Building Oracle Gym - A Gamified Crypto Trading Arena with Pyth Network

Koushik Ghosh — Mon, 30 Mar 2026 07:31:22 +0000

Ever wondered if you'd survive a crypto flash crash? What if you could practice trading with real market data without risking actual money?

That's why I built Oracle Gym - a gamified trading platform powered by Pyth Network that lets you replay actual market conditions.

The Problem

Most paper trading platforms use fake data. When you practice on simulated prices, you're not learning how real markets behave.

Oracle Gym solves this by using real Pyth Network price feeds to create authentic training scenarios.

What is Oracle Gym?

Oracle Gym is a gamified crypto trading trainer where you:

Replay real markets - Choose from 13 crypto assets with actual Pyth price data
Face random shocks - Powered by Pyth Entropy V2 smart contract
Make decisions - Buy, sell, hold, reduce, wait, or hedge at checkpoints
Get scored - AI coaching analyzes your performance
Level up - Earn XP, build streaks, compete on leaderboards

Think of it as a flight simulator for crypto traders.

Tech Stack

Built with:

Frontend: Next.js 15, TypeScript, Tailwind CSS v4
Blockchain: Wagmi v3, Viem v2, Base Sepolia
Pyth: Hermes API (real-time + historical), Entropy V2 (on-chain randomness)
Wallets: 300+ supported (MetaMask, WalletConnect, Coinbase Wallet)

How It Works

1. Price Feed Integration

Oracle Gym uses Pyth's Hermes API for real-time and historical price data. Each feed includes:

price - The actual price value
expo - Exponent for decimal handling
publish_time - Timestamp verification
conf - Confidence intervals

2. Entropy V2 for Random Shocks

The "boss fight" feature uses Pyth Entropy V2 to generate verifiable on-chain randomness. This creates unpredictable market shocks that are:

Unpredictable - No one can predict the outcome
Verifiable - Anyone can verify on-chain
Fair - Cannot be manipulated

Smart contract deployed at: 0x148123bc5b719a7e169ee652a72be387c964b6f4 on Base Sepolia

3. Scenario Types

Six different scenarios using real Pyth historical data:

Breakout - Trend compression snaps into expansion
Crash - Violent downside unwind
Chop - Whipsaw price action
Fakeout - Momentum looks clean until the floor disappears
Slow Bleed - Long, draining descent
Volatility Spike - Wide candles and sharp reversals

The Gamification System

XP & Leveling

Players earn 500 XP per level with unlimited progression.

7-Day Streak System

Build streaks by claiming daily rewards with multipliers from ×2 to ×14:

Day	Base Reward	Multiplier	Total Potential
1	100 coins	×2	200 coins
2	150 coins	×4	600 coins
3	200 coins	×6	1,200 coins
4	250 coins	×8	2,000 coins
5	300 coins	×10	3,000 coins
6	350 coins	×12	4,200 coins
7	500 coins	×14	7,000 coins

The multiplier also applies to session earnings.

Supported Assets

13 major cryptocurrencies using official Pyth price feeds:

BTC, ETH, SOL, BNB, XRP, ADA, DOGE, AVAX, POL, LINK, UNI, ATOM, PYTH

Key Challenges Solved

Challenge 1: Hydration Errors

Problem: Server-rendered HTML didn't match client-rendered HTML for wallet-dependent UI.

Solution: Added mounted state to only render wallet UI after client-side hydration.

const [mounted, setMounted] = useState(false);

useEffect(() => {
  setMounted(true);
}, []);

// Only render wallet-dependent UI after mount
{mounted && isConnected && (
  <div suppressHydrationWarning>
    {balance} coins
  </div>
)}

Challenge 2: Entropy Callback Timing

Problem: Entropy V2 callbacks take 1-2 blocks, creating UX uncertainty.

Solution: Implemented polling with status tracking to show real-time progress.

Challenge 3: Rate Limiting

Problem: Free Pyth API has rate limits.

Solution: Upgraded to Pyth Pro for 1,000+ requests per minute.

What I Learned

Pyth Network is Production-Ready

The Hermes API is fast, reliable, and well-documented. Pyth Pro's enhanced rate limits made development smooth.

Entropy V2 is a Game-Changer

Having verifiable on-chain randomness opens up possibilities for fair gameplay mechanics that weren't possible before.

Gamification Works

Adding XP, levels, and streaks transformed a simple replay tool into an engaging training platform.

Results

After building Oracle Gym:

✅ 13 Pyth price feeds integrated
✅ Entropy V2 smart contract deployed and working
✅ 300+ wallet support via Wagmi v3
✅ Sub-second page loads with Next.js 15
🎮 Average session: 8-12 minutes
📈 Most popular assets: BTC (45%), ETH (30%), SOL (15%)

Try It Yourself

Want to test your trading skills?

orcale-gym.vercel.app

Connect your wallet, get 1,000 free coins, and start training!

What's Next

Future improvements planned:

Multiplayer mode - Compete against other traders in real-time
More assets - Add forex, commodities, and stocks
Custom scenarios - Let users create and share scenarios
Mobile app - React Native version
Tournament system - Weekly competitions with prizes

Conclusion

Building Oracle Gym taught me that real data makes all the difference. When you practice with actual market conditions, you learn patterns that matter.

Pyth Network's price feeds and Entropy V2 made it possible to create a training platform that feels authentic. The combination of real data, verifiable randomness, and gamification creates an engaging learning experience.

If you're building anything that needs reliable oracle data or on-chain randomness, check out Pyth Network.

Built for the Pyth Community Hackathon 2026

What would you build with Pyth Network? Drop a comment below! 👇

Forem: Koushik Ghosh