Every claim we have tested
62 of them. Folklore people repeat, reports this industry sells, and the questions we had to answer about our own method. Each one links to the page that owns it, and each one carries the two things that decide whether a percentage means anything: how much data it stands on and what it was measured against.
What a report here has that a report elsewhere does not
Nothing in this list is a chart with a number under it. Every comparison is run against a control — a matched move, a permuted grid, a direction-randomised day, an era the selection never saw — because the control is usually where the finding dies. "Weekend gaps fill 88% of the time" is true; a size-matched midweek move fills 88% too, and that second number is the whole report.
13claims that hold up
26that do not survive their control
23where only part of it holds
Should You Avoid Trading on Monday?
“Never trade Monday — the market is still waking up from the weekend.”
Mostly false — and backwards for indices.
19 markets · 344,411 hourly bars
Is Early Monday Morning Too Thin to Trade?
“Early Monday specifically is thin and choppy before London opens.”
True — on movement.
19 markets · 100,568 hourly bars
Is Friday Afternoon Too Choppy to Trade?
“Avoid Friday afternoon — position squaring makes it choppy and risky.”
True, but mis-explained.
19 markets · 92,758 hourly bars
Is the Asian Session Worth Trading?
“The Asian session is not worth trading — low liquidity, no movement.”
Half true — and it depends entirely on your pair.
19 markets · 576,962 hourly bars
Are Tuesday to Thursday the Best Days to Trade?
“Tuesday to Thursday are the most active, trending days.”
False.
19 markets · 1,037,802 hourly bars
Is the Santa Claus Rally Real?
“December rallies — the Santa Claus effect lifts markets into year end.”
Not answerable for stocks. Real, and in the wrong asset, for metals.
19 markets · 135,267 hourly bars
Is August Really Dead in Forex?
“August is dead — desks on holiday, thin liquidity, nothing moves.”
Half right — and quieter is not the same as safer.
19 markets · 151,231 hourly bars
Is the First Hour After the London Open a Fakeout?
“Avoid the first hour after the London open — it is a violent fakeout.”
False — with one real exception.
19 markets · 146,477 hourly bars
Is Tuesday 120–130% More Volatile Than Monday?
“Tuesday volatility runs 120-130% of Monday.”
False — and it is the rare claim stated with a number.
19 markets · 346,410 hourly bars
Is Friday Really Worse Than Monday?
“Monday and Friday are the two worst days, and Friday is worse of the two.”
Mostly supported — the closest any folklore came to surviving whole.
19 markets · 310,202 hourly bars
Should You Avoid the Wednesday Triple-Swap Rollover?
“Wednesday carries triple swap, so its rollover is the one to avoid.”
Unproven — held loosely on purpose.
18 markets · 25,647 hourly bars
Should You Wait for Tokyo Instead of Trading the Sunday Open?
“The Sunday open gaps and spreads are wide - wait for Tokyo.”
True — and it is the best example of why a big number can be untradeable.
18 markets · 49,609 hourly bars
Does "Sell in May and Go Away" Work?
“Sell in May and go away - Nov-Apr beats May-Oct.”
We cannot tell you — and that is the honest answer.
19 markets · 864,971 hourly bars
Should You Stop Trading From June to August?
“Cut trading frequency from June to August - the whole summer is dead.”
Quieter, yes. Biased, no.
19 markets · 444,809 hourly bars
Is the London Kill Zone the Most Important Window?
“ICT London kill zone (02:00-05:00 NY) sets the directional move for the day.”
No — it is an ordinary hour.
19 markets · 219,541 hourly bars
Is the New York Kill Zone Worth Trading?
“ICT New York kill zone (07:00-10:00 NY) - the second most important window.”
Yes — and it outranks London, which is the reverse of what is taught.
19 markets · 219,646 hourly bars
Does the Asian Kill Zone Just Range?
“ICT Asian kill zone (20:00-22:00 NY) ranges rather than trends.”
No — it is the most directional of the four.
19 markets · 143,660 hourly bars
What Actually Drives the ICT London Close Kill Zone?
“ICT London close kill zone (10:00-12:00 NY) is a tradeable reversal window.”
Real, but named after the wrong end of itself.
19 markets · 146,394 hourly bars
Are ICT Kill Zones Real?
“The kill zones are the only hours worth trading.”
Yes — the framework holds in all 19 markets.
19 markets · 729,241 hourly bars
Does the 4pm London Fix Move the Market?
