Frequently asked questions.

AlphaNet is a quantitative AI futures trading platform that brings strategy research, automated trading, risk controls and algorithmic execution into one system. Its purpose is to make a systematic trading stack accessible through a wallet-connected interface.

An AI trading bot and a quantitative trading system are not mutually exclusive categories. A bot is software that automates activity; a quantitative system describes the research and decision process behind that activity. A bot can be the execution component of a sophisticated quantitative strategy.

No — not directly, and not well as a predictor even indirectly. A language model generates text. It doesn’t observe live prices, doesn’t hold a position, and doesn’t place orders. Anything that lets ChatGPT “trade” is a wrapper that turns its text into API calls, and the quality of the trades is the quality of the text.

Some of it is. Most of what’s marketed under the name isn’t — either the AI is a moving-average rule with a new label, or the returns exist only in a backtest, or both.

Quant trading strategies are systematic approaches to trading where every decision comes from a model or rule that’s been specified and tested in advance. They fall into a handful of families, each exploiting a different regularity in how markets behave.

A quant fund is an investment fund whose trading decisions are made by mathematical models. Researchers build the models, systems execute them, and the portfolio is the aggregate of thousands of small, statistically-driven positions.

A quant is someone who uses mathematics, statistics and computation to make decisions in financial markets. Instead of forming a view and trading it, a quant builds a model, tests it against data, and lets the model decide.

Agentic trading is trading carried out by an AI system that is given a goal and works toward it autonomously — deciding when to act, how much to trade, and when to stop, rather than executing a fixed instruction.

AI crypto trading uses machine-learning methods within a cryptocurrency trading process. Models may estimate returns, classify market conditions, assess volatility or help execute orders. The phrase describes the use of a technology, not a particular level of profitability.

An AI hedge fund is a fund whose trading decisions come primarily from machine-learning models rather than from human judgement or from hand-built statistical rules. The models learn patterns from data instead of being told what to look for.

An AI trading agent is a system that observes market conditions, decides what to do, and executes trades — repeatedly, over time, in pursuit of an objective a person has given it. Agentic trading is the approach; the agent is the thing that does it.

An automated trading system places and manages trades without a person clicking. It watches the market, applies a decision process, and sends orders when its conditions are met. The person’s role is to build it, monitor it, and switch it off.

Quantitative trading uses data, mathematical models and explicit decision rules to evaluate market opportunities. It turns a trading idea into a process that can be tested, measured and applied consistently. A strategy may be simple or complex; using AI is not a requirement.

Systematic trading makes decisions by a defined process — rules or models specified in advance — rather than by a person’s judgement in the moment. Every entry, exit and position size can be traced to the process. Nothing depends on how the trader felt that morning.

Leverage trading creates market exposure that is larger than the capital supporting a position. In crypto derivatives, collateral serves as a financial buffer while gains and losses depend on the position’s notional size. Leverage amplifies losses as well as gains.

Cross margin and isolated margin describe how collateral supports leveraged positions. With cross margin, eligible positions share a pool of collateral. With isolated margin, a particular collateral balance is assigned to an individual position.

[Interactive widget: Entry price · Leverage · Long/Short · Maintenance margin rate → Liquidation price and distance from entry]

Last price is the price of the most recent trade on the venue. Mark price is an estimate of fair value derived from spot prices on external exchanges. They’re usually close. When they diverge, it matters which one your venue uses for what.

Spot trading and perpetual futures provide different forms of market exposure. A cash-funded spot purchase acquires the asset itself. A perpetual futures position creates exposure through a derivative contract without a fixed expiry date. The contract’s collateral and settlement rules define the position.

Liquidation is a venue’s process for closing or reducing a leveraged position when the supporting account or position no longer meets its maintenance-margin requirement. It is a risk-management mechanism for the trading venue, not a guarantee that the trader’s losses will be small.

A funding rate is the periodic payment exchanged between long and short holders of a perpetual future. It exists to keep the perpetual’s price close to spot, since there’s no expiry to force convergence.

A perp DEX is a decentralized exchange or trading venue for perpetual futures. “Perps” are derivative contracts without a fixed expiry date. Their design typically uses margin and a funding mechanism to keep contract prices linked to an underlying reference market.

A perpetual future — a perp, or perpetual swap — is a derivative contract that tracks an asset’s price and never expires. You can hold it for a minute or a year. Unlike a traditional future there’s no settlement date, no rollover, and no convergence to spot at expiry.

