Methodology
Thirteen gates, and why each one is there
A threshold without a reason is superstition. Every gate below is paired with the thing it rejects and where the number came from. Any candidate failing any single gate is not a trade.
Direction must agree with trend
Checked before anything else, because it is the cheapest filter and the most violated.
- A call requires price above a rising 50 day simple moving average, with 14 day RSI between 45 and 70.
- A put requires price below the 50 day.
This exists because of two specific errors. A call was once surfaced on a name trading at $210.96 against a 50 day of $212.91 with RSI 43.9, on the argument that a supertrend reading had flipped positive. Supertrend direction alone is not trend confirmation. Separately, a put was held on a stock sitting 5.6% above its rising 50 day with RSI 56, which is simply a bet against something that is going up.
Both lose for the same reason: the option was asked to overcome the underlying’s own drift before it could make money.
Contract gates
| Gate | Requirement | What it rejects |
|---|---|---|
| Delta | 0.50 to 0.65 | Cheap far strikes that need an 8% to 24% move. See strike selection. |
| Break even | within 3% of spot | Contracts whose arithmetic requires an outlier, not a normal move. |
| Spread | 5% of mark or less | The only cost that is not compensated by anything. See spread is the edge. |
| Implied volatility | under 40%, over 70 rejected outright | Paying a volatility premium that decays whether or not the thesis works. |
| Theta | under 5% of mark per day | Positions where the clock beats the thesis. Expressed as a percentage of premium, never in dollars, because dollars are meaningless across different contract prices. |
| Open interest | 1,000 or more | Chains nobody trades. |
| Volume | 500 or more today | Stale open interest from a crowded strike that has gone quiet. |
| Bid size and ask size | 20 or more, each | The trap described in liquidity traps. This one is invisible in both spread and open interest. |
| Days to expiry | 21 to 60, third Friday monthly preferred | The expiry week theta cliff, and the thin spreads on weeklies. |
| Earnings | must fall after expiry | Holding an unreleased print. See event volatility. |
| Recent high | no 52 week high in the last three sessions | Entering into a move that has already happened. |
| Break even ceiling | must sit below the 52 week high | Trades that need a new all time high just to return the premium. |
| Size | 5% to 8% of capital | Position sizes that make one trade decide the year. |
Pacing
Separate from whether any individual trade is good.
- At most two new positions per week, at most one per day
- At most six open contracts at once
- Exit at +50% on the contract, or five days, whichever comes first
Pacing limits exist because the failure mode is not holding a loser. It is opening a seventh position because the first six are uncomfortable. The weekly cap is a hard gate in the automation, not a guideline: when it is reached, the screen reports no ideas rather than producing a list that cannot be acted on.
Most runs report no candidates. Thirteen conjunctive gates on a universe of roughly 500 names, intersected with a requirement for a specific dated catalyst, will usually produce nothing. A screen that returns something every day has gates that do not bind.
Why single leg only
Maximum loss equals premium paid, always, with no assignment risk and no margin interaction. It also keeps the thing being tested small enough to attribute: when a trade loses, the cause is the underlying move, the volatility, the spread, or the clock. Four things, all measurable. A multi leg structure would obscure which one it was, and the point of the log is to find out.