SPX 0DTE morning-trend credit spread
Each morning at 10:30 ET it sells a same-day SPX credit spread on the side the index has moved away from since the open, and holds it to expiry unless a stop at twice the credit fires first.
- Total return
- +25.4%
- A year
- 5.8%
- Max drawdown
- 3.2%
- Trades
- 882
SPX, 14 Sep 2022 to 14 Sep 2026, starting with $100,000. Simulated, after slippage and commission.
What it does
It lets the first hour play out. At the 09:30 ET open it notes where SPX is, and at 10:30 it looks at which way the index has gone since.
Then it sells against that move. If SPX is up, it sells a put credit spread below the market; if SPX is down, a call credit spread above it. The short strike goes 1.25 times the at-the-money straddle away from spot, so it sits further out on days the options market expects a bigger move, and the long strike sits 50 points beyond it.
It trades one contract, at a limit of the two legs' mid prices less 0.10. Once that fills, it places a stop to buy the spread back at twice the credit. There is no profit target and no time exit: a spread the stop never reaches is held to expiry and settles in cash.
That makes at most one trade a day, and none on a day SPX is exactly where it opened.
Build it yourself
This is the prompt, word for word. Sent to Xavi in a new conversation on 22 Oct 2026, it wrote the strategy, saved it as “SPX 0DTE Smooth Morning-Trend Spread” and ran the backtest on this page. Paste it into your own conversation and you get the same strategy.
Build an SPX 0DTE strategy called "SPX 0DTE Smooth Morning-Trend Spread". Follow these rules exactly and do not add anything: Universe: SPX and SPXW options (name both roots), same-day expiry only (0 DTE). Stream the 25 nearest strikes around spot. Check every minute. Each trading day: 1. Record the SPX price at the 09:30 ET open. This is the reference price. 2. At 10:30 ET, try one entry per day. Skip the day if there is no reference price, if a position or order is already open, if the chain is empty, or if SPX is exactly at the reference price. 3. Compute the ATM straddle: the mid of the at-the-money call plus the mid of the at-the-money put, today's expiry. 4. If SPX is above the reference price: sell the put with the strike nearest to spot minus 1.25 × straddle, and buy the put nearest to 50 points below that short strike. If SPX is below the reference price: sell the call with the strike nearest to spot plus 1.25 × straddle, and buy the call nearest to 50 points above that short strike. Skip if a leg has no mid, or if the long leg is not on the far side of the short leg. 5. Enter as one 2-leg credit spread, 1 contract, limit credit = short mid − long mid − 0.10. Skip if that is 0 or less. Use a DAY order. 6. When the entry fills, place a GTC stop order to buy the spread back at a debit of 2 × the credit received. 7. No profit target and no time exit. Otherwise hold to expiry and let it cash-settle. Plot the reference price, the 10:30 spot, and the short and long strikes. Log the entry credit and the stop level. Then backtest it from 2022-09-14 to 2026-09-14 with $100,000 starting capital, the Slippage fill model at 0.05 per leg, and $0.65 commission per contract.
The four-year backtest needs an account. Without one, the free preview takes a shorter version of the same rules and tests it over recent data:
The backtest
On SPX minute data from 14 Sep 2022 to 14 Sep 2026.
- Total P/L
- $25,404
- Ending equity
- $125,404
- Max drawdown
- 3.2%
- Sharpe
- 1.37
- Sortino
- 1.94
- Trades
- 882
- Win rate
- 60.4%
- Profit factor
- 1.21
The fills used the Slippage model at $0.05 a leg, with $0.65 commission per contract, so $13,408 of slippage and $1,599 of commission are already taken out of every figure here.
Year by year
| Year | P/L | Return |
|---|---|---|
| 2022 (from 14 Sep) | +$775 | +0.8% |
| 2023 | +$7,459 | +7.4% |
| 2024 | +$4,747 | +4.4% |
| 2025 | +$3,235 | +2.9% |
| 2026 (to 14 Sep) | +$9,188 | +7.9% |
What the backtest does not tell you
What it ties up, and what it lost at worst
At its peak the run had $5,000 of margin in use, 5.0% of its starting capital. Its largest loss on a single trade was $795.10.
One spread can lose $5,000
A spread 50 points wide can lose 50 points, less the credit it took in: $5,000 a contract on SPX. The stop is what kept this run's losses smaller, and a stop is not a guarantee.
The stop fills at the market, not at its price
The stop arms when the spread's mid price reaches twice the credit, and then buys the spread back across the bid and the ask. On a same-day spread in a fast market that can be well past the trigger, and a real market can be worse than the minute the backtest sees. How the backtest fills a stop →
It is one stretch of history
Every whole year in this run made money, but not evenly: the best made +7.4% (2023), the worst +2.9% (2025). Nothing says the next four years will look like these.
Nobody traded this
The results come from replaying historical option quotes, with the costs above. This page shows what the software does with a prompt. It is not advice to trade the strategy.
Build your own
Describe a strategy the same way and Xavi writes it, backtests it and keeps every version, so you can see what each change did. The data every backtest runs on comes with the account.
Simulated results, not a track record.