Measured, not guessed · 2020..2026 of H1 data

When does SPX500 actually rise?

Every hour below is a real measurement: how often SPX500 closed higher than it opened in that hour of the week, across 39,343 hourly bars. Not signals — base rates, with the sample size on every figure.

Read this one differently. Across SPX500's whole grid, the number of windows clearing our significance bar is 0 — against 0.31 that a grid with no hour-of-week structure at all produces at the same settings. In other words we cannot demonstrate that this market keeps an hourly schedule. Everything below is measured honestly and is still worth reading as description, but it is not evidence of a pattern, and with 7 years here we could only have detected a lean of about 8.1 points — larger than the effects we find in the markets that do show structure. How this was tested →
52.2%closed higher overall
39,343hourly bars measured
5yof history
0hour-buckets beyond chance (|z|≥3)
3.0×its busiest hour against a typical one

When does SPX500 actually move?

We cannot demonstrate which way SPX500 leans in any hour — that is the caveat above, and it is honest. How far it moves is a different question, and this one it answers.

Window (UTC)Typical movevs its own average hourRound trip
Friday 14:0011 pts 3.0×0.7 pts
Thursday 14:009.5 pts 2.7×0.7 pts
Tuesday 14:009.3 pts 2.7×0.7 pts
Thursday 15:008.5 pts 2.4×0.7 pts
Wednesday 14:008.1 pts 2.3×0.7 pts

A typical hour in SPX500 ends up 3.5 pts from where it opened — and covers rather more than that on the way, since a move out and part-way back still finishes near its start. Every figure in this section is that net distance; its quietest (Monday 04:00) moves 1.7. The spread on a round trip is about 0.7, so the gap between the busiest hour and the quietest one is the difference between a move worth paying for and one that is mostly cost.

Does the London–New York overlap actually deliver?

Ask when to trade anything and you get the same answer everywhere: the London–New York overlap, 12:00–16:00 UTC. Those pages do cite numbers — the overlap is roughly 19% of the day and carries over a third of daily volume, and spreads there are genuinely the tightest. Every one of those is a figure about how busy the hour is. None is a figure about which way it goes. So here is that number for SPX500, from 39,343 hourly bars:

WindowClosed higherSampleAvg move
London–NY overlap (12:00–16:00 UTC)52.7%n=6,878+0.20 pts
Every other hour52.1%n=32,465+0.07 pts

The gap is 0.6 percentage points — inside what chance produces at these sample sizes. On direction alone, the overlap is not a better bet for SPX500 than the rest of the day. That does not make the advice useless: the overlap is where liquidity and spreads are best, which matters for cost. It just is not an edge in which way price closes.

SessionClosed higherSample
Sydney51.6%n=13,742
Tokyo52.1%n=15,486
London52.7%n=15,472
New York52.5%n=15,287

Sessions overlap by design, so these windows share bars.

The strongest and weakest hours

Averaged across the trading week, the hour that has leaned up hardest is 18:00 UTC — SPX500 closed higher 54.3% of 1,671 times there, though that lean is thinner than it looks. The weakest is 21:00 UTC, closing higher only 46.3% of 551 times.

Hour × weekday, the whole picture

Colour is the up-rate — red leaned down, teal leaned up. The brass dot marks buckets whose lean is statistically beyond chance at |z|≥3. Most cells have no dot. That is the honest shape of the market: mostly noise, with a few stubborn corners.

0003060912151821MonTueWedThuFriSun

No single weekday-hour bucket of SPX500 clears the |z|≥3 bar. That is a finding, not a failure — most markets, most of the time, are indistinguishable from chance at this resolution.

Days and seasons

WeekdayUp-rateSample
Monday54.0%n=7,766
Tuesday50.7%n=7,976
Wednesday52.1%n=7,931
Thursday52.3%n=7,900
Friday52.0%n=7,197
Sunday51.5%n=573

Monday has been the friendliest day (54.0% of 7,766), Tuesday the least (50.7%). By calendar month, November leads at 52.9% and December trails at 50.5% — month buckets carry roughly a twelfth of the sample each, so hold them more loosely than the hourly figures.

Did any of it survive a real backtest?

Colours are description. A different question is whether a simple rule built on them — enter at the open of a bucket, exit at its close, pay a realistic spread — would have made money, and kept making it after a realistic spread is charged. We ran that test across 4,248 rules on the whole board: 200 finished net positive — 4.7%.

Read "net positive" as a filter that was applied, not as evidence that survived. Those rules were chosen on the first 70% of the history and charged a spread on that same era, so the count above is the selection restated. Until September 2026 they were chosen on the whole history and then scored on its last third, which is not a holdout at all; we rebuilt the search and published the correction rather than quietly fixing it. The last 30% now selects nothing. The longest-standing clean holdout here is section 8 of the evidence page — windows picked on pre-2020 candles only, scored on 2020 onward — and it is worth +0.54 points in ordinary hours, which is less than a round trip.

No published rule runs on SPX500 once spread costs are charged. The strategy board shows that too — a leaderboard that only shows winners is lying by omission.

Watch it live

These are the same numbers, updating as each hour begins — free to embed on your own site from the widget builder:

Read it without fooling yourself

The guide walks through all four traps with live numbers.

Where SPX500 sits

Two things worth knowing before you treat any of the above as SPX500-specific.

These markets keep almost the same clock

Correlating the shape of the trading day — each hour's distance from that market's own average hour — against every other market we hold:

Tested claims that land differently here

Of the 21 pieces of timing folklore we tested across all markets, these are the ones where SPX500 itself moves beyond chance:

What we have tested that covers SPX500

Each of these is a claim this industry sells, measured against a control. ⚠ Only the ones whose measurement actually included SPX500 are listed — several of our tests exclude markets outright, and offering one here that skipped this symbol would be the opposite of the point.

All three blocks are computed from the same files this page is, so they cannot drift from it. The full picture — every market against every hour, with the noise count attached — is on the screener, and the tests behind the significance bar are on /evidence.

Open SPX500's strongest hour in the dashboard →

Explore SPX500 yourself — every hour, day and month, filterable, with the sample size on every figure. The free account covers the full 5-year history.

Open the free dashboard →