Three Contracts That Look Good on Screen and What the Numbers Actually Say

Open any options chain, sort by premium, and the top of the list looks like free money. That sort is the most expensive habit in put selling, because premium does not measure opportunity. It measures what the market thinks is about to happen, and the market is usually right.

The question worth asking is never which contract pays the most. It is which contract pays more than the risk it carries. Those are different questions and no column in the chain answers the second one.

What follows is three cases where a contract reads well on screen and badly under examination. None of them is exotic. All three turn up near the top of a premium sort on any given morning.

The open interest ghost town

Open interest counts the contracts at a strike and expiry that are currently held and not yet closed. It is a stock figure rather than a flow figure, so it tells you how many positions exist, not how many changed hands today. Volume tells you the second thing. You want both.

A strike carrying a handful of contracts and no volume is not a market. It is a quoted price with nobody standing behind it.

The usual warning about thin strikes is that you will struggle to get out. That warning does not apply here. Every contract we score runs to expiry, and the outcome is settled by where the underlying closes against the strike. If you never intend to trade out of a position, why should the size of the crowd matter?

Because thin open interest changes what the quoted price means.

A liquid strike has been priced by many participants arguing with each other on both sides. The mid is what is left when that argument settles. A strike with almost no interest has been priced by one market maker's model and corrected by nobody. Maybe the number is fair. Maybe it is stale, or wide by default, or bent by a single resting order from somebody with reasons of their own. You cannot tell which from the chain, and there is no crowd to check it against.

There is a nastier version of the problem. Implied volatility is not observed, it is reverse-engineered from the option's price. So the IV at that strike comes from the same unexamined quote, and anything you build on top of the IV inherits the flaw without showing it. A probability estimate, an expected move, a safety buffer. Each step renders the number more confidently than the last while the original uncertainty sits underneath, now invisible.

One check costs nothing. Look at the implied volatility of the strikes either side of the one you are considering. Across strikes at a single expiry, IV follows a curve, and it is smooth because participants price those strikes against each other. A strike whose IV jumps sharply off that curve for no economic reason is usually telling you about its own illiquidity rather than about the stock. The curve is the crowd's opinion. The outlier is one model's.

The wide spread

The bid-ask spread is the gap between what buyers are offering and what sellers are asking. On a heavily traded contract that might be a couple of cents. On a neglected one it can be a third of the contract's value.

Here is the part that gets skipped. Selling a put, you are almost never filled at the mid. You are filled somewhere between the mid and the bid, and on a wide market that gap is real money handed over at the moment of entry.

Work an example. A contract quoted 1.00 bid, 1.40 ask has a mid of 1.20. Sell at 1.05, which is an ordinary fill on a market that wide, and you have collected 87% of what the screen said the thing was worth. The rest is gone before the trade has done anything at all.

Put that against the return the trade was supposed to produce. Against 4,500 in collateral over 30 days, 120 in premium annualises to roughly 32%. The same position at 105 annualises to roughly 28%. Four points of annualised return, lost at entry, entirely invisible in the chain you screened from. If your process ranked that contract above an alternative on a mid-price assumption, the ranking may simply be wrong.

This matters more for put selling than for most things, because the premium is the whole return. There is no dividend, no appreciation, no second payment later. The best possible outcome is that the contract expires worthless and you keep exactly what you collected on day one. Whatever the spread takes at entry comes straight off the maximum the position can ever produce.

Judge spreads as a percentage of the mid rather than in cents. Five cents on a contract worth 4.00 is about 1.25%. Five cents on a contract worth 0.35 is over 14%. Identical cent figure, completely different trade. The same logic applies to the underlying, which is why our execution scoring scales spread tolerance to liquidity tier rather than applying one number to everything.

Selling into earnings without knowing it

Earnings are the cleanest case of premium being expensive for a reason.

In the days before a scheduled announcement, implied volatility rises, sometimes dramatically. The options market is pricing a wider range of outcomes because something genuinely unpredictable is about to resolve. Sell a put into that and the premium is fat. You have also taken on overnight gap risk, and distance from the strike does considerably less for you against a discontinuity than it does against ordinary drift. The stock does not travel through the intervening levels. It simply appears somewhere else at the open.

This is not automatically a bad trade. It is a different trade. Selling volatility into an event and harvesting steady premium in quiet conditions are two strategies with two risk profiles, and the damage comes from running one while believing you are running the other.

Which is why our two engines invert the same rule. Patrol requires a contract to expire before any scheduled earnings date, so no position is ever exposed to an overnight repricing. Strike does the opposite, requiring an imminent announcement and rich implied volatility, holding through the release to capture the volatility collapse that follows. Same scoring core, opposite earnings condition. One rule flipped is the entire reason there are two products instead of one.

The failure case is neither. It is selling a 30-day put on a company that reports in nine days without having checked, collecting premium priced for an event you knew nothing about, then calling the result bad luck when the stock gaps through your strike. The position was never mispriced. The market told you exactly what it was worried about. You did not read that part.

Checking is harder than it sounds. Most options data providers, including the ones supplying institutional-grade chains and volatility surfaces, do not publish exact forward-dated earnings in their APIs. What circulates instead is estimates inferred from last year's timing, and those are frequently wrong by several days. Several days is the entire margin that decides whether a contract expires before or after the event. We carry a separate vendor for that one field, which is a permanent exception to an otherwise consolidated data stack, and it exists because an estimated earnings date is not good enough to gate on.

The shape underneath all three

Every one of these has the same structure. A number that is easy to see stands in for a number that is hard to see, and the swap goes unnoticed because the easy number is right there in the chain and the hard one is not.

Screening well means refusing the swap. Is this price real. Will I actually receive it. Is there something known and scheduled inside the contract's life that explains the generosity. Three questions per contract, across several thousand names with dozens of strikes and expiries each, before the open.

That is the arithmetic this platform does. Not to tell anyone what to trade, but to make the premium column mean something.