A transparent look at the signal engine that scans, scores and filters the crypto market — and turns it into informational long/short verdicts you can track.
btccampus runs a disciplined pipeline every 5 minutes: scan the market, score candidates with rule-based technical analysis, pass them through strict risk gates and an AI review, then publish the survivors as signals with reference take-profit and stop-loss levels. Most candidates never survive the gates — by design.
Every 5 minutes, all liquid USDT perpetual markets are pulled from public exchange market data with fresh 1-hour candles, price and 24h volume. Coins under $10M daily volume or outside the 0.5%–12% move window are dropped immediately.
EMA 9/21/50 trend structure, RSI, MACD, Bollinger position and volume combine into a LONG / SHORT / NEUTRAL candidate with a 0–100 strength score.
Cooldowns, blacklists, over-extension vetoes and the news filter all get a veto. A market-wide AI layer can downgrade or pause new signals entirely when conditions turn hostile.
Each surviving candidate is published as a LONG / SHORT verdict with ATR-based reference take-profit and stop-loss levels, its strength score and the gates it passed.
Signals, price alerts, whale wallets and your on-chain portfolio are tracked continuously — you decide what, if anything, to do with them.
Every number in the system comes from a real, named source — nothing is simulated. These are the feeds the engine and the dashboard actually use.
Live 1-hour OHLCV candles and tickers from public exchange market data (OKX, BingX)
Liquidations, long/short ratio and funding rates from Coinglass
Crypto news reviewed by an AI layer for per-coin red flags, plus the Fear & Greed index
Large transfers from Whale Alert and BTC exchange-flow analytics from CryptoQuant
Total market cap, BTC dominance and coin metadata from CoinMarketCap
A Claude model reads the data above and produces the market bias and risk verdict
Market data refreshes every few minutes and is cached between runs. We show real numbers at honest intervals — never fabricated “real-time” noise.
Rather than relying on a single indicator, the engine combines four independent perspectives. Each one can veto or weaken a signal, and their agreement is what produces a high-strength candidate.
Deterministic math over 1-hour candles: EMA 9/21/50 alignment and slope for trend, MACD cross and histogram for momentum, RSI(14) for exhaustion, Bollinger position for volatility context, ATR(14) for exit distances.
Watches the conditions that wreck leveraged positions: liquidation cascades, extreme funding rates and volatility spikes. When the environment turns hostile it downgrades signals or pauses new ones entirely.
An AI layer reads crypto headlines and flags coins with fresh red flags — hacks, delistings, regulatory hits — so the engine skips them. Whale transfers and BTC exchange flows add context on where big money is moving.
A Claude model reviews the combined picture — market bias, risk conditions, news — and issues a verdict the engine obeys: publish normally, flag elevated risk, or stand down until conditions improve.
Indicator confluence produces a 0–100 strength score for every candidate, alongside a confidence value. Strength does not just rank candidates — it tells you how much the indicators actually agree.
Strength 80+ marks strong confluence; 60–79 moderate; below 60 weak. Confidence must be at least 50 or the candidate is dropped. And counter-intuitively, a sky-high score is treated with suspicion — a coin already up more than 7% in 24h is vetoed for LONG signals, because chasing is how accounts die.
Markets behave differently under different conditions. Our system continuously classifies the current market regime and adapts its analysis accordingly. This regime-awareness helps prevent false signals that often occur when applying a one-size-fits-all approach.
Strong directional movement with momentum confirmation
Sustained bearish pressure with lower highs/lows
Consolidation between defined support and resistance
Elevated price swings with unpredictable direction
Compressed ranges often preceding major moves
Market shifting between regimes with mixed signals
Regime awareness feeds the engine in concrete ways: altcoin signals are checked against what Bitcoin is doing, ATR widens stop distances in volatile regimes, and the AI risk layer stands the whole system down when conditions turn hostile.
Not every high-scoring candidate becomes a signal. Before anything is published, a candidate must clear every gate below — a single failure drops it and logs the reason, which you can inspect in AI Coin Analysis.
Active Risk Filters Include:
A candidate that survives every gate is published with a complete structure — nothing is left open-ended, and everything is visible in your dashboard. Signals are informational, not instructions to trade.
The reference price at the moment the signal was generated
Calculated from volatility, never improvised
So you can judge the signal, not just follow it
Tracked against live prices, reported honestly
Track the market with a disciplined, transparent signal engine — plus alerts, whale tracking and your on-chain portfolio. You stay in control of every decision.