Polymarket FastLoop Trader

Trade Polymarket's 5-minute BTC fast markets using real-time price momentum from Binance.

Polymarket only. All trades execute on Polymarket with real USDC. Use --live for real trades, dry-run is the default.

How it works: Every cycle, the script finds the current live BTC fast market, checks BTC price momentum on Binance, and trades if momentum diverges from market odds.

This is a template. The default signal (Binance momentum) gets you started. Your agent's reasoning is the edge — layer on sentiment analysis, multi-exchange spreads, news feeds, or custom signals to improve it.

⚠️ Fast markets carry Polymarket's 10% fee (is_paid: true). Factor this into your edge calculations.

When to Use This Skill

Use this skill when the user wants to:

• Trade BTC sprint/fast markets (5-minute or 15-minute)
• Automate short-term crypto prediction trading
• Use CEX price momentum as a Polymarket signal
• Monitor sprint market positions

Setup Flow

When user asks to install or configure this skill:

1. Ask for Simmer API key

   • Get from simmer.markets/dashboard → SDK tab
   • Store in environment as SIMMER_API_KEY

2. Ask about settings (or confirm defaults)

   • Asset: BTC, ETH, or SOL (default BTC)
   • Entry threshold: Min divergence to trade (default 5¢)
   • Max position: Amount per trade (default $5.00)
   • Window: 5m or 15m (default 5m)

3. Set up cron or loop (user drives scheduling — see "How to Run on a Loop")

Quick Start

# Set your API key
export SIMMER_API_KEY="your-key-here"

# Dry run — see what would happen
python fastloop_trader.py

# Go live
python fastloop_trader.py --live

# Live + quiet (for cron/heartbeat loops)
python fastloop_trader.py --live --quiet

# Live + smart sizing (5% of balance per trade)
python fastloop_trader.py --live --smart-sizing --quiet

How to Run on a Loop

The script runs one cycle — your bot drives the loop. Set up a cron job or heartbeat:

Every 5 minutes (one per fast market window):

*/5 * * * * cd /path/to/skill && python fastloop_trader.py --live --quiet

Every 1 minute (more aggressive, catches mid-window opportunities):

* * * * * cd /path/to/skill && python fastloop_trader.py --live --quiet

Via OpenClaw heartbeat: Add to your HEARTBEAT.md:

Run: cd /path/to/fast market && python fastloop_trader.py --live --quiet

Configuration

Configure via config.json, environment variables, or --set:

# Change entry threshold
python fastloop_trader.py --set entry_threshold=0.08

# Trade ETH instead of BTC
python fastloop_trader.py --set asset=ETH

# Multiple settings
python fastloop_trader.py --set min_momentum_pct=0.3 --set max_position=10

Settings

Example config.json

{
  "entry_threshold": 0.08,
  "min_momentum_pct": 0.3,
  "max_position": 10.0,
  "asset": "BTC",
  "window": "5m",
  "signal_source": "binance"
}

CLI Options

python fastloop_trader.py                    # Dry run
python fastloop_trader.py --live             # Real trades
python fastloop_trader.py --live --quiet     # Silent except trades/errors
python fastloop_trader.py --smart-sizing     # Portfolio-based sizing
python fastloop_trader.py --positions        # Show open fast market positions
python fastloop_trader.py --config           # Show current config
python fastloop_trader.py --set KEY=VALUE    # Update config

Signal Logic

Default signal (Binance momentum):

1. Fetch last 5 one-minute candles from Binance (BTCUSDT)
2. Calculate momentum: (price_now - price_5min_ago) / price_5min_ago
3. Compare momentum direction to current Polymarket odds
4. Trade when:
   • Momentum ≥ min_momentum_pct (default 0.5%)
   • Price diverges from 50¢ by ≥ entry_threshold (default 5¢)
   • Volume ratio > 0.5x average (filters out thin moves)

Example: BTC up 0.8% in last 5 min, but fast market YES price is only $0.52. The 3¢ divergence from the expected ~$0.55 → buy YES.

Customizing Your Signal

The default momentum signal is a starting point. To add your own edge:

• Multi-exchange: Compare prices across Binance, Kraken, Bitfinex — divergence between exchanges can predict CLOB direction
• Sentiment: Layer in Twitter/social signals — a viral tweet can move fast markets
• Technical indicators: RSI, VWAP, order flow analysis
• News: Breaking news correlation — use your agent's reasoning to interpret headlines

The skill handles all the Simmer plumbing (discovery, import, trade execution). Your agent provides the alpha.

Example Output

⚡ Simmer FastLoop Trading Skill
==================================================

  [DRY RUN] No trades will be executed. Use --live to enable trading.

⚙️  Configuration:
  Asset:            BTC
  Entry threshold:  0.05 (min divergence from 50¢)
  Min momentum:     0.5% (min price move)
  Max position:     $5.00
  Signal source:    binance
  Lookback:         5 minutes
  Min time left:    60s
  Volume weighting: ✓

🔍 Discovering BTC fast markets...
  Found 3 active fast markets

🎯 Selected: Bitcoin Up or Down - February 15, 5:30AM-5:35AM ET
  Expires in: 185s
  Current YES price: $0.480

📈 Fetching BTC price signal (binance)...
  Price: $97,234.50 (was $96,812.30)
  Momentum: +0.436%
  Direction: up
  Volume ratio: 1.45x avg

🧠 Analyzing...
  ⏸️  Momentum 0.436% < minimum 0.500% — skip

📊 Summary: No trade (momentum too weak: 0.436%)

Source Tagging

All trades are tagged with source: "sdk:fastloop". This means:

• Portfolio shows breakdown by strategy
• Other skills won't interfere with your fast market positions
• You can track fast market P&L separately

Troubleshooting

"No active fast markets found"

• Fast markets may not be running (off-hours, weekends)
• Check Polymarket directly for active BTC fast markets

"No fast markets with >60s remaining"

• Current window is about to expire, next one isn't live yet
• Reduce min_time_remaining if you want to trade closer to expiry

"Import failed: Rate limit exceeded"

• Free tier: 10 imports/day. Pro: 50/day
• Fast market trading needs Pro for reasonable frequency

"Failed to fetch price data"

• Binance API may be down or rate limited
• Try --set signal_source=coingecko as fallback

"Trade failed: no liquidity"

• Fast market has thin book, try smaller position size