Quant

Abstract

Quant is a browser extension that turns market noise into a short list of backtested, ready-to-execute trade setups. Each setup includes an Entry, Stop-Loss, and Take-Profit, scored by expected value and validated by rolling backtests across approximately 150 Hyperliquid perpetual markets. Users review, edit, and sign every order. Quant does not trade for you. It narrows chaos into a small set of high-confidence decisions and leaves execution firmly in human hands.

1. Problem

Retail perpetual traders face a compounding set of problems that grow worse as markets get faster:

Quant addresses these problems with a single interface layer that sits on top of the websites traders already use.

2. Solution

Quant is a browser extension for Chrome and Chromium-based browsers (Brave, Edge, and others on Windows and macOS). Once installed, it overlays a small interactive avatar on every page. Clicking the avatar opens a panel with a curated list of trade setups for the coin and timeframe the trader is viewing.

Core principles

What Quant is not

Quant is not copy-trading (you are not mirroring another account), not auto-trading (there is no "set and forget"), and not chat-to-trade (prompting a chatbot is not how most traders want to act). It is a decision-support layer, not a decision-maker.

3. Who Quant Is For

Quant is built for wallet-native retail traders who discover ideas on social platforms, check data on aggregators like CoinMarketCap or CoinGecko, and want structured entries, stops, and take-profits without spending all day on charts.

4. Trade Setups

A setup is a trade plan that has earned the right to be shown. It is generated by a tested strategy, scored by its backtest performance, and filtered by risk/reward guardrails before reaching the user.

What a setup contains

• Entry price. The price at which the strategy triggered.
• Stop-loss. The invalidation level.
• Take-profit. The target level.
• Confidence score. Derived from expected value per trade on recent backtest data. Display thresholds: green (excellent), blue (good), yellow (moderate), red (low).
• Rationale. A plain-language explanation of why the strategy triggered.
• Position side. Long or short.

The trader can edit any level or adjust position size before execution. When two strategies disagree on direction for the same coin and timeframe, both setups are shown so the trader can evaluate each independently. A long-vs-short balance bar indicates the current directional tilt.

Setup validity

Setups do not live forever. Two rules govern expiration:

Time-to-live. Each setup has a shelf life matched to its timeframe:

Maximum entry drift. If the current price has moved too far from the setup's entry, the setup is hidden. Drift thresholds scale with timeframe:

Newer setups also receive a freshness bonus in ranking, ensuring the list favours the most current opportunities.

5. Strategy Engine

Quant runs approximately 50 technical-analysis strategies, roughly evenly split between long and short, across approximately 150 Hyperliquid perpetual markets and seven timeframes (1m, 5m, 15m, 30m, 1h, 4h, 1d). The library is designed to scale; new strategies can be added at any time without downtime.

How it works

1. Data ingestion. Price and volume data streams continuously from Hyperliquid.
2. Indicator computation. Each strategy is composed of conditions built from standard indicators (moving averages, RSI, ADX/DMI, stochastics, and others).
3. Signal evaluation. When entry conditions are met, a setup is created with computed Entry, Stop-Loss, and Take-Profit levels.
4. Refresh cadence. Signals are re-evaluated on a cycle matched to their timeframe. A 5-minute strategy refreshes far more often than a daily one.
5. Scoring. Rolling backtests run continuously. The confidence score you see is the expected value per trade from the most recent backtest window.
6. Filtering. Setups that fail EV, TTL, or drift checks are removed before they reach the user.

The aim is not to predict the future. It is to compress the decision space so traders spend attention only where the odds and risk/reward look reasonable.

Expected Value

Quant's goal is to surface only setups with positive Expected Value (EV).

• WinRate: the share of trades that hit take-profit before stop-loss over the backtest window.
• LossRate: the share that hit stop-loss first (1 minus WinRate).
• AvgWin / AvgLoss: average size of wins and losses measured in R, where 1R equals the distance from entry to stop-loss.
• Fees/Slippage: the average all-in cost per trade (venue fees, funding, typical slippage).

You do not need a 50% win rate. With a 2:1 risk-to-reward ratio, break-even is roughly 33%. Above that, after costs, EV turns positive. Backtests use the maximum historical depth available from Hyperliquid and re-run on a rolling basis to keep scores current.

6. Contextual Information

Every setup includes a rationale explaining why the strategy triggered and what the backtest data shows. This helps traders quickly decide whether a setup fits their thesis without interpreting raw indicator values. A richer information layer is in development. Planned capabilities include:

• Trade explainers. Deeper context behind each setup's trigger and the current market environment.
• Macro and sentiment cues. Relevant external events surfaced alongside setups when they may affect the trade thesis.
• Warnings. Flags when conditions may undermine a setup's assumptions.

This layer will never place, modify, or cancel trades. It provides context, not action.

7. Safety, Custody, and Control

• Non-custodial. Quant cannot move funds without the user's signature.
• Wallet. Created inside the extension via Privy (one wallet per user). Private keys remain on the user's device.
• On-chain signing. When executing on Hyperliquid, the user signs every order. Paper-mode trades are simulated and do not touch the chain.
• Execution safeguards. Slippage buffers, retry with exponential backoff, and system telemetry protect order quality.
• Leverage. Enabled alongside live Hyperliquid execution. Intentionally absent during the paper-only beta so traders can build familiarity without capital risk.
• Paper mode. Available from day one and remains accessible after live execution launches. Traders can practise and validate strategies with simulated funds before committing real capital.

