> ## Documentation Index
> Fetch the complete documentation index at: https://docs.askfutures.com/llms.txt
> Use this file to discover all available pages before exploring further.
# For options traders
> Use AskFutures as the underlying-behavior tool behind your options book — what to type, what you get back, and the four requests it will refuse.
AskFutures backtests **futures, not options**. There are no chains, no Greeks,
no implied vol. What there is: six years of minute-level data on the
underlyings your options settle into, and a plain-English way to ask what those
underlyings actually do over your DTE, in your vol regime, at your hedge band.
This page is the translation layer.
Past performance does not guarantee future results. Always test before you trade.
Want the long read? [The options trader's series](/tutorials/options-traders-series)
walks through each row below as its own part, with the reasoning.
## What it will refuse
These come back as "not supported" — there is no options data or economic
calendar behind the engine. Each has a futures-side question inside it; the
table below shows how to ask that instead.
* `sell 15 delta strangles, 1 week out` — no options legs of any kind
* `sell $3 wide call spreads when CL is up $2` — no spreads
* `when IVR is above 50%` — no implied-vol series
* `on FOMC days` / `on NFP release days` — no economic calendar (the strategy
still builds, on **every** day; the assumptions card says so)
## Say this, not that
Left column is what an options trader reaches for first. Middle is what to
type. Right is what comes back and how to read it through an options lens.
| You'd say | Type this instead | What you get, and how to read it |
| ----------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| *What's the realized 1-week move on crude?* (strike selection for a 7-day strangle) | `Show me the distribution of the weekly high-minus-low range on CL since 2020, in dollars per contract` | A percentile ladder. Put your short strikes against the 84th/16th (one sigma) and 95th/5th (the tails). If the chain's implied move sits outside the 84th, you're being paid for a move it usually doesn't make. |
| *How wide is the RTH session?* (0DTE) | `Show me the distribution of the 09:30 to 16:00 range on ES since 2022, in points` | Same ladder, session-scoped. Ask in the units you trade — points, ticks, dollars, percent. |
| *What has it done lately, not on average?* | `List the last 20 overnight gaps on NQ` | "List" returns the raw values instead of a distribution — the recent regime at a glance. |
| *Sell put spreads after a 20-day low — does it bounce?* | `On ES daily bars, buy at the close when the close is below the lowest low of the prior 20 days. Hold 5 days, 2× ATR(14) stop. Long only. Since 2020` | A backtest of the futures position your spread is a proxy for, held for your DTE. Win rate ≈ how often you'd finish on your side of the strike; avg win vs. avg loss is the asymmetry to size width against. |
| *Add the call-spread side* | `Add the short side: sell when the close is above the highest high of the prior 20 days, same hold and stop` | The mirror is built from the opposite boundary (highest high), not a copy-flipped operator that fires every bar. |
| *Is the setup an edge at all, or am I just selling premium?* | `Show this strategy's entry condition as a setup scorecard` | Forward-return distributions of the raw condition — no exits, no stops — vs. the unconditional baseline (1h and 5h intraday; 10 sessions daily). If conditional looks like baseline, there's no directional edge; move to the vol row below. |
| *Only when IV rank is above 50* | `Only trade when the 14-day normalized ATR is above its 80th percentile over the last 250 days` — then your rule | A rolling percentile of realized vol: the same shape as IV rank, computed on what the underlying actually did. Combine with a z-score fade to test "sell the spike, lean toward reversion". |
| *Only in a high-VIX regime* | `Only take trades when VX is above its 50-day simple moving average` — then your ES rule | Cross-market gates work: the condition reads VIX futures, the trade is in ES. `VX below 16` and `VX above its 20-day high` work too. |
| *After a squeeze, does the breakout follow through?* (long straddle) | `On GC daily bars, when the Bollinger Band width (20, 2) is below its 50-day median, buy a close above the upper band and sell a close below the lower band. Hold 10 sessions, 2× ATR stop` | The realized-path half of a long-vol trade over your expiry. No theta or vega here — only whether price moves. |
| *On NFP Fridays, trade the 8:30 range break* | `On MES 1-minute bars, Fridays only: buy a break above the high of the 08:30–08:45 range and sell a break below its low. Stop 1× the range, target 2×, only between 08:45 and 11:00, first trade of the day, flat at the close` | Weekday gate + clock-anchored range + time window are all native — the calendar isn't, so this is **all** Fridays. If it pays across all Fridays, the morning structure is the edge, not the print. Clock times are US/Eastern. |