Funding rate arbitrage — the crypto carry trade — buys an asset on spot and shorts an equal amount on the perpetual. Price exposure nets to zero. What’s left is the funding payment, which flows to the short whenever the perp trades above spot.

Margin trading is borrowing funds to take a position larger than your capital. Deposit $1,000, borrow $4,000, trade $5,000. Your gain or loss is on the full $5,000; your capital at risk is the $1,000.

Open interest is the total number — or total notional value — of derivative contracts currently open. Every perpetual position that hasn’t been closed counts. It measures how much money is committed to the market right now.

A non-custodial crypto trading platform is designed so that users retain control of their wallet keys rather than handing those keys to the platform. That does not mean every asset remains untouched in the wallet while trading takes place.

AlphaNet documents separate controls for stopping strategies, changing allocations and withdrawing funds. These preserve user choice but do not guarantee immediate completion at a particular price.

AlphaNet describes its trading model as non-custodial. The intended distinction is that users retain control of their wallets rather than giving AlphaNet their private keys. Trading capital can nevertheless be deposited into the infrastructure supporting a strategy.

Wallet connection lets AlphaNet identify the account interacting with the platform. The documented workflow separates connecting a wallet from enabling trading and depositing funds; connecting alone should not be confused with deploying a strategy.

A non-custodial exchange lets you trade while your assets stay in a wallet you control. There’s no deposit into an exchange account, no balance held on your behalf, and no withdrawal process controlled by a company. You sign each trade; settlement happens on-chain.

A wallet drainer is malicious code that gets you to sign a transaction granting an attacker the ability to move your assets. You don’t hand over your seed phrase. You approve something that looks routine, and the approval is the theft.

Self-custody means you hold the private keys that control your assets. No exchange holds them for you, no company can freeze them, and no one can move them without your signature.

Autopilot is AlphaNet’s mode for running a user-selected quantitative strategy automatically. The user chooses the strategy and its capital allocation; the strategy then applies its own trading process. The documented controls also allow allocation changes and stopping.

Some crypto trading bots can be profitable over particular periods, while others lose money. Automation alone creates no guaranteed edge. The outcome depends on the strategy, the market environment, the risk taken and the full cost of trading.

Backtesting applies a strategy’s rules to historical data and records the trades it would have made. The output is a simulated equity curve and the statistics that describe it.

AlphaNet seeks to reduce trading slippage through execution models informed by liquidity and market conditions. Its Hackworth V3 materials describe Dynamic TWAP improvements aimed at execution quality. The goal is cost control, not a promise of zero slippage.

Both are execution algorithms that split a large order into smaller pieces to reduce market impact. They differ in how they schedule the pieces.

Trading execution algorithms manage how an order is submitted and filled. Their task can include dividing an order, choosing when to submit each part and adjusting to available liquidity. They are different from the research that decides whether the trade should exist.

A DCA bot automates dollar-cost averaging: buying a fixed amount at regular intervals, or buying an additional tranche each time price falls a set percentage. The average entry price improves as price drops, and the timing decision disappears.

A market regime is a persistent state of the market that determines how prices behave and which strategies work. Trending or ranging. High volatility or low. Risk-on or risk-off. A strategy tuned for one regime typically loses in another.

Mean reversion bets that a price which has moved sharply away from its recent average will move back toward it. Buy the dip, sell the spike, on the assumption that the extreme was an overreaction.

Momentum trading buys assets that have been rising and sells ones that have been falling, on the evidence that recent performance tends to persist over the next weeks or months.

Trend following enters a position once a trend is established and holds until the trend shows signs of ending. It doesn’t predict; it reacts. The core assumption is that markets move in persistent directions more often than a random walk would suggest.

VWAP — volume-weighted average price — is the average price an asset traded at over a period, weighted by how much volume traded at each price. “VWAP strategy” means two different things depending on who’s saying it.

Algorithmic trading in crypto means using software to apply defined trading logic. The logic can govern an entire strategy or only one part of the workflow, such as converting an already-approved trade into smaller orders. Algorithmic trading does not necessarily use AI.

Alpha decay is the erosion of a strategy’s excess return over time. An edge that earned 2% a month last year earns 1% this year and nothing next year. The signal that predicted price no longer does.

