# Canonical: https://www.uddalak.xyz/hyperliquid-academy/article/101-robust-price-indices-bad-manipulated-prices
# Author: Uddalak Das
# Last source: https://www.uddalak.xyz/hyperliquid-academy/article/101-robust-price-indices-bad-manipulated-prices

---
title: "How Robust Price Indices Handle Bad or Manipulated Prices"
description: "Hyperliquid's robust price design uses multiple source and candidate prices so one bad exchange print, local trade, or model input is less likely to control the risk engine."
pillar: "Markets & Trading"
track: "Markets & Trading"
cluster: "Markets & Trading"
level: "Advanced"
primary_icp: ["trader", "market maker"]
secondary_icp: ["institutional decision-maker"]
content_type: "risk"
reading_time: "3 min"
last_reviewed: "2026-09-07"
last_verified: "2026-09-07"
volatile: true
prerequisites: ["99-how-hyperliquid-calculates-mark-prices", "100-how-oracle-prices-feed-hyperliquid-risk-engine"]
related: ["108-extreme-volatility-on-hyperliquid", "37-adl-platform-solvency"]
---

# How Robust Price Indices Handle Bad or Manipulated Prices

Price manipulation defense is an aggregation problem. If a risk engine trusts one source, corrupting that source can corrupt liquidations. Hyperliquid reduces this dependency by combining validator oracle submissions and multiple mark-price candidates, then using median-style selection. Robustness means one outlier is less powerful, not that pricing can never fail.

## Source-level robustness

An oracle can aggregate several external spot venues rather than trust one exchange. Weighted-median logic is less sensitive to one extreme observation than a simple average.

Validator aggregation adds another layer: the chain consumes a consensus process around reference submissions rather than a single centralized feed.

## Mark-level robustness

Even a correct oracle can lag a fast derivatives market. Hyperliquid therefore combines oracle and market-derived candidates for mark price. A median among candidates reduces the influence of a single distorted local or external component.

## What if the whole market is broken?

No statistical method can infer a uniquely correct fair price when the underlying asset itself is undergoing a halt, depeg, exploit, migration, or fragmented price discovery. At that point, market operations and explicit risk decisions matter.

Robustness buys resistance to isolated bad data. It does not create truth from a market with no reliable reference.

## How traders should monitor it

Track oracle-versus-mid and mark-versus-oracle deviation. Sudden persistent divergence can indicate stress, thin local liquidity, or source-market disruption.

For large leverage, treat unusual reference divergence as a reason to reduce size rather than wait for the liquidation engine to demonstrate which price wins.

## How this connects

Start with [How Hyperliquid Calculates Mark Prices](/hyperliquid-academy/article/99-how-hyperliquid-calculates-mark-prices), [How Oracle Prices Feed Hyperliquid's Risk Engine](/hyperliquid-academy/article/100-how-oracle-prices-feed-hyperliquid-risk-engine) if the underlying concept is unfamiliar. Related reading: [What Happens to Hyperliquid Markets During Extreme Volatility?](/hyperliquid-academy/article/108-extreme-volatility-on-hyperliquid), [Auto-Deleveraging and Platform Solvency](/hyperliquid-academy/article/37-adl-platform-solvency).

## Sources

- https://hyperliquid.gitbook.io/hyperliquid-docs/trading/robust-price-indices