How the strategy works:
RSI Divergence
A public laboratory that documents, in a verifiable way, how a classic technical strategy actually behaves on EUR/USD. Every result — win or loss — is published.
⚠️ Educational experiment · not financial advice. Past results do not guarantee future ones.
01 The core idea
The strategy looks for the moments when the market's push is running out of steam. The tool we use is the RSI (Relative Strength Index), a classic indicator that measures the "strength" of a price move on a scale from 0 to 100.
The signal does not appear when price and RSI agree, but precisely when they contradict each other. That contradiction is called a divergence, and it's the heart of everything.
02 The two signals
Price falls and prints a new low, lower than the last one. But the RSI, instead of following, forms a higher low. Price keeps dropping, yet with less and less force: the decline is tiring, and a bounce often follows.
The mirror image. Price rises and prints a higher high, but the RSI forms a lower high. The climb loses momentum, and a drop often follows.
Every trade has a fixed 5-minute expiry: the entry price is compared with the price 5 minutes later to determine the result, WIN or LOSS.
03 The time window matters (a lot)
This is the most important — and most honest — thing to understand. Testing the strategy on three months of historical data, we found that it does not work the same way at every hour. It only really performs in one specific window:
7:00 AM – 12:00 PM in New York (EST)
It's no accident: this is when the London and New York sessions are open at the same time. It's the moment of highest volume and liquidity in the currency market, when moves are more orderly and "clean" — the ideal conditions for a strategy like this to be at its best. Outside this window, the signals become far less reliable.
04 Why you see signals outside the window too
The system runs 24 hours a day, so you can watch it work whenever you connect. But the signals are labelled:
05 The AI quality read
For every signal, an AI model analyses the chart and assigns a score from 1 to 10.
Important: the AI does not predict whether you'll win. It only judges how well the setup matches the strategy's ideal conditions. Part of the experiment is exactly this — checking, over time and with public data, whether that score really helps tell the better signals apart.
06 Our transparency promise
We don't promise winning trades. We publish every result — wins and losses — without cherry-picking the good ones. This is a public laboratory that verifiably documents how a classic technical strategy actually behaves.