⚔️ Duel

@polybacktest Small-body edge on direction is noisy from macro. On a pair candle (BTC/ETH spread) shared macro cancels, body-size filter should look cleaner. Duel has 5-min BTC/ETH and ETH/SOL markets live, tick data available if useful: duel.market

56.4% win rate with one rule: ignore the big candles. Trade the small ones. 1,931 BTC 5-minute markets on Polymarket. Binance 1-second OHLC data. We built the pre-market 1-minute candle and tested every pattern. First, the classics: Hammer (long lower wick, bet UP): 52.9%. 104 trades. Shooting star (bet DOWN): 44.1%. 118 trades. Strong bull candle (bet UP): 48.9%. 440 trades. Strong bear candle (bet DOWN): 50.3%. 453 trades. Doji (bet reversal): 44.6%. 240 trades. All dead. Not a single classic pattern breaks even. Then we bucketed by body size. Tiny body (0-10% of range): 54.6% continuation. 130 trades. Small body (10-20%): 56.4% continuation. 110 trades. Medium body (40-60%): 47.8%. Big body (80-100%): 46.0%. 567 trades. The weaker the candle, the more it follows through. The "strong conviction" full-body candle? 46%. Below coin flip. Big moves overshoot and reverse. Small moves are the ones that keep going. Does this hold on ETH and SOL markets, or is it BTC-specific? Worth checking. You can backtest your own strategies with polybacktest. It has sub-second historical data going back a month

What's next: → Stocks: AAPL vs MSFT, TSLA vs NVDA → Stocks vs crypto: BTC vs NVDA → Commodities: gold vs silver, oil vs gas → Sports: xG soccer, basketball, more. Outperformance is the signal. Correlation is the engine. cc @jessepollak @jessexbt_ai @buildonbase

Infrastructure is open → API: quote any pair, pricing is public → Docs: duel.market/api-docs.html → Feedback or a new market you want to see? support@duel.market Looking for market makers, builders, and traders with ideas.

Markets, head-to-head ⚔️ Duel is the first prediction market for outperformance. Now live on @base Pick which asset wins over a fixed time window. Binary contracts that settle at $1 if your side outperforms, $0 if not. duel.market 🧵