Crypto Winner Scanner | Multi-Asset Intelligence

CRYPTO SCANNER v2 Upgraded Feb 12: BTC regime gate, raised thresholds (75+), BTC correlation rejection, RSI overbought filter, 8h resolve window, 2:1 TP/SL ratio. Prior results: 8.3% WR. Tracking new performance from this point forward. See learned algorithms for our best crypto system (87.5% WR).

Win Rate

Resolved signals

Avg P&L

Per signal

Total Signals

Best Trade

Unique Pairs

Last Scan

Current Winners loading...

Loading winners...

Win Rate by Confidence Tier

Recent Signal History

Loading history...

Full Scan Log — All Analyzed Coins iEvery coin that was deeply analyzed in a scan, including those that didn't make the 70-point threshold. Shows all 7 indicator scores so you can see exactly why each coin was accepted or rejected.

Loading scan log...

Understanding Confidence Tiers iThe scanner assigns a verdict based on the total score. Higher tiers mean more indicators agree.

▼

Each coin is scored 0-100 across 7 independent indicators. The total score determines which confidence tier it falls into. Higher tier = more indicators agreeing = higher conviction. v2 raised the minimum threshold from 70 to 75 and removed the Lean Buy tier (which had the worst win rate).

Strong Buy

Score 85-100

6-7 out of 7 indicators are strongly bullish. Momentum, volume, RSI, moving averages, MACD, price patterns, and breakout position are nearly all firing. Target: +3-6% within 8 hours. Risk: -1-2%. TP:SL ratio >= 3:1. This is the highest conviction level.

Buy

Score 75-84

5-6 indicators are bullish, with maybe 1-2 being neutral or weak. Strong agreement but not unanimous. Target: +2.5-5% within 8 hours. Risk: -1-2%. TP:SL ratio >= 2:1. Solid setup with most factors aligned.

Below 75: Not shown as a winner. The previous threshold was 70, which let in too many marginal signals. Lean Buy tier (70-74) was removed after performance analysis showed it contributed the most losses.

Glossary of Indicators iEach of the 7 factors explained in plain language — what it measures, why it matters, and how it's scored.

▼

Multi-Timeframe Momentum 0-20 pts

What it measures: Is the price going up on BOTH the 1-hour and 5-minute charts? If yes, the uptrend is consistent across different time scales.
How it scores: +10 if both timeframes are positive. +5 more if 4-hour momentum exceeds 1%. +5 more if short-term is accelerating (0.5%+).

Why it matters: A coin going up on multiple timeframes is more likely to keep going than one moving up on just one chart.

Volume Surge 0-20 pts

What it measures: Is the current trading volume higher than normal? Volume is how many coins are being bought/sold. A "surge" means more people are trading than usual.
How it scores: +8 if volume is 1.5x average. +6 more if 2x. +6 more if 3x or higher.

Why it matters: Price moves with high volume tend to sustain. Price moves on low volume often fizzle. Volume confirms conviction.

RSI Sweet Spot 0-15 pts

What it measures: The Relative Strength Index (RSI) is a number from 0-100 that shows if a coin is "overbought" (too high, might drop) or "oversold" (too low, might bounce). 50-70 is the sweet spot — momentum is strong but it's not dangerously overheated.
How it scores: 15 pts for RSI 50-70 (ideal). 8 pts for RSI 40-50 (warming up). 5 pts for RSI 70-80 (hot but risky). 0 pts for RSI above 80 or below 40.

Why it matters: Buying something with RSI above 80 often means you're buying at the top. The sweet spot catches momentum early, before it gets dangerous.

Above Moving Averages 0-15 pts

What it measures: Moving averages smooth out price noise. The SMA-20 (average of last 20 hours) and SMA-50 (last 50 hours) act as support levels. If the price is above both, the trend is up.
How it scores: +8 if price is above SMA-20. +4 if above SMA-50. +3 if SMA-20 is above SMA-50 (called a "golden cross" structure — a classic bullish signal).

Why it matters: Professional traders watch these levels. When price is above its moving averages, buyers are in control.

MACD Bullish 0-10 pts

What it measures: MACD (Moving Average Convergence Divergence) compares fast and slow momentum. When the MACD line crosses above its signal line, it indicates strengthening upward momentum.
How it scores: +5 if the MACD line is above the signal line. +5 more if the histogram (difference between the two) is positive AND growing — meaning momentum is accelerating.

