Atlas
My own trading platform, built with AI coding agents.
PAPERP-003 runs through Nov 20.
Why I'm building it
By hand
I entered trades on emotion.
Atlas
Runs my strategy the same way, every single time.
I started trading forex and crypto in 2018 and lost money by hand. I never traded one strategy long enough to know its real win rate.
Where it stands
Rules Done
Experiment Ongoing
Paper Now
Live Not built
What I built
- Strategy Studio
Run a test.
- Experiments
Compare tests.
- Trade panel
Replay any trade.
- PAPER page
Follow a paper run.
Paper run P-003
- Run 1: GBP/USDOct 25 to Nov 6
- Run 2: GBP/JPYNov 8 to Nov 20
- Account
- OANDA practice
- Target
- 1.5R (1.5 times the risk)
“Paper tests whether Atlas runs my strategy exactly as written, not whether the strategy makes money.”
EXPERIMENTAug 2025 to Sep 2026
| Pair | Trades | Won | Net |
|---|---|---|---|
| GBP/USD | 769 | 28.2% | −230.9R |
| GBP/JPY | 783 | 28.6% | −228.0R |
Every trade
P-003GBP/USD
PAPER12 trades, 5 won, net +0.55R.
P-002
PAPER1 trade, 0 won, net −1.0R.
P-001
PAPERNo trades.
How it got here
- Jul 31First version started
- Sep 10Rebuilt from scratch
- Sep 27First paper run
- Sep 28First paper trade: stopped out at −1R
- Oct 18Coding agents restricted from trading
- Oct 22GBP/USD cleared for paper
- Oct 25P-003, the first monthly PAPER run, starts
Backtest and paper results are simulated. Past results, including live ones, don't predict future results. Atlas is a tool. Nothing here is financial advice. I am not a registered financial or investment adviser.