Entry timing: data and coverage

The same test for every coin

At each observed hour, compare the price with 24 hours earlier. Accept the first rise of at least 20%, then allow no further signal for that coin for 72 hours. Compare its price 24 hours after the signal with the signal price: lower, or flat/higher.

Exclude the first seven days of each available series and require all hourly observations from 24 hours before through 48 hours after the signal. Missing hours are not filled. These are sampled prices, not intrahour lows or guaranteed trade execution prices.

Which 20 coins?

The first 20 rows of the main homepage’s expanded Inflation list, in its default order, observed Sep 9 2026. The selection is fixed as of that date, not reconstructed historically. Today’s surviving list introduces selection bias. The separate 500-coin inflation monitor was not substituted.

Coverage and counts

The requested period begins Jan 1 2020 and ends after Sep 8 2026 UTC. Later-listed coins begin with their available history. Across 547 signals, 309 ended lower the next day and 238 ended flat or higher. The table records missing hours inside each observed series; affected windows were excluded.

Price sources and checks

Sixteen coins use spot-exchange hourly USDT opens; XMR and LEO use another exchange’s USD opens. HYPE and Canton use hourly USD aggregate prices. Each coin uses one source throughout the extension. The original three-coin chart keeps its earlier study period and sources, identified on the charts.

All 547 signal timestamps and outcomes were recalculated independently from the saved hourly data. Recent exchange events received 324 reference-price comparisons against an aggregate feed. Small source differences can change threshold-crossing hours: two of the 108 checked next-day directions differed between sources. Counts therefore describe the designated price series, not an exact universal count across all venues.

Additional spot checks corroborated flagged ADA and UNI prices, all seven Canton events, and five recent HYPE events. Earlier HYPE spot history was unavailable at that comparison venue. One ZEC comparison contained only stale, zero-volume candles and was rejected; the primary exchange and aggregate feed both still showed a qualifying rise and next-day decline.

What this can tell us

The results describe past prices after this specific signal. Coins have different observation periods and sample sizes; Bitcoin and LEO each had only four signals. Events across coins can reflect the same market shock and are not independent trials. The pooled total gives more weight to coins with more signals.

This study does not compare against ordinary trading hours, identify the best waiting period, or test returns from buying after a fixed delay. It does not include spreads, fees or slippage. A lower entry price can keep falling.

The ARB and HYPE illustrations were selected after observing the data. The outcome charts include all signals satisfying their fixed rule, rather than only the illustrated episodes.

Download the 20-coin counts and coverage (CSV)