Why it matters: MACD catches trend changes early. An expanding histogram means the up-move is gaining steam, not losing it.

Higher Highs & Higher Lows 0-10 pts

What it measures: The most basic definition of an uptrend: each peak is higher than the last (higher highs) and each dip is higher than the last (higher lows). Checks the last 6 hourly candles.
How it scores: +5 if at least 2 consecutive higher highs. +5 if at least 2 consecutive higher lows.

Why it matters: This is Trend Analysis 101. Higher highs + higher lows = textbook uptrend. The absence of this pattern often precedes reversals.

Near 24h High 0-10 pts

What it measures: Where is the current price relative to its 24-hour high and low? If the price is near the top of its daily range, it suggests buying pressure and potential breakout.
How it scores: 10 pts if price is in top 15% of its daily range. 6 pts if in top 30%. 3 pts if in top 45%. 0 if in the bottom half.

Why it matters: Coins trading near their 24h high often break through to new highs. Coins in the bottom half of their range are weaker.

Our Methodology — How Picks Are Made iStep-by-step explanation of how we go from 600+ coins to a handful of winners.

▼

Here's exactly what happens every 15 minutes, end to end. No black box — you can verify every step.

Fetch all tradable pairs — We query the Crypto.com Exchange public API for every instrument. We filter to USDT spot pairs only (the most liquid, easiest to trade). This typically yields 600-700 pairs.

Pre-filter for activity — We discard any pair with less than $20,000 in 24h trading volume (too thin to trade safely) or negative 24h price change (we only look for upward momentum). This typically cuts to 30-150 candidates depending on market conditions.

Rank by momentum, take top 60 — Remaining candidates are sorted by 24h change. We deep-analyze the top 60 to keep scan time under 90 seconds (each pair requires 2 additional API calls for candlestick data).

Deep analysis with 7 indicators — For each of the top 60, we fetch 1-hour candles (48 hours of history) and 5-minute candles (4 hours of micro-structure). We then run all 7 technical indicators described in the glossary above. Each indicator scores independently — no indicator knows what the others scored.

Confluence threshold: 70/100 — Only coins scoring 70+ out of 100 are promoted to "winner" status. This means at least 5 of 7 indicators must be meaningfully bullish. The principle is confluence — one bullish signal is noise, but 5+ agreeing signals form a pattern.

Record signal with targets — Each winner is saved to the database with its exact entry price, score, all factor details, a profit target (+1.5% to +3% depending on tier), and a risk level (-1.5%). This creates an auditable record.

Resolve 4 hours later — A separate job checks every signal after 4 hours. Did the price hit the target? Did it hit the stop? Did it land somewhere in between? The outcome is recorded as win, partial_win, loss, or partial_loss — and shown publicly on this page.

Why we stand by this approach

Multi-factor confluence is well-documented in academic literature. Individual technical indicators are weak predictors (most hover around 50-55% win rates in isolation). However, when multiple independent indicators agree simultaneously, win rates historically improve to 60-70%+ in momentum-favorable conditions. Our approach is based on the same principle used by quantitative trading firms: don't bet on one signal, bet on the alignment of many.

Full transparency. Every single signal is recorded and publicly resolved. We don't hide losses, we don't cherry-pick wins. The stats section at the top of this page shows the actual, real-time win rate. If the scanner is underperforming, you'll see it.

Is the Algorithm Self-Learning? iHonest answer: partially. Here's what adapts and what doesn't.

▼

What IS Fixed (by design)

The 7 indicators — Momentum, Volume, RSI, MAs, MACD, HH/HL, and Near-High are always used. These are fundamentals of technical analysis that have worked across decades of markets.
The scoring weights — Momentum and Volume are weighted highest (20 pts each) because they're the strongest short-term predictors. This weighting is based on published research, not gut feeling.
The 70-point threshold — We intentionally set a high bar so the scanner stays selective rather than flooding you with marginal signals.

What Adapts Over Time

Outcome tracking — Every signal's result is recorded. Over time, the leaderboard reveals which score tiers and which specific coins have the best track records.
Weekly analysis — A scheduled job runs every Sunday analyzing patterns: which tier wins most? Which coins repeat as winners? This data informs whether thresholds should change.
Market-aware filtering — The pre-filter (positive 24h change + volume minimum) naturally adapts to market conditions. In a bear market, fewer coins pass → fewer signals → the scanner stays quiet rather than forcing bad calls.

