Cheap SPX Same-Day Options Are the Worst Ones to Buy

We priced the same-day SPX straddle at the open on every session since April 2023. That is the at-the-money call plus the at-the-money put, both expiring that afternoon, divided by where SPX opened. Each leg is taken at its first trade of the session, which lands at or just after the bell. On the sessions where we can check it against the 9:30 minute bar, the two agree to within 2.5% of the straddle price and pick the same strike every time.

We then repriced it every hour of the day. That second pass covers 88 sessions rather than 847, because hourly option prices only reach back about four months while daily ones go back years.

Two things came out of it. The straddle costs about half of what VIX says it should, at every level of VIX. And when it looks cheap against its own recent history, it is the worst thing to buy, not the best.

What it costs

The cost rises with VIX, which is no surprise. What matters is the last column. It compares what you actually pay against what VIX implies for a single day.

The gap is not an error in the pricing. VIX is built to describe a full calendar day, and it carries the overnight with it. A straddle bought at 9:30 and held to the close only has to cover the hours the market is open. On top of that, 30-day volatility sits above same-day volatility when things are calm, so the shorter contract prices lower.

The steadiness is the interesting part. From VIX 13 through 20 the ratio barely moves. It sits between 0.53x and 0.57x across more than 650 sessions.

Then it widens. Above VIX 20 it goes to 0.65x, and above 30 it reaches 0.73x. Same-day options get relatively more expensive than VIX suggests exactly when people most want to own them.

The cost falls all day

Every hour that passes takes value out of a contract that expires at 4pm. By the last hour the straddle costs a quarter of what it did at the open. This matters for anyone trying to judge whether it is cheap, because a reading at 3:30pm cannot be compared to one at 10am.

The last column is how far SPX still travelled between that hour and the close. Put the two together and the cost sits above the move at every hour of the day.

Cheap is the worst thing to buy

The next question is whether the price carries information. We ranked every hourly reading against the same hour on the previous 20 sessions, so a 10:30 reading is only compared to other 10:30 readings. Nothing from the future is used in the ranking.

The result runs opposite to the obvious guess.

A cheap straddle returned 0.70 of what it cost and paid off on 33.3% of readings. An expensive one returned 0.98 and paid off on 48.1%.

The reason is that volatility clusters. A reading is cheap because the session has been quiet, and quiet sessions tend to stay quiet through to the close. A reading is expensive because something is moving, and that keeps going too. Buying the cheap one is paying a small price for an even smaller move.

This is not simply a VIX effect. Dividing each reading by VIX before ranking it, which strips out the volatility level of the week, leaves the same shape. Cheap readings won 37.3% and expensive ones 46.5%.

Only one hour holds up

Splitting the same test by hour shows the effect is not spread evenly. It lives in the early afternoon, and at the open it points the other way.

At 1:30pm an expensive straddle paid off on 60.0% of sessions and a cheap one on 16.7%. Seven hours were tested, and testing seven things means one of them clearing an ordinary bar is fairly likely on luck alone. Raising the bar to account for that leaves 1:30pm still standing. Every other hour is noise.

The honest caveat is the reversal at 9:30, where the cheap third won more often. It is not a significant result on its own, but a pattern that flips at one end of the day is a weaker pattern than one that holds throughout.

What this does not say

None of this points at direction. The typical signed move to the close was between -0.014% and -0.042% across cheap, middle and expensive readings, and each group split close to evenly between up and down sessions. The reading describes size, not which way.

It also does not make the buyer whole. Across all 847 daily sessions the straddle paid more than it cost on 42.4% of them. The typical outcome to a buyer was -0.063% of SPX, while the average was +0.004%. That difference is April 2025 doing the work. On April 9 the straddle cost 3.25% and SPX moved 9.90%. A handful of days like that pay for a great many small losses, which is the whole shape of owning options.

Two limits on the hourly work are worth naming. It rests on 88 sessions, because that is as far back as hourly option prices reach. And the seven readings inside one session move together, so the effective sample is closer to the number of days than the number of readings.