I've run grid bots on ranging pairs for years. The pattern is always the same: months of smooth 5-10% monthly returns, then one trending week erases a year.
Is there any configuration that survives long-term — capped levels, equity stops, regime filters — or is the blow-up mathematically guaranteed and I'm just trading the timing?
Guaranteed, given enough time. A grid is short volatility and short tail risk — the equity curve is a carry trade: steady carry, occasional -80% day.
You can extend lifespan with a hard equity stop at -25%, trading only inside measured ranges, and capping grid levels. But you are trading the timing of the blow-up, not avoiding it. Size accordingly.
What worked for me: treat grid profits as withdrawals, not compounding. I skim 50% of profits monthly into a separate account. When (not if) the grid dies, I've already banked more than the account loses. Changes the psychology completely.
Add a regime filter: kill the grid when price breaks a 200-day range or when realized vol doubles its 90-day median. Ugly chop is survivable for a grid; a repricing trend is not. This one filter would have saved every grid blow-up I've witnessed.