What an hour of the day is worthWhat an hour of the day is worthAn earlier test measured the hour each of 666 markets was dearest, found the distribution flat, and read it as noise. Flat is equally what a real cycle in every market at its own hour looks like — which was the owner's original suggestion — so that test could not tell the two apart. This one splits each market's league in half, picks the hours on the first half and takes the return on the second. The hours repeat at 10.7% against a 1.6% chance rate, so the profiles are not noise. What they are worth is another matter, and it splits sharply by what a market is quoted in.
Hours repeated
10.7%
of markets, against 1.6% by chance
Chaos-quoted
+1.74%
median out of sample, 70% profitable
Divine-quoted
0.00%
median out of sample, 46% profitable
Roughly a hundred chaos-quoted markets carry this and pay about 1.7% on a median day, seventy percent of the time. Divine-quoted markets have profiles just as stable and no amplitude worth taking: zero out of sample, profitable less often than a coin. The top of the list reaches twenty percent, which is where the work is.
349 markets deep enough to profile
| Market | Buy → sellBuy → sellHours of the UTC day, and they stay UTC. An hour-of-day bucket is a property of the data rather than a moment: converting it to your clock would make one row mean different things to two readers, and neither would match the analysis it came from. Everything on this site that is a *moment* — when a digest arrived, when a signal fired — is on your clock instead. | Out of sampleOut of sampleHours chosen on the first half of the league, return measured on the second with those hours fixed. This is the number to believe. The column beside it is the same figure over the whole window, which was chosen on the data that measures it and runs consistently higher for that reason. | Whole window | Hours heldHours heldThe second half put the cheapest and dearest hours within one of where the first half did. Chance is about 1.6% for both ends; the population manages 10.7%. A market that did not hold may still show a return, and that return is the one to distrust. | CauseCauseChaos has a daily cycle of its own — measured separately, strongest at 14-15h UTC and weakest at 22-23h — so every market chaos denominates inherits it. About a hundred rows here are that one finding seen a hundred times. They can all be worked; they cannot all be counted as separate evidence, and sizing them as independent bets is the mistake this list invites. | Div / hour |
|---|---|---|---|---|---|---|
| 10h→22hUTC | 25.3% | 30.1% | moved | chaos clock | 127.8 | |
| 0h→8hUTC | 23.2% | 22.0% | moved | chaos clock | 198.6 | |
| 5h→15hUTC | 21.5% | 35.2% | moved | chaos clock | 73.2 | |
| 11h→1hUTC | 21.0% | 35.8% | held | chaos clock | 632.5 | |
| 7h→23hUTC | 20.0% | 23.6% | moved | chaos clock | 96.9 | |
| 8h→2hUTC | 16.1% | 27.1% | moved | chaos clock | 818.9 | |
| 23h→7hUTC | 15.8% | 20.3% | held | chaos clock | 401 | |
| 11h→0hUTC | 15.3% | 22.9% | moved | chaos clock | 196.1 | |