8. Using Quant

1. Install the extension from the Chrome Web Store (compatible with Brave, Edge, and other Chromium browsers on Windows and macOS).
2. Sign in. Quant creates a wallet and displays a small 3D avatar overlay ("Pip") on the page.
3. Open any supported coin page and click the avatar. The panel shows the top setups for your chosen timeframe.
4. Review a setup: check the confidence score, rationale, and risk/reward. Edit size or levels if needed.
5. Execute with one click. Manage positions in the sidebar: active positions, P/L, trade history, and summary statistics.

Execution details

• Routing. All live orders go to Hyperliquid (HyperCore).
• Order composition. One click places an order with attached stop-loss and take-profit.
• Fees. Users pay standard Hyperliquid trading fees plus a 5 basis-point (0.05%) builder fee that goes to Quant. This fee is charged transparently via Hyperliquid's onchain builder-code program. The user approves the maximum builder fee once; it is enforced onchain and can be revoked at any time. There are no subscription fees, hidden costs, or platform withdrawal fees.

About Hyperliquid

Hyperliquid is a fully onchain order-book exchange for perpetual futures. It settles trades on its own L1 chain (HyperCore), offers deep liquidity across 150+ markets, and supports programmatic order routing, which is what Quant uses to deliver one-click execution.

9. Roadmap

10. Business Model

Near-term target (Q2 2026): Serve approximately 1,000 daily active traders, facilitate approximately $10M in routed volume per day (approximately $300M per month), and sustain approximately $150K in monthly recurring revenue at approximately 5 bps.

11. Points and Referrals

Quant includes a points program that rewards active participation. Points accrue through trading activity, feature usage, and community contributions. A multi-tier referral system is planned. Referrers earn points based on referee activity, and higher tiers unlock revenue sharing from referees' trading fees. Full details will be published as the program rolls out.

12. Architecture

Quant is built as a set of lightweight, independently deployable services:

• Extension. A browser-based interface built with React and TypeScript that renders the overlay, sidebar, and trade controls.
• Strategy engine. A cloud-hosted service that evaluates strategies, generates signals, and runs rolling backtests in real time.
• API gateway. Handles authentication, signal delivery, position tracking, and account management.
• Data pipeline. Streams price and volume data from exchanges, computes indicators, and caches results for sub-second response times.

Performance targets: Strategy response under 300ms (p50). UI render under 1 second.

13. Privacy, Security, and Legal

• Non-custodial. Keys stay on the user's device. Quant cannot move funds without explicit signature.
• Audit. A security audit is planned before public rollout. The closed beta runs with guardrails and limited exposure.
• Geography. Quant follows venue rules. If Hyperliquid or any future integrated venue is geo-blocked in a region, Quant does not bypass those restrictions.
• Disclaimer. Quant is not financial advice. Backtested results are simulated; past performance does not guarantee future outcomes. All strategies carry risk, including total loss of capital.

14. Agent Arena

The Agent Arena is a public benchmarking space, separate from Quant's core trading product, where AI-powered agents compete in paper trading to demonstrate their performance over fixed horizons.

Binary Options Agents

Binary Options Agents analyze market data, news sentiment, and technical indicators to forecast whether BTC will go up or down over a 1-hour or 24-hour time window. Each prediction simulates a $500 binary bet at 50/50 odds: correct = +$500, incorrect = -$500, hold = $0. All agents start with $10,000 virtual capital. This model is aligned with directional prediction use cases such as Polymarket up/down BTC markets and Hyperliquid's upcoming HIP-4 on-chain options.

Trading Agents

Trading Agents execute automated trades based on Quant's backtested signals on BTC perpetuals. Users configure strategy selection (from 50+ backtested strategies across 7 timeframes), leverage (1x to 50x), position sizing (fixed USD or % of portfolio), take-profit/stop-loss levels (strategy defaults or custom overrides), daily loss limits, maximum drawdown thresholds, and maximum concurrent positions. All trading agents start with $10,000 virtual capital in paper trading mode. A fee simulation mode displays performance net of Hyperliquid exchange fees (4.5bps taker entry, 1.5bps maker exit) and builder fees (5bps).

How it works

Multiple AI agents, each powered by a different model and data configuration, submit structured predictions or execute signal-based trades. Agents are scored against real outcomes. Equity curves, accuracy rates, and leaderboards are updated after each round. An Agent Builder role lets community members create, configure, and submit their own agents to compete.

Relationship to Quant

Forecasting and execution are kept deliberately separate. Arena outputs are informational and experimental. They do not place real trades or override Quant setups. Over time, top-performing agents and their underlying approaches may inform Quant's roadmap, including the longer-term vision of autonomous trading capabilities. Any such integration will be introduced carefully, with human oversight preserved.

15. Ecosystem

Quant is built within the XBorg ecosystem and benefits from its infrastructure, community, and go-to-market support. Early adoption is supported by partners including SwissBorg.

16. Glossary