Combine indicators that answer different questions, then define risk independently of the signal.
Give every indicator one role
A useful trading stack is not a vote among as many indicators as possible. Give each component a job: a trend filter determines directional context, a volatility envelope identifies unusual expansion, and price structure defines invalidation. Indicators derived from the same closing-price series are correlated, so three momentum oscillators rarely provide three independent facts.
Timeframe matters as much as formula. A 20-period calculation on 15-minute candles describes a different market horizon from the same calculation on four-hour candles. Crypto trades continuously, but candle boundaries and venue data still affect values.
The chart logic implemented here
The /charts service retrieves OHLCV candles for 15-minute, 1-hour, and 4-hour intervals. Its Ichimoku conversion line is the 9-candle high-low midpoint; the base line uses 26. Span A averages those lines, while Span B uses the 52-candle midpoint; both are displaced 26 candles. Price above the cloud grants long permission, below grants short permission, and inside produces no-trade permission.
Keltner Channels use a 20-period exponential moving average of closes. The envelope is that average plus or minus two times a 20-period Average True Range. In this implementation, a long breakout requires long permission and the latest two closes above their Keltner upper bands; a short breakdown is the mirror condition. Confidence can be raised when channel slope agrees with direction and cloud thickness is meaningful relative to price. This is educational status logic, not an order or performance claim.
Separate confirmation from risk
Waiting for a second close can filter a one-candle excursion, but enters later and cannot prevent reversals. ATR measures recent movement size, not direction. Place an invalidation level, then size exposure so loss there fits a predetermined budget.
The chart service’s signal is distinct from the home scanner’s heuristic pressure and from DeepBook order-book observations on /dashboard. DeepBook bid or ask notional is resting price × quantity, not a count of longs and shorts. Server alerts and UI labels can also differ, so rely on the explanation displayed with a signal rather than assuming shared exact thresholds.
Know what is educational only
RSI, MACD, Bollinger Bands, funding, and liquidation data are not part of the inspected chart calculation. The app also has no social/news sentiment feed. If consulted externally, document the provider, coverage, formula, and timestamp.
- Use closed candles consistently when comparing confirmations.
- Record the selected symbol and timeframe before interpreting a chart.
- Treat missing or newly initialized values cautiously; long lookbacks need enough history.
Worked Ichimoku–Keltner decision
Consider fictional one-hour values solely to demonstrate the implemented decision sequence.
- Latest close and cloud
- $105; cloud top $101; cloud bottom $98
- Keltner values
- Middle $100; upper $104; lower $96
- Previous candle
- Close $104.50; previous upper $103.80
- Status
- Above cloud; two closes above respective upper bands
The chart logic would grant long permission and mark a confirmed breakout. That does not estimate win probability. An illustrative invalidation below $101 creates $4 risk per unit from $105; a $200 risk budget would imply 50 units before fees, slippage, and gap risk.
Common interpretation traps
- Stacking several price-derived indicators and mistaking correlated agreement for independent confirmation.
- Reading ATR or a Keltner boundary as directional prediction rather than a description of recent volatility.
- Treating chart permission, home pressure, DeepBook notional, and server alerts as one unified signal system.
Frequently asked questions
Why can Ichimoku values be missing on early candles?+
The lines require sufficient 9-, 26-, and 52-candle history, and cloud spans are displaced. A valid long-lookback value cannot be formed immediately.
Does two-candle confirmation eliminate false breakouts?+
No. It is a rule that demands persistence beyond the channel, but price can reverse after any number of confirming closes.
Are indicator settings universal?+
No. The described values match this app’s implementation. Different platforms may initialize averages, source candles, or displacement displays differently.