Grid trading places a ladder of buy orders below the current price and sell orders above it, at fixed intervals. Each time price moves up through a level, a sell fills; each time it moves down, a buy fills. Every up-and-down oscillation completes a pair and captures the gap.

A market maker posts both a bid and an ask, continuously, and earns the difference when both sides fill. Buy at $99.98, sell at $100.02, pocket four cents. Do it thousands of times a day.

Order flow trading is the practice of reading executed trades and resting orders to infer what large participants are doing, and positioning accordingly. Instead of asking where price has been, it asks who is buying, who is selling, and how aggressively.

Pairs trading takes a long position in one asset and a short position in a closely related one, betting that the gap between them will close. It’s the simplest form of statistical arbitrage: one pair, one spread.

Statistical arbitrage — stat arb — trades temporary deviations between assets that historically move together, betting the relationship will reassert. Long the one that’s cheap relative to the relationship, short the one that’s rich, wait for convergence.

TWAP stands for time-weighted average price. In trading, a TWAP execution approach spreads an order across a period of time rather than attempting to fill the entire amount immediately. The time-weighted benchmark and the actual achieved execution price are not necessarily identical.

Walk-forward analysis fits a strategy’s parameters on a window of historical data, tests them on the following window, then rolls both windows forward and repeats. Every test period is out-of-sample — the strategy never saw that data when its parameters were chosen.

A good Sharpe ratio reflects attractive excess returns relative to return variability, but there is no universal number that makes a strategy safe or suitable. The ratio compares average excess return with its standard deviation over a defined period.

Sharpe = (average return − risk-free rate) ÷ standard deviation of returns

Sharpe and Sortino are both risk-adjusted performance ratios, but they use different measures of variability. Sharpe uses the standard deviation of excess returns. Sortino focuses on downside deviation relative to a specified target or minimum acceptable return.

Win rate is the proportion of completed trades classified as profitable under the reporting method. A strategy with 60 winning trades out of 100 completed trades has a 60% win rate. The treatment of break-even trades and costs should be stated.

The risk–reward ratio compares what you stand to lose on a trade if the stop hits to what you stand to gain if the target hits. Risking $100 to make $300 is 1:3.

A stop loss is an order or trading rule intended to reduce or close exposure after a specified adverse condition occurs. It expresses an exit condition; it does not guarantee that the position will close at an exact price or within an exact loss amount.

A trailing stop is a stop-loss order that moves with price. Set 5% below a long position, it rises as price rises and stays put when price falls. If price retraces 5% from its high, the stop triggers.

Alpha is the part of a return that can’t be explained by exposure to the market. If BTC rises 20% and a BTC strategy returns 25%, roughly 5% is alpha — the rest is beta, the return you’d have earned by holding.

Look-ahead bias is when a backtest uses information that wasn’t available at the moment the simulated trade was made. The strategy appears to predict the future because, in the test, it could see it.

Maximum drawdown is the largest decline from an equity peak to a later trough during a specified observation period. It describes the depth of a historical loss sequence, rather than simply the return between the first and last dates.

Profit factor is total gross profit from winning trades divided by total gross loss from losing trades.

Risk of ruin is the probability that a trader loses enough capital to be unable to keep trading — not necessarily to zero, but below the point where the strategy can operate.

Risk-adjusted return evaluates performance in relation to the risk taken to produce it. It is an approach to comparison, not one universally defined statistic. Different measures emphasize overall volatility, downside variability or peak-to-trough losses.

The Calmar ratio is annualised return divided by maximum drawdown over the same period.

The information ratio measures how much return a strategy earns above a benchmark, per unit of risk taken relative to that benchmark.

The Kelly criterion is a formula for the fraction of capital to risk per trade that maximises long-term compound growth.

The Sortino ratio is a variant of Sharpe that only counts downside volatility as risk.

A high-win-rate strategy can lose money when its losing trades are larger than its winners or when trading costs consume its gross gains. Win rate measures how often a trade wins, not how much the entire sequence earns.

AlphaNet charges a fee on each filled trade, calculated on trade notional. A dynamic model that factors in performance determines the fee rate, which may change over time. Its Points documentation describes a distribution pool funded from a portion of platform revenue.

Crypto trading fees reduce the amount of a gross trading gain that remains in the account. The full economic cost can also include the bid-ask spread, slippage and funding payments on perpetual contracts. These items are related, but they are not the same charge.

