A/B/C test on a BTC/ETH 50/50 rebalance bot, Aug 2025 divergence window (ETH +24%, BTC -7%):
Variant A — 2% threshold + 1min timer: 3 trades, -2.10% ROI
Variant B — 5% threshold, no timer: 1 trade, +0.13% ROI
Variant C — 1% threshold + 30min timer: 12 trades, -2.05% ROI
The pattern: wider threshold + no time trigger = least damage. Tighter triggers forced repeated sells of ETH into a falling BTC - 12 times for Variant C.
HODL benchmark for the same window: ~+8.5%. All three variants underperformed it. 📊
Rebalancing needs mean reversion to work. This was a directional breakout, not noise - and the bot couldn't tell the difference.
Backtest before risking capital...
#Bitcoin #CryptoGates #Altcoins
Variant A — 2% threshold + 1min timer: 3 trades, -2.10% ROI
Variant B — 5% threshold, no timer: 1 trade, +0.13% ROI
Variant C — 1% threshold + 30min timer: 12 trades, -2.05% ROI
The pattern: wider threshold + no time trigger = least damage. Tighter triggers forced repeated sells of ETH into a falling BTC - 12 times for Variant C.
HODL benchmark for the same window: ~+8.5%. All three variants underperformed it. 📊
Rebalancing needs mean reversion to work. This was a directional breakout, not noise - and the bot couldn't tell the difference.
Backtest before risking capital...
#Bitcoin #CryptoGates #Altcoins