“The 4pm London fix hour behaves differently - the benchmark banks were fined for rigging.”
Yes — and it is measurably calmer since the 2015 reform.
19 markets · 73,180 hourly bars
Should You Avoid the Daily Rollover?
“The 21:00–22:00 UTC rollover is thin liquidity to be avoided.”
True — and it produces the biggest numbers you should least trust.
19 markets · 108,799 hourly bars
Is the grid data-mined?
185 significant buckets against 5.9 from a permutation null that keeps each market's own structure.
Section 1 of the evidence page
Has the edge already decayed?
70 significant buckets in the early era against 79 in the late one, at matched power.
Section 2 of the evidence page
Does anything we found have a reason to exist?
The strongest windows sit where liquidity is thinnest, which is a reason to distrust them rather than a mechanism.
Section 3 of the evidence page
Does a hot streak mean anything?
A hot streak is as predictive at four weeks' lag as at one, so it is drift rather than memory — worth 0.9 points.
Section 4 of the evidence page
Does the weekend gap fill?
Weekend gaps fill 88% of the time within a day. A size-matched midweek move fills 88% too.
Section 5 of the evidence page
Does yesterday's high or low matter?
A break of yesterday's extreme holds 50.4% of the time against 51.0% for a matched arbitrary level.
Section 6 of the evidence page
Does the opening range set the day?
The opening range is broken on essentially every session, so it selects nothing at all.
Section 7 of the evidence page
Chosen on one era, scored on another
Windows picked before 2020 and scored after it are worth +0.56 points in ordinary hours — real, and less than a round trip.
Section 8 of the evidence page
The largest surviving effect is one tick
Every market that trades the whole clock closes down at the New York close, by about one tick. It reads as settlement, not direction.
Section 9 of the evidence page
The axis we do not have
The trading day of the month — this category's signature axis — produces 0 of 285 slots beyond chance.
Section 10 of the evidence page
Why n ≥ 300 and |z| ≥ 3?
Every defensible threshold setting was run, not just ours: 22 of 25 beat noise by five times or more.
Section 11 of the evidence page
What cost does the edge die at?
The median alertable rule clears its assumed spread by 1.29×, and 9 of 12 stop paying at 1.5× the assumption.
Section 12 of the evidence page
Does the size axis survive out of sample?
The timetable picked before 2020 still ranks the hours after it at 0.925 — and 0.859 inside the busiest four hours.
Section 13 of the evidence page
Is the timetable the same in August as in November?
Which hours are biggest holds all year. How big they all are moves about 5.7% in August, and reverses in 3 markets.
Section 14 of the evidence page
Is UTC the right clock?
The whole grid rebuilt on four clocks: no wall clock finds more than UTC, and Sydney is the control.
Section 15 of the evidence page
Should you only take a window when the market is busy?
A flagged window is LESS reliable after a big hour, not more — and neither state clears its round trip.
Section 16 of the evidence page
Are the weeks inside a window independent? Our own maths assumes so
The weeks inside a window are not quite independent, so our |z| ≥ 3 bar is really about |z| ≥ 3.06.
Section 17 of the evidence page
Crypto trades all week, and we have been measuring it with a five-day instrument
Crypto's weekend moves 0.85× a weekday hour — an ordinary trading period whose timetable does not carry across it.
Section 18 of the evidence page
When does the day's high and low actually form?
Move the day boundary and not one of 19 markets keeps its answer about where the day's high forms.
Section 19 of the evidence page
Does the first hour set the day?
A green first hour is worth -0.12 points over a control that re-pairs it with another day, and the initial balance is broken on 99.7% of sessions.
Section 20 of the evidence page
Are nineteen markets nineteen experiments?
Correlated markets inflate the null behind section 1 by only 1.11×, so our 19 markets are worth 17 independent ones — the headline holds, the agreement count does not.
Section 21 of the evidence page
Do round numbers act as support and resistance?
Hours that break a round number finish PAST it, not short of it — the opposite of the folklore — and the edge clears a round trip in 0 of 19 markets.
Section 22 of the evidence page
Does a narrow range precede an expansion?
The day after the narrowest day in seven is 1.53× that day and 0.944× an ORDINARY day — the expansion is regression from a selected minimum, and the tightest coils are followed by the smallest days.
Section 23 of the evidence page
Does a candlestick pattern add anything?
