Start with the event, not the chart
An earnings preview is only an earnings preview if the date is real. Verify the announcement from the company's investor-relations page or the SEC filing record, not from a calendar aggregator alone. Two details matter beyond the day: whether the release is before the market opens or after it closes, and which timezone the company states.
The session decides which close belongs to the reaction. For a before-market report, the reaction session is the announcement day; the prior close is the reference for a position exited before the news. For an after-market report, the reaction session is the next trading day, and the announcement-day close is the pre-event reference. If the session is unknown, the honest treatment is to qualify the comparison rather than assume one.
Also distinguish the results release from the conference call. A company may publish numbers before the open and hold the call later. A trade that intends to exit before the call is not the same as one that intends to hold through it.
Finally, separate a confirmed date from an estimated one. A date derived from last year's pattern is a planning estimate. It is reasonable to research around it; it is not reasonable to label it confirmed or to build an event article that presents the estimate as settled.
Check the quote before you check the thesis
Every option number needs a timestamp and a session. A daily preparation quote from yesterday's close is useful evidence, but it is not the price available now. A weekend does not create a new quote. A holiday does not either.
Ask three questions of any number on a research page:
- As of when? The exact session or timestamp, not "recently."
- From which provider? Historical and current implied volatility can come from different vendors and need not be methodologically identical.
- Is it eligible? A quote can be present and still fail a liquidity or quality gate. A wide market, a stale mark or a missing strike should be shown as unavailable, not silently replaced.
The same rule applies to the stock price used in an implied-move calculation. If the spot is from a prior close, say so. If the option quote is from a different moment than the spot, the resulting move is an approximation, not a live measurement.
This is not pedantry. An earnings decision made on a stale quote is a decision about a market that no longer exists. The error is usually larger than the difference between two reasonable strategy views.
Make the expiration contain the event you mean
An option only participates in an earnings move if it is still alive when the news arrives. Check the expiration date against the announcement date and session, including the possibility that the report moves.
For a before-market report, a contract expiring that same day still contains the announcement, but its remaining time after the release is only the trading session. For an after-market report, a contract expiring on the announcement day does not contain the reaction, because the market has already closed. The first expiry that contains the reaction is later.
This is the most common silent mismatch in earnings options. The trader believes the position is an event trade; the contract has already stopped being one.
A related mistake is assuming every expiry prices the event equally. The first expiry after the announcement is usually the most event-sensitive. Later expiries contain the event plus ordinary time, so the same implied volatility change has a smaller proportional effect. Neither is automatically better. They answer different questions.
Read liquidity as a cost, not a footnote
The bid-ask spread is a real cost and it is usually widest exactly when event interest is highest. A wide market also makes the displayed mid an unreliable mark: the midpoint of a $0.20 bid and a $1.40 ask is not a price anyone can trade.
Before using a quote, check the spread width relative to the premium, the size available, and whether the strikes you need are quoted at all. If the chain is thin, the research can still show historical evidence, but a current executable candidate should be blocked or explicitly qualified. A blog article must not turn a historical leg into a suggested live trade.
Costs compound with the strategy. A long option pays the spread once; a multi-leg structure pays it on every leg, and the mid-price payoff diagram can look materially better than the executable result. When comparing structures, compare the cost-adjusted outcomes and state the fill assumption. A midpoint simulation is not an executed trade.
If the spread is wide enough that the thesis needs an unusually large move to overcome it, that is information about the trade, not an inconvenience to omit.
Ask what the historical sample actually contains
A statement like "the stock moved more than the implied move in seven of the last eight quarters" sounds precise. It is only meaningful with the cohort, the matching rule and the treatment of missing data attached.
Check the period covered, the number of distinct earnings events, how many observations were valid, and how many were rejected. Several entry days from one quarter are not several independent events. A quarter with a missing quote is not a zero outcome. A comparison that silently changes the entry offset is not matched.
Also check the metric. Historical win frequency after costs is not a forecast probability. An average return without the median, the downside and the worst observations hides the shape of the distribution. A maximum-loss return on an undefined-risk strategy is not a return on capital in any comparable sense.
If the sample is small, the honest presentation is descriptive: here are the observations, here is the count, here is the dispersion. That is more useful than a significance claim the data cannot support.
Write the exit before the entry
The exit is part of the trade, not an afterthought. Decide in advance whether the position will be closed before the announcement, into the reaction, or held to expiration. Each choice has a different exposure.
- Before the announcement: the position is a bet on the pre-event volatility and drift, not on the result. It normally exits at the prior session close for a before-market report, or the announcement-day close for an after-market report.
- Into the reaction: the position carries the event and the volatility repricing. The relevant close follows the session convention above.
- To expiration: the position carries the event plus the remaining time. The payoff is the expiration payoff, not the post-event mark.
For multi-leg structures, the exit also has to state which legs are closed and when. A calendar spread has two expirations and cannot be described by a single payoff diagram at one date. An early assignment or an exercise changes the position, sometimes before the intended exit.
A written exit also protects against the most expensive habit in event trading: reclassifying a losing event trade as a long-term investment after the report.
A five-minute pre-trade checklist
- Date and session confirmed from a primary source, with the timezone and the source link.
- Fiscal period verified from the company's own reporting, not inferred from the calendar month.
- Quote cutoff visible and consistent between spot and options; stale or missing data labelled as such.
- Expiry contains the event under the actual session convention, not the assumed one.
- Liquidity checked on every leg, with the spread treated as a cost.
- Historical cohort stated with counts, matching rules and missing observations.
- Capital basis named — premium at risk, defined maximum loss, or secured cash are not interchangeable.
- Exit written down with the price reference and the session it belongs to.
If any item cannot be answered, the position is not ready. That is a useful outcome: the checklist is designed to make the missing evidence visible before it becomes a loss.
For live company evidence, start from the earnings research pages. The expected-move calculator shows how the straddle estimate is computed, and the implied versus actual move guide explains how to compare the estimate with observed outcomes without treating the comparison as a forecast.