Dashboard
Sharp / dumb / whale net positioning · 30d hourly · hyperdash, refreshed every 2h
Net % = (long − short) / (long + short). Latest snapshot per group with 30d sparkline + 24h/7d delta.
Pearson cross-correlation between hyperdash’s sharp cohort and two laggers (our curated SmartMoney list + hyperdash’s size-whale cohort). Bars show r across lags from −48h to +48h; positive lag = sharp leads the comparator. Peak marked with a dashed line. |r| < 0.3 is reported as “no clear relationship”.
When the three cohort lines diverge, the cohorts disagree. Sharp and dumb crossing each other is a regime-shift signal. The dashed light-gray line on the secondary right axis shows BTC’s rolling 7-day percent change for macro context.
Cleaner than sharp-vs-rekt because the market comparator isn’t PNL-defined — it includes everyone (~95K traders). Market netPct is computed from trader counts, not notional (perps net to zero in $-terms by construction, but the count of longs vs shorts CAN differ meaningfully — typically the herd is long-skewed). Gap shaded green when sharp leans more long than the herd, red when more short. Sustained gap = contrarian signal worth flagging.
Stockanswers “where is each cohort positioned right now?” A net % near +100 means the group is almost entirely long; near −100 means almost entirely short. The 24h / 7d delta tells you which way they’ve been moving recently.
Lagtests whether sharp money moves first. We resample the sharp series and each comparator to a shared hourly grid, then compute Pearson r across discrete lags. A peak at positive lag with r ≥ 0.3 means sharp leads the comparator by that many hours. Overlap chip counts wallets that appear in both our curated list and hyperdash’s labels for that cohort.
Flow plots all three cohort groups on one axis so you can see when they diverge. BTC’s rolling 7-day percent change is overlaid on a secondary right axis so cohort moves can be contextualized against macro crypto. Daily BTC values are linearly interpolated onto the hourly cohort grid.
Sharp vs Market compares sharp cohort positioning to the GLOBAL market average (all ~95K hyperdash-tracked traders, not just a PNL bucket). The gap is shaded so contrarian regimes — sharp leaning against the herd — jump out visually. Replaces the older sharp-vs-dumb divergence panel since the global market is a more neutral comparator than the PNL-defined rekt set.
Data sources: hyperdash HistoricalCohortPositioningV2 (cohort positioning, every 2h) + GetPerpsMarketParticipation (market totals, every 2h) + our internal smart_money_daily (curated wallets, every 30min). Cohorts grouped: sharp = extremely_profitable + very_profitable; dumb = rekt + very_unprofitable; whales = apex + whale.