Which indicator signals actually work, coin by coin.

1. 01

   RSI crossed below 30

   expects a rise

2. 02

   RSI crossed below 25

   expects a rise

3. 03

   RSI crossed below 20

   expects a rise

4. 04

   RSI crossed below 15

   expects a rise

5. 05

   RSI crossed above 70

   expects a fall

6. 06

   RSI crossed above 75

   expects a fall

7. 07

   RSI crossed above 80

   expects a fall

8. 08

   RSI crossed above 85

   expects a fall

9. 09

   MACD crossed above signal

   expects a rise

10. 10

    MACD crossed below signal

    expects a fall

11. 11

    SMA50 crossed above SMA200

    expects a rise

12. 12

    SMA50 crossed below SMA200

    expects a fall

13. 13

    Price crossed above EMA200

    expects a rise · approximate

14. 14

    Price crossed below EMA200

    expects a fall · approximate

15. 15

    Supertrend flipped bullish

    expects a rise · approximate

16. 16

    Supertrend flipped bearish

    expects a fall · approximate

17. 17

    Close broke below the lower Bollinger band

    expects a rise

18. 18

    Close broke above the upper Bollinger band

    expects a fall

19. 19

    Volume spiked above 3× normal

    expects either

1. 01Pick the signal and the timeframeThe masthead is a sentence: after RSI crossed below 30 on the 4h chart. Change any word from its menu. Picking a bearish signal turns the board to face falls, so the leaders are the coins that dropped most.
2. 02Set the window and the lookbackHow long after the signal to measure — a day to a month on the 4h — and how far back to count firings, up to 180 days. The board refuses a pair that cannot be answered, like a 30-day window over a 30-day lookback, and says why.
3. 03Read the podium against the fieldThree leaders by edge, then every qualifying coin. Simple view shows five columns with bars on one shared scale; Detailed adds baseline, run-up, best and worst, percentiles and the last firing. The field row is the line to beat.
4. 04Open a row, then see the exact barA coin's row opens its ledger — every firing, dated, with its outcome. Click one and it is drawn on the chart: the bar it fired on with the stored indicator value, the threshold, and the window's end. Then hand the signal to an indicator alert on that coin.

Indicator Edge — common questions

Does RSI actually work in crypto? How would I know?

By checking what happened after it fired, not by trusting the textbook. The Indicator Edge board measures indicator performance the honest way: it takes every time RSI crossed below 30 on a coin — or any of 18 other signals — and records where the price was one day, three days or a week later, over hundreds of past firings. It then subtracts that coin's own baseline: its median move over a window of the same length, measured from every bar in its history rather than only the ones that followed a signal. Note what that means when you narrow the lookback — the signal sample shrinks to the period you picked, while the baseline stays the coin's long-run drift, so edge is a comparison against normal, not against that period. What is left is the edge: how much the signal actually added, per coin, on the timeframe you trade. Some coins bounce reliably from RSI 30; others keep falling. The board shows which is which.

How is the edge of an indicator signal calculated?

Edge is the median move after the signal minus the coin's own baseline — its median move over the same window measured from every bar in its history, signal or not. Ranking raw post-signal moves would rank beta and drift: in a one-directional market every coin's RSI crosses 20 in the same few hours, and the highest-beta coin tops every board whatever the indicator did. Subtracting the baseline means a coin that rises anyway earns no credit for rising. The headline number on every row is that excess, and a field row at the bottom shows every coin combined so each result is read against the whole board rather than in isolation.

Why does each row show two signal counts?

Because firings are not the same as observations. One sustained drawdown can push RSI under 30 a dozen times in a row, and each firing opens a forward window over almost the same bars. Counted naively that reads as twelve signals with a perfect hit rate for what is really one event. The board de-overlaps those windows and reports independent signals alongside the raw firings — 24 of 41, say — and the minimum-signals filter uses the independent count, never the raw one. Win rates are greyed out under 20 independent signals because small samples make that number jumpy.

Which indicators and signals does the board cover?

Nineteen threshold-cross signals across seven indicators. RSI crossing below 30, 25, 20 or 15 and above 70, 75, 80 or 85; the MACD line crossing above or below its signal line; the 50-period simple moving average crossing above or below the 200-period (golden and death crosses, on whichever timeframe you pick); price crossing above or below the 200-period EMA; Supertrend flipping bullish or bearish; a close breaking below the lower or above the upper Bollinger band; and volume spiking above three times its normal level. Each carries the direction it is supposed to predict, so a bearish signal is judged on how far the coin fell.

Which coins, exchanges and timeframes are scored?

Every coin CoinBeacon holds candles for on five venues — Binance spot and the Binance, Bybit, OKX and Gate.io perpetual-futures markets — which is 700+ coins, one row per coin with the venue chosen by sample size. Four timeframes: 15m, 1h, 4h and 1d, each with its own set of forward windows and lookbacks. Signals are detected on closed candles only, so nothing repaints, and every firing is stored the moment it happens — the ledger currently holds more than three million scored signals and grows with every candle.

Is this a backtest?

It is the honest half of one. Every historical firing of a signal is recorded, and the coin's forward return at fixed horizons after it is measured and stored, so what you see is the real distribution of outcomes for that signal on that coin. It is not a strategy simulation: there are no stops, no position sizing, no fees or slippage, and no compounding. Two other things worth knowing. A settled outcome is never re-scored, so the record cannot drift after the fact. And the two indicators that carry state from wherever their history began — EMA200 and Supertrend — are marked approximate, because a reconstructed reading can differ slightly from a live one on the same bar.

Will the board tell me which indicator to trade, or which coin to buy?

No. It reports what followed each signal in the past, per coin, with the sample size next to it — nothing more. A positive edge over 30 independent signals is evidence a signal has meant something on that coin; it is not a prediction that the next firing will. The point is to let you check a signal's record before you build an alert around it, and to stop you trusting an indicator on a coin where it has never worked. Everything on this page and in the app is market data, not financial advice.

Is the Indicator Edge board free?

Yes, and you can look before you sign up: the board runs for anyone, showing the top twenty coins for whichever signal and timeframe you pick. A free account opens the rest — the full ranking rather than the head of it, coin search, venue pinning, and the ledger behind every row: each individual firing, dated, with what price did after it. No card to enter and no exchange API keys to connect.