| *FOMC afternoons* | `On ES 1-minute bars, buy a break above the high of the 14:00–14:05 range and sell a break below its low, only between 14:05 and 15:30, max 2 trades per day, 20-tick trailing stop, flat at the close` | Eight event days in \~250. Run the same rule at 10:00–10:05 to see whether 2pm is special at all. |
| *Gamma-scalp around my long straddle* | `On CL 5-minute bars, buy when the 20-bar z-score of the close is below −2 and sell above +2. Exit when it crosses back through 0. Max 4 trades per day, alternate directions, flat at the close. $1.25 commission per side, 1 tick slippage` | The futures leg of a long-gamma book. Total P\&L is what the hedging leg costs or earns; average hold is your realized hedging frequency; max consecutive losses is the whipsaw streak. |
| *Delta-hedge at the wing — what does the whipsaw cost?* | `On ES 1-minute bars, buy when the close crosses above the prior session's high and exit when it crosses back below; short when it crosses below the prior session's low and exit when it crosses back above. Max 6 trades per day, flat at the close` | Every trade is a hedge you'd have put on and taken off at that band. Trade count is how often; P\&L is the cost. |
| *When should I roll?* | `On CL daily bars, buy at the close above the 20-day SMA and exit 3 days before roll` — compare with `exit 1 day before roll` | "Days before roll" and "days before expiration" are conditions; AskFutures knows each contract's dates. |
| *At what band width does the whipsaw beat the premium?* | `Propose an optimization over the z-score threshold and the stop multiple` — then `Run it` | Ranges centered on your values, up to 500 combinations, nothing runs until you approve. Look for a plateau, not a peak; sweep the hold (3, 5, 10 sessions) as your DTE ladder. |
| *How far past my short strike does it go before coming back?* | `Show me the trades, with MAE and MFE, and the 10 worst` | **MAE** (max adverse excursion) per trade is the column options traders are missing: the 95th percentile of MAE is your wing. MFE is where a take-profit or an early assignment would have hit. |
| *Reduce commissions to \$0.75 a side* | Say it **when you create** the strategy: `… assume $0.75 commission per side and 1 tick slippage` | Commission is dollars per side; slippage is ticks per round turn. |
| *Test one year back* | `Since 2020` or `last 3 years` | The default is already the trailing 12 months, so "1 year back" changes nothing. |
## Ground rules
* **Name the futures root**, not the ETF or index: SPX/SPY → `ES` or `MES`;
QQQ → `NQ`/`MNQ`; USO/crude → `CL`/`MCL`; GLD → `GC`/`MGC`; TLT → `ZB`,
`ZN`; VIX → `VX`. Ask *"what symbols do you support?"* for the roster.
* **Clock times are US/Eastern.** NFP is 08:30, FOMC is 14:00, the RTH open
is 09:30.
* **"Hold for 5 days"** is a calendar-day hold on intraday strategies and a
5-bar hold on daily ones. Say *sessions* if you mean sessions.
* **One strategy per chat.** Edits create a new version in the same chat; a
second strategy opens a sibling chat.
* **Every backtest counts** against your monthly allowance — a 500-combination
sweep is 500 of them. Propose first, trim, then run.
## A worked example: pricing a short strangle's wings
```text Your prompt theme={null}
Show me the distribution of the 5-day close-to-close move on CL since 2020, in dollars per contract
```
Note the 84th and 95th percentiles. Those are your one-sigma and tail.
```text Your prompt theme={null}
On CL daily bars, only trade when the 14-day normalized ATR is above its 80th percentile over the last 250 days. Sell at the close when the close is more than 2 standard deviations above its 20-day mean, buy when more than 2 below. Exit after 5 sessions or on a 2× ATR stop. Since 2020, $1.25 commission per side
```
Profit factor above 1 means the underlying tends to revert after spikes —
the directional half of your thesis holds.
```text Your prompt theme={null}
Show me the trades with MAE, sorted by the worst
```
The 95th-percentile MAE is how far past your short strike it typically
travels before reverting. That's the wing.
```text Your prompt theme={null}
Propose an optimization over the hold and the stop multiple
```
Three holds are three expiries of the same trade. Pick the plateau.
A futures backtest isolates the **directional and realized-path** component
of a premium trade. It has no theta and no vega — read it as "what did the
underlying do", never as the P\&L of the options position itself.
## Next steps
Every series you can gate on — ATR, NATR, z-scores, percentiles, VWAP, ranges.
The metrics card, the trade list, MAE and MFE.
Propose, run and apply a parameter sweep.
The full roster, including `VX`.