Crypto research
that shows its work.
We test market ideas with historical coverage that includes delisted coins, then publish plain verdicts. Including the ideas that failed.
NewStudy 01: calm coins beat wild coins ›Alt Radar. Every alt against the market.
TOTAL, TOTAL2, TOTAL3 and OTHERS rebuilt from scratch, then each coin measured against them: how closely it follows, how hard it falls, and whether it was breaking away in the saved window.
Rebuilt, not borrowed
Cap-weighted TOTAL indexes from exchange prices and supply, each checked against the published versions.
Leave-one-out
Each coin is compared with the index rebuilt without itself, so big coins don't grade their own homework.
Who's moving alone
When a coin's 30-day correlation breaks from its 1-year norm, it gets flagged.
Calm versus wild.
See the comparison.
Across: how wild each coin has moved over 30 days, compared with the rest. Up: its 30-day move. In the 2025–26 comparison, the calmest fifth beat the typical coin 68% of the time; the wildest fifth, 32%. This is not a Bitcoin beat rate.
Select a coin to inspect its measurements.
View as table
The lottery effect, on clean data.
Every day we ranked the 100 most-traded coins on Binance by how wildly they'd moved over the past month, then watched the next 30 days.
Wild coins feel like the opportunity. Most of the time they're the trap. A few moon, and those rare wins keep people buying lottery tickets.
The group-level difference appeared in every tested year, but largely disappeared in strong alt-season months. The effect is known in finance. Calm alts did not beat Bitcoin over the compounded test.
View as table
| Year | Calm | Wild |
|---|---|---|
| 2020 | 58% | 45% |
| 2021 | 56% | 42% |
| 2022 | 64% | 38% |
| 2023 | 59% | 38% |
| 2024 | 63% | 41% |
| 2025 | 71% | 33% |
| 2026 Jan–Sep | 65% | 34% |
We publish the misses too.
Most research only shows the winners. Knowing what fails is half the edge.
| Idea | Tested on | What we found | Verdict |
|---|---|---|---|
| Calm beats wild › | 680 Binance pairs, 2020–26 | Calm coins beat the typical coin 68% of the time in 2025–26, wild coins 32%. Held every year. | ✓ HOLDS |
| Bottoming signals › | Same universe, exchange data | Coins with all three bottoming signs did worse than coins with none: 33–42% vs 44–47%. | ✕ FAILED |
| Chart fingerprints › | 100 coins, hourly, 2021–26 | Shazam-style matching recognised charts 99% of the time. Its forecasts lost to simple rules. | ✕ FAILED |
| Mood gauges timing BTC › | Funding, flows, breadth | No measure predicted BTC's direction over 7 or 30 days. | ✕ FAILED |
| Planetary cycles › | BTC and alts, 2019–26 | No link to direction, 30–60 day mood, or alt season. The sign flipped year to year. | ✕ FAILED |
Dead coins count
Studying only today's survivors flatters everything. The stored Binance catalogue includes histories that ended; it is not every global token.
Rules before results
Registered checks and exploratory work are labelled separately. Once we inspect a later period, it is no longer fresh for new choices.
Plain verdicts
Holds, failed, or too early to tell. With the numbers, in words anyone can follow.
Get the next
research note.
Email requests available · automated newsletter not connected