In the past three weeks, I have replaced the team's derivatives routing with a @Pyth Network pull oracle. The most intuitive change is that the workflow of 'update first, then match' has eliminated many gray areas: in the same transaction, the price is written on-chain first, then read for execution, which has relieved both users and risk control. Previously, we were led by the heartbeat price, either lagging behind or leaving disputes with old prices; now, turning 'when to use the new price' into a business parameter, this 'timing sovereignty' brings more interpretability than just being 'faster'.
Cost and Receipts: Update on demand who uses who pays, no longer burdened by fixed heartbeat fees across multiple chains. We reuse the most lightweight requests with the latest 'Qualified Price', only triggering updates when price age or confidence exceeds the threshold, significantly converging the fluctuations of settlement costs.
Disputes and Rollbacks: Write 'Update Failure = Overall Rollback' into the critical path, rejecting the compromise of 'Complete Transaction then Supplement Price'. After two weeks of comparison, complaints about slippage/old prices in the ticketing system have visibly decreased.
$PYTH
The three handles of risk control are indeed useful.
Price Age: Key Operations (Open Position, Borrowing, Liquidation, Redemption) Forced Age ≤ N seconds, otherwise must be updated first.
Confidence/Price Ratio: When divergence increases, limit opening new positions, only allowing reduction or increasing margin.
EMA Reference Price: Extreme volatility used as a 'conservative boundary' for liquidation discount and collateral estimation, normal transactions still read instant price.
These three handles combined are our own 'Market Integrity Wall'. It doesn't block trading, only dirty trading.#PythRoadmap
Cross-Chain Consistency and Monitoring
After multi-chain deployment, we created a delay panel and consistent liquidation window for each chain: if a specific chain experiences abnormal tail delays, it temporarily becomes read-only or only reduces positions, avoiding 'Inter-Chain Time Arbitrage'. This set of SRE scripts (Forced Update Failure → Rollback, Abnormal Chain → Read-Only) has saved us during two severe fluctuations.
Bonus Points for Supporting Capabilities
Express Relay: We have integrated some liquidation tasks into priority auctions, making it easier to match clean liquidity during extreme moments, and more controllable liquidation discounts.
Entropy Random Number: Verifiable random numbers for point lottery/task distribution, the front end has implemented a 'Request-Result Availability' state machine, automatically retrying on failure, providing a smooth user experience.
Workshop Practical: The details of 'Same Order Update + Consumption' asked in the ETHGlobal New Delhi integration workshop, we copied back, saving a lot of potential pitfalls.@Pyth Network
Real-time news inspires business boundaries.
Pyth Pro has turned 'Institution-Level Data' into a subscription model, selling freshness and coverage like Spotify sells songs; institutions like Jump have already joined the early experience, which is a direct signal for us to explore the 'Freshness Package (Basic/Professional/Institutional)'.
Blue Ocean ATS has joined the network, providing 24/5 post-market pricing for US stocks; our US stock-related products finally have 'source trustworthy' late-night prices, allowing us to tighten the price age threshold instead of a one-size-fits-all closure.
Hyperliquid's native stablecoin USDH also has real-time price feeds, crucial for cross-domain collateral and index basket rebalancing.#PythRoadmap
Next Steps for 30–60–90 Days
30 Days: Turn 'Freshness Package' into a billing tier (Age/Confidence Threshold/SLA), transparently display 'Whether to Force Update, Refusal Reason' on the front end.
60 Days: Batch multi-asset updates to reduce costs, improve multi-chain consistency window; liquidation robots default to carrying updates.
90 Days: Implement governance on thresholds and new categories on-chain, publish 'Public Transaction and Refusal Rate Weekly Report' monthly, turning auditability into a growth asset.
Tokens and Governance
PYTH serves as both a certificate for participating in governance and a link connecting 'Data Subscription → Real Income → Community Distribution'. In the long run, whoever turns the auditable fairness into a commodity can retain better market-making and institutional traffic.@Pyth Network
