen trade while you're a piece up. You were 60% sure, and you didn't find it.

Move 24: you played Rf2 at +2.10 and lost 1.90.
EXAMPLE · Stockfish 17, depth 18. From one of your own 30 games.

It ends. That's the point.

WEEK 07 · EXAMPLEMON 14 DEC

CHESS · RAPID · 1400

Conversion from +2

73%71%Target
26 games · at Target 2 of 2
Next Constraint
Blunders with under 60 s left
BROKEN
  1. Two Reviews in a row, neither one skipped.
  2. The Target met in both.
  3. The Constraint is Broken. The next one is picked from your games.

Most training is a pile of advice with no finish line. This is a queue of one.

POKER

Poker, same loop.

Upload PokerStars cash-game hand histories, at least 300 Hands. Same six fields, one method, two games. Your $5 covers both.

WEEK 01 · EXAMPLEMON 2 NOV
POKER · 6-MAX CASHOpening too wide from early seats
42%≤38% TARGET
1,840 Hands · 312 spotsBelow target
5 drillsNext Review Mon 9 Nov
EXAMPLE · Hands opened from UTG and HJ. Filled = opened.

Hand #2481 · UTG · K♠9♦ · raise 3bb · fold to 3-bet · −3bb

  • Your screen name only.
  • Opponents' names are stripped as the files are read.
  • The files are deleted after parsing.
  • Your report is private unless you share it.
  • Get a free poker report
From: One Move · Mon 9 Nov, 7:00Your chess Review: Conversion from +2

52% / 71% target · 28 games

You're relaxing a move too early. In 6 of your 13 lost +2 positions, the swing came on a move you played in under 5 seconds.

This week's Drill set is five of those positions. Take ten seconds before each move.

Last week you scored 3/5; you predicted 4/5 at 60%.

Do this week's Drill set
EXAMPLE REVIEW · WEEK 02

04 · REVIEW

Monday, 7am.

The email is the cue. There's no app to remember and no streak to maintain.

Some weeks you won't play 10 games (or 300 Hands). We'll send a Skipped review, carry the Evidence over, and not pretend it counted.

Stockfish, the hand parser and the drill grader are code. A language model writes the one short paragraph, from the computed numbers.

PRICING

$5 a month. Both games. Cancel any time.

One plan. Five dollars.

USD $5/MONTH
The weekly Review, the Drill set, chess and poker, and your whole history.
FREE
The whole first report. No account.
SMALL PRINT
Billed by Stripe. Cancel from your account. If you stop, your history is kept.
Get my free report firstSubscribe from any Free report.

We give a straight answer.

Is this AI?
Mostly code. Stockfish, the parser, the Constraint picker and the grader are code. A language model writes one short weekly paragraph from the numbers. It doesn't chat or coach.
How is this different from free game review?
Game review is per game. This looks across 30 games, names the one pattern, and keeps score week after week.
What does Broken mean?
You met the Target in two non-skipped Reviews in a row. We pick the next Constraint from your games.
What if I don't play enough?
Under 10 games or 300 Hands, the week is a Skipped review. It's still sent, it doesn't count, and the Evidence carries over.
Do I need an account?
Not for the Free report. Subscribing creates one from your Stripe email; you sign in with a link.
What happens to my data?
Chess reports are public at their link. Poker reports are private, opponents' names are stripped, and the files are deleted.
Why $5?
It covers the analysis and one email a week. There's nothing else to sell you.
Can I cancel?
Any time, from your account or any email. Your history stays.

One thing. Until it's fixed.

Free. No account. The whole report, not a preview.