Rules-based model v1.1. The score starts at 50. Each factor adds or subtracts up to half its weight, with a signal between −1 and +1. Weights total 100. Missing data contributes zero and lowers coverage; its weight is never redistributed. This is an analytical weighting, not a calibrated probability or a verified forecast success rate.

24-hour weights: trend 25%, momentum 10%, buying pressure 10%, structure 10%, derivatives 20%, market context 15%, unlocks 10%. Seven-day weights: 30%, 10%, 10%, 15%, 10%, 10%, 15%, respectively. FDV, liquidations, account ratios and the order book provide context without additional score points.

The score and needle are hidden after a failed refresh, if the price is older than 10 minutes, if the primary candle series is stale or discontinuous, if fewer than 60 primary candles are available, or if coverage is below 55%. Missing trend timeframes retain their reserved subweights.

Trend uses EMA 50/200 and ATR. Momentum combines RSI and MACD. OI change is compared with the same contract’s mark price over the same interval. Funding uses its reported interval; eight hours is not assumed. Falling OI does not prove a price reversal.

Historical zones use daily and weekly pivots confirmed by three closed candles on each side, clustered using price and ATR. Up to 1,000 daily and 520 weekly candles are requested, subject to available listing history. Gate is used when Binance is unavailable; candles, ticker and the Bitcoin comparison stay on the same spot venue. Gate does not provide candle taker-buy volume, so that factor remains unavailable.

Auto-refresh runs every two minutes while this page is visible. Source caching is approximately one minute for prices, 1–5 minutes for candles and three minutes for OI. Gate price timestamps indicate retrieval time. All sources may be delayed or unavailable. Scenarios remain conditional.