Bottom line: The core scoring logic is fixed and rule-based — not a neural network or black box. We believe this is a feature, not a bug. Fixed rules are transparent, auditable, and don't overfit to recent noise. The self-learning layer sits on top: it tracks which configurations work best and can inform future parameter adjustments. Think of it as "the rules are stable, but we measure how well they're working."

Known Limitations & Areas for Improvement iWe believe in full transparency — including what this tool can't do.

▼

No tool is perfect. Here's what we know about the scanner's weaknesses and what we're working to improve:

Current Limitations

⚠Momentum bias. The scanner only looks for coins going up. It cannot predict sudden reversals caused by news, exchange delistings, or whale dumps. A coin scoring 90 can still drop 20% if Elon tweets about it.

✔~~4-hour resolve window.~~ FIXED — Continuous resolve. We now walk through every 5-minute candle during the 4-hour window and detect if the target or stop price was hit at any point. No more "false losses" from temporary spikes that reversed.

⚠No sentiment or news analysis. The scanner is purely technical (charts and numbers). It doesn't know about upcoming token unlocks, regulatory announcements, partnership news, or social media hype. Fundamentals matter.

⚠Single exchange. We only scan Crypto.com Exchange. Price movements on Binance, Coinbase, or Kraken might differ. Arbitrage opportunities or exchange-specific liquidity issues aren't captured.

⚠Fixed weights. Momentum and Volume are always worth 20 points each. In certain market conditions (e.g., low-volatility sideways markets), other indicators like RSI or MACD might be more predictive. The weights don't adapt to regime changes.

⚠Top-60 cutoff. We only deep-analyze the top 60 candidates by 24h change. A coin with modest 24h change but perfect technical structure might be missed if 60 other coins had bigger moves that day.

⚠No historical stress tests. Our algorithms haven't been backtested against the 2008 crash, COVID crash, or other black-swan events. Self-learning only trains on recent trades, so extreme tail-risk scenarios are uncharted territory.

✔~~No statistical significance testing.~~ FIXED — Binomial significance. We now run binomial hypothesis testing (normal approximation with Wilson confidence intervals) on win rates. The Stats & Leaderboard panels show whether performance is statistically significant or just noise.

⚠No real-money track record. Everything is paper-traded or backtested. We have no audited live-money performance. Paper trading doesn't account for slippage, partial fills, or the psychological pressure of real capital.

⚠Recency bias in learning. The self-learning system optimizes against recent closed trades. If market regime shifts dramatically (bull → bear, low vol → high vol), learned parameters may become stale before the system adapts. There's no mechanism yet to detect regime changes proactively.

✔ Recently Implemented

✔Continuous resolve. Now walks through every 5-minute candle in the 4-hour window. Detects if target or stop was hit at any point, not just at the end. Peak price and hit timestamps are recorded.

✔Volatility-adjusted targets. Targets and risk levels are now based on each coin's ATR (Average True Range) instead of fixed percentages. A volatile meme coin gets wider targets; a stable large-cap gets tighter ones.

✔Correlation filtering. After scoring, the scanner checks each winner's move against BTC. If BTC pumped >2% and the coin is just following BTC (within 30% correlation), it gets a 5-point score penalty. Independent movers rank higher.

✔Statistical significance testing. Binomial hypothesis testing with Wilson confidence intervals on all win rates. Stats panel now shows whether performance is statistically meaningful or potentially noise.

✔Discord alerts. Strong signals are automatically posted to Discord via webhook when the scanner finds winners. No need to check the page constantly.

Still Planned

○Multi-exchange data. Cross-reference signals with Kraken and Coinbase order book depth to confirm liquidity and validate price action across venues.

○Adaptive weight tuning. Use the outcome data we're collecting to gradually adjust indicator weights. If Volume Surge consistently predicts winners better than MACD in certain conditions, the system should learn that.

Before You Do Anything With This Page

What this page does: Every 15 minutes, our scanner checks 600+ cryptocurrencies on Crypto.com Exchange. It looks at 7 different technical indicators (momentum, volume, RSI, moving averages, MACD, price patterns, and breakout proximity). If 5 or more indicators agree that a coin is likely to keep going up, it appears here as a "winner."