A perp DEX principally provides a market for perpetual contracts: orders, liquidity, margin and settlement. AlphaNet’s distinguishing focus is the quantitative strategy and execution process used to trade those markets. The venue and the strategy layer solve different problems.

A maker places an order that rests on the book — a limit order that doesn’t fill immediately. A taker places an order that fills against a resting one — a market order, or a limit that crosses the spread. Exchanges charge takers more, and sometimes pay makers.

AlphaNet Points are the platform’s participation-based reward units. The published framework covers users, guild leaders and content creators, with eligible activity contributing to a participant’s points balance. The design connects rewards to a pool funded by platform revenue.

A builder code lets an application that routes orders to Hyperliquid attach a small fee to those orders, paid by the user to the application. It’s how third-party interfaces, bots and platforms earn revenue on the venue without holding user funds.

A high-water mark is the highest value an investment has previously reached. When a performance fee uses one, the fee is charged only on gains above that peak — never on recovering ground that was already paid for.

Hyperliquid is accessed through a web interface that works on mobile browsers, plus third-party mobile clients built on its API.

Not natively. Two routes exist: depositing into a user vault, and third-party tools that mirror a chosen wallet’s trades into your own account.

Yes, through the testnet. Mock USDC, the live interface, real liquidations — no capital at risk.

You can trade perpetual futures that track US equity and index prices, including the S&P 500. You cannot buy shares.

Not for trading. Placing an order, cancelling one and being liquidated all cost zero gas. You pay maker or taker fees on fills and nothing for the transactions themselves.

Hyperliquid publishes a leaderboard in its interface and exposes position and fill data for any address through its API. The leaderboard endpoint itself is not in the documented API surface, which is why searches for it keep surfacing.

No. There is no identity verification, no account creation and no document upload. You connect a wallet and trade.

Not natively. Hyperliquid is a perpetual futures and spot venue. There are no listed calls or puts on the protocol’s own order book.

Deposit USDC through the Arbitrum bridge. If your funds are on another chain, either move them to Arbitrum first or use an integrated cross-chain route.

Any EVM-compatible wallet works. Open the interface, choose Connect, approve the connection in your wallet. There’s no account to create — the wallet is your identity, your balance and your signing authority.

The Hyperliquid explorer shows blocks, transactions, addresses and the trading activity settled in them. HyperEVM has a separate explorer for contract-level activity.

A position is liquidated when account equity falls below the maintenance margin for that position. The trigger and the price that measures it are both worth understanding before you size a trade.

Spot markets run on the same on-chain order book as the perpetuals, under the HIP-1 token standard. You buy and hold the token itself rather than a contract tracking its price.

You delegate HYPE to a validator and earn staking rewards. For traders, the more useful effect is that staked HYPE reduces your trading fees.

Entering a referral code when you first trade gives you a discount on trading fees for an initial volume window. The referrer earns a share of the fees you pay.

Payments between longs and shorts that keep the perpetual price tethered to spot. Hyperliquid pays them every hour, at one eighth of the computed eight-hour rate. Most venues settle every eight hours.

No deposit fee. A flat 1 USDC withdrawal fee, regardless of amount.

A 10% performance fee on depositor profits, paid to the vault leader. No deposit fee, no withdrawal fee, no management fee.

A dollar-cost-averaging bot builds a position in fixed increments — on a schedule, or triggered each time price falls a set percentage. Average entry improves as price drops.

A grid bot places a ladder of buy orders below price and sell orders above it, then repeats. Each oscillation within the range completes a pair and captures the gap.

HIP-3 is the standard that lets parties other than the protocol deploy perpetual futures markets. A deployer stakes HYPE, defines the market and its oracle, and the contract lists alongside the validator-operated perps.

The Hyperliquid Liquidity Provider vault — a protocol-run pool that makes markets and acts as backstop liquidator. Anyone can deposit USDC and take a share of its profit and loss.

Hyperliquid runs on a purpose-built Layer 1 using HyperBFT consensus, producing blocks in well under a second and reaching finality in the same window.

A central limit order book where every order, cancellation and fill is recorded on-chain. Buyers match against sellers by price and time priority, as on a centralised exchange — but the book lives on the blockchain, not on a company’s servers.

A parallel deployment of Hyperliquid using mock funds. Same interface, same order types, same margin mechanics, no real money.

Market, limit, stop-market, stop-limit, take-profit, trailing stop and TWAP, with post-only and reduce-only modifiers.