Does activity move between coin groups?
Apparent handoffs repeat, but timing shuffles produce plenty of repetition too.
The verdict
◊ INCONCLUSIVEApparent handoffs repeat, but timing shuffles produce plenty of repetition too.
A possible handoff
Imagine one coin rises with increased trading activity. A week later, it fades while another strengthens. That is consistent with a rotation story—but it also happens when unrelated traders act independently.
The screen
Across the stored database, we looked at weekly periods from 2020 onward. A strengthening coin needed at least a 10% price gain, a gain at least five percentage points above Bitcoin, volume at least 1.5 times its earlier average, and trade counts at least 1.25 times their earlier average. Minimum history and volume rules excluded many thin listings.
The following week, the leader had to stop rising and lose at least 30% of its share of measured spot trading volume. A different coin had to meet the strengthening conditions. This measures activity, not dollars transferred between accounts. OI was not required in this broad screen.
A real example
| Leader week ends | BEL rises | Following week ends | BEL then | AXS then | AXS volume |
|---|---|---|---|---|---|
| 2021-01-23 | +50.5% | 2021-01-30 | -7.7% | +21.0% | 2.33× |
| 2021-04-03 | +45.4% | 2021-04-10 | -0.3% | +49.2% | 2.76× |
| 2021-07-17 | +35.8% | 2021-07-24 | -6.5% | +127.3% | 3.73× |
| 2025-07-12 | +28.1% | 2025-07-19 | -1.8% | +10.0% | 1.56× |
| 2026-04-18 | +11.2% | 2026-04-25 | -3.4% | +40.5% | 6.88× |
BEL → AXS was selected after counting repeated pairs. The table shows five qualifying sequences, not evidence that BEL traders bought AXS.
Could repetition happen by coincidence?
We moved the receiver groups between weeks within the same year, quarter and rough Bitcoin-return conditions. About one in five shuffled timelines produced at least as many later-recurring pairs. That is weak timing evidence.
The screen generated many overlapping pair combinations. These are not independent money-flow events, and the shuffle is a descriptive check rather than proof of causality.
What comes next
Useful follow-up work would compare entire groups, separate market-wide movement from coin-specific activity, and test a fixed handoff rule on fresh periods. Tracing actual funds would require different data, such as linked on-chain transfers; most centralized-exchange trading is not visible that way.
Inspect the overlap timeline ›