What the score means: Each coin gets a score from 0-100. Scores 85+ = "Strong Buy" (very high confidence), 75-84 = "Buy" (high confidence), 70-74 = "Lean Buy" (moderate confidence). We track every single signal and publicly show the actual win rate — no hiding losses.

What "win" means: We check every signal 4 hours later. If the price went up by the target % = win. If it dropped past the risk % = loss. We show the real numbers, not cherry-picked ones.

The honest truth: No scanner in the world can predict crypto prices with certainty. Crypto can drop 10-50% in hours. This tool finds patterns that have historically worked more often than not — but "more often than not" still means you WILL lose sometimes. Never put in money you can't afford to lose. This is a research tool, not a money printer.

System Analysis DIAGNOSIS

How the crypto scanner works:
Scans 600+ pairs on Crypto.com Exchange every 15 minutes. Each coin gets a composite score (0–100) from 7 technical indicators: volume surge (unusual activity vs 24h average), momentum score (rate of price change), RSI positioning (50–70 sweet spot), MACD crossover strength, Bollinger Band breakout, support/resistance proximity, and market cap weighting (larger caps = more reliable signals). Scores above 85 = Strong Buy, 75–84 = Buy, 70–74 = Lean Buy.

v2 fixes applied (Feb 12, 2026):
• BTC regime gate: Scanner pauses entirely when BTC is below its 24h SMA (bear regime). This is the #1 fix from the live_signals system which uses the same gate.
• Raised threshold: 75+ score required (from 70). Lean Buy tier removed.
• BTC correlation rejection: Coins moving in lockstep with BTC are fully rejected (not just -5 penalty). Only independent movers pass.
• RSI overbought filter: Coins with RSI > 75 are rejected (buying at the top).
• Volume floor raised: $100K minimum (from $20K). Low-volume coins are manipulation targets.
• Pump-and-dump filter: Coins up >30% in 24h are excluded (likely to reverse).
• 8h resolve window: Extended from 4h to give moves more time to develop.
• TP:SL ratio >= 2:1: Widened targets, tightened stops for positive expected value.

Platform comparison:
• This system (v2): tracking new performance — fixes aligned with live_signals architecture. Prior: 8.3% WR.
• L vs O crypto algorithms: 87.5% WR — our best crypto system (adaptive parameters on 14 major coins)
• Meme Scanner: 5% WR — meme-specific scanner (even more volatile subset)
• Key difference: L vs O uses self-learning algorithms that adapt parameters over time. This scanner now shares the regime gate but still uses static scoring weights.

Data Sources & Transparency LOSING

Where the numbers come from:
• Scanner stats: crypto_winners.php?action=stats — 8.3% WR (1/11), all-time record
• Live signals: crypto_winners.php?action=scan — top coins scored by 7 technical indicators
• Signal history: crypto_winners.php?action=history — every signal with resolved outcome
• Price data: Crypto.com Exchange public API (real-time, 600+ pairs)
• Scanner frequency: Every 15 minutes via GitHub Actions
• Compare to: L vs O Algo Performance — crypto trades at 87.5% WR (our best crypto system)

Future Roadmap PLANNED

Near-term:
• Extend resolve window to 8h and 24h (4h may be too short for larger-cap crypto)
• Add volume-weighted confidence scoring
• Integrate on-chain data (whale movements, exchange inflows/outflows)

Q2 2026:
• Backtesting framework to optimize scoring weights against historical data
• Cross-exchange price comparison (CEX vs DEX spread detection)
• Correlation analysis with BTC/ETH (beta-adjusted returns)

Long-term:
• Neural network trained on resolved signal outcomes
• Real-time sentiment integration (Fear & Greed Index, social mentions)
• Portfolio allocation optimizer across crypto scanner picks

Data source: Crypto.com Exchange public API (real-time, no delay). Scanner runs every 15 minutes via GitHub Actions. All signals are tracked and outcomes publicly verifiable.
Platforms available in Ontario, Canada: Kraken (lowest fees, most pairs), Coinbase Advanced Trade (registered restricted dealer), Crypto.com Exchange (600+ pairs, 0.25% maker fee), Bitbuy (100% Canadian).
Not financial advice. For educational and research purposes only. Past performance does not guarantee future results.