Target vol
BTC daily — always long, exposure = 30% of target vol ÷ realized vol 30 bars, capped at 100%.
agent−1.79%
BTC held+0.38%
gap−2.17 pt
worst loss−9.87%
A sentiment score computed in code from our own measurements, with each component shown beside it. Descriptive, not predictive: the crowd's mood flips sign two quarters out of five.
What the components don't already show: price, drawdown from the high, open interest, realized volatility, stocks, the dollar, rates, stablecoins.
What the exchanges that publish their addresses hold (source: DefiLlama). When BTC or ETH falls, coins are leaving the platforms: net withdrawals. Stablecoins are money ready to buy with. Coinbase and the ETF custodians do not publish their addresses and are not included.
The rank of crypto apps in Apple's top 100, overall and Finance, in six countries: the attention of the general public. A crypto app entering the overall top has often accompanied euphoria. Google Play does not publish an open ranking.
The other data our agents read, each for its own markets. All of it is descriptive: none of it has been tested as a trading signal.
our data · descriptive, not predictive
The full charts, across four scales and all markets, are on the page Technical.
| Timeframe | Structure | RSI | MACD | ATR % | Bollinger | Support | Resistance | Volume |
|---|
What BTC did 1h, 4h and 24h after each 2026 US release covered by our hourly candles: close-to-close move from the last candle before the release time, minus the usual drift over the same horizon during the previous ten days. Each figure carries its n ; below five events we draw no conclusion on direction.