Methodology

An impressive win rate isn't enough to judge a strategy. That's why, before publishing anything, we tested six different strategy families — multi-indicator technical (EMA/RSI/MACD/ADX/VWAP), breakouts, liquidity-sweep reversals, funding rate, intraday mean reversion, and momentum. Each one was measured on its real R per trade — actual exits, costs deducted — and tested with rolling windows and out-of-sample on data never seen during tuning. Five families failed this first filter, with a negative average real R; for two of them (multi-indicator technical and intraday mean reversion), a compounded-capital simulation — 1% of capital at risk per position, with up to 102 positions open at once for the first and 24 for the second — confirmed the loss, all the way down to zero. Only momentum got through, and then passed the decisive step: a real compounded-capital simulation with multiple positions open at once — the step that changes everything.

Five of these six approaches didn't survive. One passed every test. That's the one we're offering you.

Simple, fixed rules, applied to the live portfolio from 6 October 2026. Every Tuesday at 00:00 UTC, at the close of Monday's daily candle, we rank the cryptocurrencies in the universe by their return over the last 7 days and keep the 10 strongest. The backtest universe has 41 pairs, including PAXG, a gold-backed token; live, it is the 50 largest cryptocurrencies by market cap, excluding stablecoins, among those tradable on OKX. These 41 pairs were chosen in September 2026, among the cryptocurrencies then listed on Binance, ranked among the largest by market cap, with a price history going back to October 2020 and sufficient trading volume: none that disappeared in the meantime is included, and the list is the same over the whole period. This survivorship bias tends to improve the backtest result; we have not yet measured by how much. The 10 positions share a risk budget of 5% of capital, split by inverse volatility (ATR14): each position's size is that risk divided by the distance between its entry price and its stop. A low-volatility crypto, with a close stop, therefore gets a large position. Sizes are computed on capital as it stood before the previous week's positions were closed (excluding their gains or losses for the week and those of the open overlay); a subscriber applying the published percentages to their account's current value will therefore get slightly different sizes.

Stops and BTC overlay. Each position's stop sits 1.5 times its ATR14 below the entry price. It is checked against each day's low: if hit, the position is sold at the stop price and the money stays in cash until the following Tuesday; the other positions are sold at the next rebalance. Each day, if Bitcoin closes above its 100-day exponential moving average (EMA100), a BTC position worth 50% of capital is added on top of the momentum positions; it is sold on the day Bitcoin closes below that average.

Leverage. These rules do not cap total exposure. In the backtest, median exposure is 87% of capital, but it exceeded 100% on more than one day in three (38% of days) and peaked at 237%, about 2.4 times capital. Replicating the strategy therefore requires an account that allows borrowing (margin) or derivatives, placing stop orders yourself, and adjusting the BTC position mid-week when the overlay switches (Pro subscribers get an email when it does). The backtest deducts 0.30% round-trip fees per position. It does not model the cost of borrowing, slippage when stops execute (an exit can fill below the stop during a sharp drop), or the risk of a margin account being liquidated.

Tested over 3 independent periods covering the 2021 bull run, the 2022 crash and the current market, the strategy rebalanced on Tuesdays is positive in all 3, with +896.5% cumulative over the full period (+48.3% per year) and a maximum drawdown of −52.6%, versus −76.6% for simply buying Bitcoin and −87.8% for a passive crypto basket. These figures include the leverage described above.

Over the full period, the strategy beats both passive benchmarks on all 6 metrics: return, risk and risk-adjusted return. This result depends on the rebalancing day and on leverage: the limitations below spell this out.

Research results (2020-2026 backtest) were calculated using Binance data. Since September 21, 2026, live tracking and allocation calculations use OKX data, consistent with the trading bot's execution platform.

Since late 2022, Bitcoin has returned more. From October 2022 to September 2024, the strategy made +110% while Bitcoin made +203%, with a lower Sharpe ratio (1.13 vs 1.40). From 10 October 2022 to 24 August 2026, it made +257% versus +313% for Bitcoin: its maximum drawdown stays much lower (−33.5% vs −53.0%) and its Sharpe is similar (1.03 vs 1.01), but its return is lower.

The result depends on the rebalancing day and on leverage. With exactly the same rules but the decision taken on another day of the week, the strategy would have returned between +277% and +545% over the same period (Sharpe from 0.77 to 0.99): Tuesday, the day used for the published figures, is the most favorable of the seven days, and on five of the seven, Bitcoin returned more. Without leverage, aiming at no more than 100% of capital at each rebalance, with no adjustment between two Tuesdays (maximum measured exposure: 104%), the same rules would have returned +459%, less than Bitcoin (+505%), with a maximum drawdown of −46%.

Leverage adds its own risks. Borrowing amplifies losses as much as gains, costs interest that the backtest does not count, and a margin account can be liquidated during a sharp move, before stops even execute: losing more than you put in is possible. If you do not want to use leverage, this strategy does not suit you as it stands.

• These figures come from a rigorous historical backtest, not yet a live trading track record — we have been running the strategy under real market conditions since September 3, 2026, and the Portfolio page always clearly distinguishes live performance from research results.
• Past performance, however strong, never guarantees future results.
• This is general information about our methodology, not personalized advice. The final decision on every trade remains entirely yours.