An engulfing candle is followed by its own direction 48.43% of the time — against 48.61% for following any day at all. Requiring the shape on top of the sequence is worth -0.48 points, and no market clears the bar.
Section 24 of the evidence page
Do Fibonacci retracement levels do anything?
Not one of 5 Fibonacci depths behaves differently from an undrawn ratio inside the same legs, and the version of the test a vendor would run flags a ratio nobody draws 29% of the time.
Section 25 of the evidence page
Does London break the Asian range and run?
London breaks the Asian range on 93.9% of sessions, so it selects nothing — and given a break, being the Asian high rather than a borrowed day's level is worth -1.03 points, positive in 1 of 12 markets.
Section 26 of the evidence page
Is "up" an artefact of where we cut the bar?
Redefine "up" as a close above the previous close and 121 of 167 flagged windows survive — against 144.3 when the same gaps are dealt onto other bars, so 23.3 of our flags turn on where in the week the gaps land.
Section 27 of the evidence page
What does one of our own flags actually read?
Of 185 windows this site flags, 150 read between 55% and 65% in the direction they lean — the range we advertise — but every one of the 10 that read higher sits in the thin band, and the strongest ordinary hour reads 59.43%.
Section 28 of the evidence page
Does the golden cross pick a better moment than a coin?
The golden cross beats its own random-date control in 1 of 16 markets where noise produces 0.8, and 50/200 ranks 89 of 108 pairs on its own grid. The 200-day trend filter is worth 1.25 points of up-rate — and price simply above its close 200 days ago, with no average at all, is worth 1.27.
Section 29 of the evidence page
Do weekend gaps fill?
"Gaps fill 90% of the time" is true and empty — a matched midweek move fills just as often. The indices do the opposite.
88% vs 88% — gaps filled within a day, against the control
Does yesterday's high or low matter?
68,688 breaks across 19 markets. The level holds 50.4% of the time — slightly less than an arbitrary level the same distance away.
50.4% vs 51% — breaks that held, against the control
Does the opening range set the day?
The range breaks on 99.8–100.0% of days. A signal that fires every session has not selected anything.
100.0% — of sessions break the range — it selects nothing
When does the day's high and low actually form?
"The high forms at the open" is the boundary you chose, not the market. Move it and every single market changes its answer.
0 of 19 — markets keep their answer when the day boundary moves
Does the first hour set the day?
A green first hour tells you the first hour was green. Not one of 17 markets reaches even a weak significance bar.
-0.12 pts — is what a green first hour is worth, against its control
The best time to trade forex
The overlap is real — as a movement and cost argument, not a direction one. The half that survives a holdout is WHEN, not which way.
2.41× — is what the overlap moves against a typical hour
Does Bitcoin move on the weekend?
The weekend is an ordinary trading period whose timetable is a different timetable — the biggest hour moves from 14:00 to 00:00 UTC.
0.85× — is what a weekend hour moves, against a weekday one in BTC
Why does every currency pair fall at the New York close?
The biggest number this archive contains, and it is worth about one tick: 12 of 12 markets fall, and 11 of them by less than the spread.
12 of 12 — markets close down in that hour — and it is worth about one tick
Does the trading day of the month matter?
The chart this whole category is built on. 285 slots tested, 0 clear the bar, and the strongest anywhere reaches z = 2.99.
0 of 285 — day-of-month slots clear our bar — the axis the category is built on
Should you only trade when the market is busy?
The most repeated filter in this business, and it points the wrong way: 14 of 14 markets are LESS reliable after a big hour.
-2.03 pts — is what waiting for a busy market does to a flagged window
What spread kills a seasonal edge?
The hour's MOVE beats a round trip in 471 of 471 liquid windows. The directional EDGE beats it in 4.
1.29× — is all the median alerted rule clears its assumed cost by
Does a hot streak mean anything?
A streak is as predictive at four weeks' lag as at one. Memory decays; drift does not — which is why there is no streak feature on this site.
0.91 pts — is what a hot streak is worth, and it does not fade with time
Why the ones that failed are still here
Most of this list is negative. The gap does not fill more than an ordinary move, yesterday's high does not hold more than an arbitrary level, the opening range fires every single day, the trading day of the month produces nothing at all, and holding a flagged window longer loses at every horizon we tried. Those were all things we wanted to find, and publishing them is the only reason to believe the handful that survived. A catalogue that only ever confirms its author is not evidence of anything except the author.