The Rules Behind
Our Trading Bots.

Every bot in the fleet trades defined-risk $1-wide spreads, where the amount risked is roughly the amount to be made — on the order of risking about $50 to make about $50 per contract. That symmetry is the design, not an accident of parameter tuning. It means no single trade is ever large enough to matter much on its own, and the results come from repetition across hundreds of small, identically-shaped bets rather than from any one position going right.

None of the bots run stop-losses, because none of them need one: the maximum loss is baked into the structure of the position at entry. A $1-wide spread cannot lose more than the width of the spread, less whatever was collected or paid up front, no matter what the underlying does afterwards. The worst case is known to the penny before the order is ever sent.

The practical consequence is that a fast move against an open trade demands no reaction. No monitoring the position through the session, no scrambling to close, no judgement call made under pressure with the clock running. The worst case was accepted — and capped — before the trade existed, which is precisely what lets these systems run mechanically from open to close without a human in the loop.

The contrast above is the whole argument in two boxes. Traditional credit-spread strategies sell far out-of-the-money spreads for their headline win rates — 90%, 95%, sometimes higher — by collecting a small premium a long way from the money. The trade-off is hidden in the ratio: a wide spread sold for $30 to $80 of credit still has the entire spread width, often $500 to $1,000, at risk behind it. One fast move puts all of it on the table at once.

In that design the stop-loss is the survival mechanism. It is the only thing standing between a routine losing day and a loss many multiples of the credit collected. And it has to execute perfectly in exactly the moments markets are at their worst — gapping, widening, thin on the bid — which is when stops are least reliable and slippage is largest. Months of small wins can disappear in a single bad exit, and the strategy's high win rate says nothing at all about that risk, because the win rate was never the number that mattered.

Our bots invert the trade-off deliberately. Published backtest win rates sit in the mid 60s to low 80s rather than the 90s, and the amount at risk on a trade is close to the amount to be made rather than many multiples of it. In exchange, the worst case on any one trade is already far smaller than a single far-OTM blowup — which is exactly why they can skip stop-losses entirely. There is no exit that has to fire correctly under stress, because there is no catastrophic outcome to defend against.

Figures above are illustrative of the two structures described, not a forecast or a quote. Nothing here is investment advice.

The pattern-day-trader rule was eliminated in 2026. Day trading no longer requires $25,000 sitting in an account, and the old four-trades-in-five-days tripwire that shaped a generation of retail strategy design simply is not there any more. For anyone building short-dated systematic strategies, that changed the arithmetic completely.

What used to demand $25k concentrated in a single account can now be spread across several small accounts of $1,000–$2,000 each, all day trading the same tickers — something that was structurally impossible before. The capital floor that kept these structures out of reach for smaller accounts is gone, and with it the reason most 0DTE spread strategies were written for larger balances in the first place.

These bots were designed for exactly that world. They start at one contract on small balances and scale as the account grows, on a rule of thumb of roughly one contract per $1,000 of account value. Because the risk per contract is small and fixed, that scaling stays flat and linear — the position size follows the account, and nothing about the mechanics of the strategy changes between one contract and ten.