Scores rank present short-side conditions — how stretched, overbought, and rolling-over each asset is right now, across 1m / 5m / 15m. This is not a forecast. 1m–15m price direction is dominated by noise and cannot be reliably predicted from past candles. Use this to locate setups and size risk — never as a signal that a drop is coming. The backtest below exists to measure whether a signal actually has an edge; a signal that shows none is a true, useful result.
Asset
Price
1m Δ
RSI14
vs EMA20
Funding
Align 1·5·15
Short Score
Conditions
70+ strong setup45–69 building<45 not set up· Align pills = per-timeframe short condition (1m / 5m / 15m); all three lit = multi-timeframe agreement
Backtest — does the signal actually have an edge?
How this is measured (no lookahead): at each closed candle the price-signal score is recomputed from
only candles up to that point. A signal fires when the score crosses (or sits above) your threshold.
From the signal's close we measure the forward price move at 5 / 15 / 30 minutes and the worst move
against a short (max adverse excursion) over that window. Hit rate = share of signals where
price was lower at the horizon (a short would have profited).
The backtest score uses the price-derived components only — RSI, EMA20 extension, momentum — renormalised to 0–100.
Funding is excluded because Bybit's kline history carries no per-candle funding, and cross-timeframe
alignment isn't applied here — the point is to isolate what past candles can actually be tested on. Horizons shorter
than one candle of the chosen timeframe are skipped. This measures a signal; it does not manufacture one. Flat results
are the expected, honest outcome for noise-dominated intraday direction.