A dashboard that watches Polymarket, surfaces interesting markets, and helps you size bets safely with the Kelly Criterion.
It comes in two flavors — pick whichever fits you better:
| Version | Folder | Best for |
|---|---|---|
| Static (browser-only) | docs/ |
Free hosting on GitHub Pages. No backend. Always live. ✨ Recommended. |
| Full-stack (FastAPI + SQLite) | backend/ + frontend/ |
Persistent local price history when running on your own machine. |
The docs/ folder is a completely standalone web app. All Polymarket data is fetched directly from the browser using their public CORS-enabled API.
cd ~/Desktop/polymarket-analyzer
# Initialise a git repo and push to GitHub
git init
git add .
git commit -m "initial commit"
gh repo create polymarket-analyzer --public --source=. --pushThen in your repo's settings:
- Open Settings → Pages
- Source: Deploy from a branch
- Branch: main, Folder: /docs
- Save.
Your dashboard will be live at:
https://<your-github-username>.github.io/polymarket-analyzer/
That's it. The page auto-refreshes every 5 minutes and remembers the last snapshot via localStorage.
- ✅ Fetches up to 500 active markets from
gamma-api.polymarket.com - ✅ Per-market 7-day price history chart from
clob.polymarket.com - ✅ Edge scoring, signal detection, category classification (all in-browser JS)
- ✅ Full Kelly Criterion calculator
- ✅ Search, sort, filter by category and minimum volume
If you want persistent price history (saved to SQLite over weeks), run the FastAPI backend:
cd ~/Desktop/polymarket-analyzer
./start.sh
# Dashboard → http://localhost:8000This version polls Polymarket every 5 minutes and accumulates a per-market price history in backend/polymarket.db.
The tool helps you avoid bad bets, not pick winners. There's no magic AI here — Polymarket markets are mostly efficient, and "predicting" them requires information edge.
-
Pick a category you actually understand. Don't bet on Korean politics if you only read Western news. Edge comes from knowing something the market doesn't.
-
Filter for inefficient markets: high liquidity (over $50K), 3–60 days to resolution, prices in the 15–85% range (not crowded extremes).
-
For each interesting market, write down your probability estimate before looking at the market price. This is the single most important habit. If you read the market first, you'll anchor to it.
-
Enter your estimate in the Kelly calculator. If your estimate differs from the market by less than 5 percentage points, skip the bet — the spread and your own uncertainty will eat the edge.
-
Bet Half Kelly, not Full Kelly. Full Kelly maximizes long-run growth if your probabilities are exact. They never are. Half Kelly survives bad estimates.
-
Never bet more than 2% of your bankroll on a single market, even if Kelly suggests more.
-
Track your calibration. Keep a spreadsheet: market, your estimate, market price, your bet, outcome. After 50 bets, see if your "70% confident" predictions actually win ~70% of the time. If not, you're miscalibrated — stop until you fix it.
| Signal | What it means |
|---|---|
| 🔥 Volume Surge | Recent 24h volume is large relative to liquidity. Something is happening — be cautious, you might be late to the story. |
| ✅❌ Near-Certain YES/NO | Price > 93% or < 7%. Hard to find edge here; tiny moves require huge stakes. |
| 🪙 Coin Flip | Price 44–56%. Most uncertain markets — also most competitive. |
| ⚠ Low Liquidity | Pool < $1K. Wide spreads will eat any edge. Avoid. |
| 💧 Deep Market | Pool > $100K. Tight spreads, but harder to beat — many sharp traders. |
| ⏰ Resolving Soon | Under 2 days. Volatile and emotional — usually a bad time to enter. |
- This is not financial advice. Prediction markets are gambling. Only bet what you can afford to lose entirely.
- Polymarket is not available in the US — verify legality in your jurisdiction.
- The "edge score" is a heuristic for how interesting a market is to analyze, NOT a prediction of direction.
- Categories are inferred from question keywords — they're approximate.