A crypto trading system that follows rules, not hunches.
Four strategies, each backtested before it trades real money, all executed automatically on futures. This site explains how they work and how they perform.
- Ichimoku Trend FollowingLive since Aug 2025
- ATR Mean ReversionLive since Feb 2026
- RSI Mean ReversionLive since Jul 2026
- RSI Red LegsLive since Sept 2026
The strategies
They look for different things: three buy sharp dips and wait for the rebound, one follows long trends. Running them side by side smooths out the bad stretches of any single one.
Ichimoku Trend Following
Rides established trends using the Ichimoku cloud. A daily Bitcoin breakout filter keeps it out of the market when the broader trend is weak.
Backtest since 2018: +105% a year on average, worst drawdown −28%
How it works in detailATR Mean Reversion
Buys sharp sell-offs that are large relative to the coin’s own volatility (ATR) and sells into the rebound, usually at the close of the same day. Up to four positions at a time.
Backtest since 2018: +47% a year on average, worst drawdown −70%
How it works in detailRSI Mean Reversion
Waits for a deeply oversold short-term RSI and scales in over up to three tranches as the dip extends, then exits once price recovers.
Backtest since 2018: +20% a year on average, worst drawdown −36%
How it works in detailRSI Red Legs
Builds a position in up to five equal legs while a sell-off continues, guided by RSI, and closes the whole ladder on the rebound.
Backtest since 2018: +24% a year on average, worst drawdown −34%
How it works in detail
EntryExit
All strategies together
The strategies are meant to run side by side. Their bad stretches rarely come at the same time, so together they lose far less in a drawdown than most of them do on their own. This backtest runs them as one portfolio.
Jan 2018 to Oct 2026. Five equal parts: Ichimoku, ATR Mean Reversion, RSI Mean Reversion, and RSI Red Legs on coins and on QQQ, each with its own rules and limits. Once a month the balance is split back into equal parts. Live trading uses a similar, but not equal, split. Starting balance $50,000, profits reinvested, fees included.
- Average return per year
- +47%
- Worst drawdown
- −20% −13% on closed trades
- Years with a loss
- None out of 9
- Longest time below a high
- 10 months on closed trades
Including open trades at their lowest pointClosed trades only
| Year | Return | Worst drawdown |
|---|---|---|
| 2018 | +9% | −6% |
| 2019 | +51% | −20% |
| 2020 | +46% | −16% |
| 2021 | +252% | −12% |
| 2022 | +1% | −12% |
| 2023 | +6% | −13% |
| 2024 | +82% | −15% |
| 2025 | +50% | −11% |
| 2026 to Oct | +18% | −9% |
Hypothetical results from a simulation on historical data, not live trading. Live results differ, and past performance is no guarantee of future results. How the backtests were made
How a trade happens
A signal fires
Each strategy watches the market on TradingView and sends a buy or sell signal when its rules are met, at the close of a candle or as soon as price hits a level during it.
The bridge checks it
A small server checks each signal before acting: position limits, free margin, duplicates and whether the coin can be traded. Refused signals are logged, never forced through.
The order goes to OKX
The order is placed on OKX Europe and every fill is logged, so results can be checked trade by trade.