| 11h→23hUTC | 14.6% | 15.4% | moved | chaos clock | 408.6 | |
| 23h→8hUTC | 14.6% | 23.4% | moved | chaos clock | 170.3 | |
| 4h→19hUTC | 14.2% | 8.5% | moved | chaos clock | 131.8 | |
| 21h→3hUTC | 14.2% | 21.1% | held | chaos clock | 765.6 | |
| 15h→2hUTC | 14.0% | 30.7% | moved | chaos clock | 1.6k | |
| 20h→7hUTC | 13.9% | 15.6% | held | chaos clock | 835.6 | |
| 10h→0hUTC | 13.8% | 18.8% | moved | chaos clock | 4.1k | |
| 8h→22hUTC | 13.6% | 19.4% | moved | its own | 1.1k | |
| 7h→22hUTC | 13.5% | 22.1% | moved | chaos clock | 51.1 | |
| 6h→0hUTC | 12.6% | 17.5% | moved | its own | 165.9 | |
| 20h→6hUTC | 12.6% | 13.0% | moved | its own | 74.5 | |
| 6h→21hUTC | 12.5% | 19.2% | held | chaos clock | 182.2 | |
| 4h→0hUTC | 12.5% | 16.9% | moved | chaos clock | 142.8 | |
| 5h→18hUTC | 12.4% | 19.2% | moved | its own | 160.9 | |
| 0h→7hUTC | 12.2% | 16.8% | moved | chaos clock | 586.3 | |
| 22h→7hUTC | 11.7% | 19.3% | moved | chaos clock | 201.4 | |
| 13h→2hUTC | 11.5% | 11.9% | held | chaos clock | 221 | |
| 10h→21hUTC | 11.4% | 13.6% | moved | its own | 85.2 | |
| 6h→22hUTC | 11.3% | 21.2% | moved | chaos clock | 96 | |
| 4h→19hUTC | 10.9% | 10.1% | moved | chaos clock | 144 | |
| 17h→10hUTC | 10.6% | 20.8% | moved | chaos clock | 195.5 | |
| 13h→0hUTC | 10.5% | 14.9% | moved | chaos clock | 354.7 | |
| 11h→23hUTC | 10.5% | 12.2% | moved | its own | 254 | |
| 13h→3hUTC | 10.1% | 10.9% | moved | its own | 76 | |
| 9h→22hUTC | 10.0% | 14.8% | moved | chaos clock | 403.4 | |
| 15h→22hUTC | 10.0% | 13.5% | moved | chaos clock | 404.3 | |
| 10h→1hUTC | 9.8% | 12.4% | moved | its own | 70 | |
| 14h→1hUTC | 9.8% | 15.9% | moved | chaos clock | 462.4 | |
| 9h→0hUTC | 9.5% | 6.0% | moved | chaos clock | 56.5 | |
| 6h→0hUTC | 9.1% | 21.4% | moved | chaos clock | 194.3 | |
| 19h→7hUTC | 9.1% | 14.4% | moved | its own | 88 | |
| 2h→13hUTC | 8.9% | 8.0% | moved | its own | 173.2 | |
| 14h→2hUTC | 8.4% | 13.5% | moved | chaos clock | 66.6 | |
| 0h→15hUTC | 8.0% | 12.8% | moved | its own | 52 | |
| 13h→21hUTC | 8.0% | 10.3% | moved | its own | 364.2 | |
| 23h→5hUTC | 7.9% | 11.4% | moved | chaos clock | 627.4 | |
| 0h→14hUTC | 7.7% | 13.0% | moved | chaos clock | 226.7 | |
| 0h→3hUTC | 7.7% | 6.8% | moved | its own | 103 | |
| 14h→21hUTC | 7.7% | 10.9% | moved | chaos clock | 1.9k | |
| 8h→3hUTC | 7.3% | 8.1% | moved | its own | 736.1 | |
| 0h→17hUTC | 7.3% | 11.1% | moved | its own | 87 | |
| 8h→16hUTC | 7.2% | 13.4% | moved | chaos clock | 2.3k | |
| 6h→1hUTC | 7.0% | 12.6% | moved | its own | 72.3 | |
| 0h→10hUTC | 7.0% | 12.0% | moved | chaos clock | 363.9 | |
| 7h→23hUTC | 6.7% | 6.7% | moved | chaos clock | 281.5 | |
| 22h→4hUTC | 6.5% | 6.7% | moved | its own | 261.1 | |
| 23h→5hUTC | 6.5% | 11.5% | moved | chaos clock | 1.2k | |
| 1h→12hUTC | 6.3% | 14.7% | held | chaos clock | 222.4 | |
| 6h→14hUTC | 6.1% | 13.6% | moved | chaos clock | 90.1 | |
| 9h→0hUTC | 6.1% | 6.6% | held | chaos clock | 399.3 | |
| 22h→9hUTC | 6.1% | 10.1% | moved | chaos clock | 2.5k | |
| 0h→13hUTC | 6.0% | 9.7% | moved | chaos clock | 77.3 |