【🔄 Multi-Head Box Auto Rotation 09-17】
▬ The system uses the same set of rules every 7 days to run rolling-window backtests on 16 coins (4 windows: to verify whether the strategy truly makes money or only performs in a specific market phase), then automatically rotates coins.
▬ Coins selected this round: $ZEC , $ETH
▬ Scoring (number of profitable windows / total profit / number of trades):
· ZEC 2/4 windows +7,754U 66 trades
· ETH 3/4 windows +1,299U 39 trades
· BTC 3/4 windows -48U 23 trades
· BNB 2/4 windows -684U 45 trades
· DOGE 1/4 windows -824U 24 trades
· TRX 2/4 windows -2,132U 52 trades
· SOL 2/4 windows -2,307U 29 trades
· ADA 2/4 windows -2,476U 24 trades
▬ Selection threshold: In at least 2 of the 4 windows it must be profitable, total profit must be positive, and trade count ≥ 6. The threshold looks simple, but it filters out most coins that look great in backtests but fall apart live.
▬ Why rotation is required: The expectation difference of the same rules across different coins is 10x or more. The market structure of coins changes—choosing once and locking it in is the laziest way and also the easiest way to lose money.
Data source: 4h box detection + rolling-window backtesting. The method and thresholds are public and reproducible. Not investment advice.
▬ The system uses the same set of rules every 7 days to run rolling-window backtests on 16 coins (4 windows: to verify whether the strategy truly makes money or only performs in a specific market phase), then automatically rotates coins.
▬ Coins selected this round: $ZEC , $ETH
▬ Scoring (number of profitable windows / total profit / number of trades):
· ZEC 2/4 windows +7,754U 66 trades
· ETH 3/4 windows +1,299U 39 trades
· BTC 3/4 windows -48U 23 trades
· BNB 2/4 windows -684U 45 trades
· DOGE 1/4 windows -824U 24 trades
· TRX 2/4 windows -2,132U 52 trades
· SOL 2/4 windows -2,307U 29 trades
· ADA 2/4 windows -2,476U 24 trades
▬ Selection threshold: In at least 2 of the 4 windows it must be profitable, total profit must be positive, and trade count ≥ 6. The threshold looks simple, but it filters out most coins that look great in backtests but fall apart live.
▬ Why rotation is required: The expectation difference of the same rules across different coins is 10x or more. The market structure of coins changes—choosing once and locking it in is the laziest way and also the easiest way to lose money.
Data source: 4h box detection + rolling-window backtesting. The method and thresholds are public and reproducible. Not investment advice.