The Hyperliquid Field Guide

The Hyperliquid Field Guide

Everything you wish someone had explained before your first trade on Hyperliquid. From zero to confident in 14 short modules. Free, no signup.

Not affiliated with Hyperliquid Labs · Educational only · Not financial advice
🌱
I'm new to Hyperliquid Start from the stack and onboarding. Track A → D in order.
📈
I already trade perps Skip to the token, HLP and the ecosystem layer. Track B onward.
🏦
I'm here for yield HLP first, then vaults, DeFi and the ranked passive playbook.

◉ LIVE: funding & open interest right now

MarketMarkFunding (1h)AnnualizedOpen interest
Loading from the Hyperliquid public API…
Fetched in your browser from api.hyperliquid.xyz, the same public endpoint anyone can use. If it fails to load, example data is shown.
Track A · Foundations · 1 of 4

The Stack: what Hyperliquid actually is

People call it "a perps DEX." That's like calling New York "some buildings." Underneath is a full financial machine, three layers, two money pipes, one flywheel. Learn the map and everything else clicks.

Figures verified: July 2026
HyperEVM Smart contracts · lending · liquid staking · DEXes · stablecoins HyperCore On-chain order books · perps · spot · vaults · staking · HIP-3 markets HyperBFT Consensus · ~21 active validators · sub-second finality shared state, contracts can read the order book
The three layers. One chain, one state, the EVM and the exchange see each other.

Layer by layer

LayerWhat it isWhy it matters to you
HyperBFTCustom high-speed BFT consensus, small validator set (~21 active)Speed comes from this; so does the venue's biggest trust assumption (see Risk Map)
HyperCoreThe native exchange state machine, order books run fully on-chainCEX-grade trading with DEX-grade transparency: every position, every fill, public
HyperEVMEthereum-compatible smart-contract layer (live since Feb 2025)The DeFi playground: lending, liquid staking, yield, composable with the exchange

Where the money comes from

Hyperliquid does more perp volume than the rest of the perp-DEX field combined and generates hundreds of millions in annualized fees. The unusual part is where those fees go:

Traders perp + spot volume Trading fees taker / maker / auctions Assistance Fund buys HYPE on market Losing / liquidated positions HLP vault market-making + backstop Depositors anyone with USDC Top row: the fee → buyback flywheel. Bottom row: the trading-PnL → vault-yield pipe.
The two money pipes every "yield" in this course ultimately drinks from.

This matters because it means passive yield on Hyperliquid is mostly backed by real cash flow from trading activity, not token printing. It also means every yield here shrinks if volumes shrink. Real revenue cuts both ways.

Checkpoint

1. Where do Hyperliquid's order books run?
Off-chain, settled on-chain later Fully on-chain, inside HyperCore On HyperEVM smart contracts
HyperCore runs the books on-chain at sub-second speed, the core differentiator vs. CEXes (off-chain books) and AMM DEXes (no books at all).
2. What backs most passive yield on the chain?
Token emissions VC subsidies Real trading-fee revenue and market-making profits
Fees, spread and liquidations are real cash flow, rare in DeFi. The flip side: yields track trading volume.
Track A · Foundations · 2 of 4

Getting On: wallets, deposits, first-week safety

The least glamorous module saves the most money. Ten minutes of setup paranoia has outperformed every trading strategy ever devised.

Figures verified: July 2026

What you need

  1. A self-custody wallet (MetaMask, Rabby, or a hardware wallet, use hardware for anything you'd mind losing). Hyperliquid is non-custodial: your keys, your funds, your responsibility.
  2. USDC: the collateral asset for trading and most yield products.
  3. Two minutes of paranoia: configured below.

Deposit routes

RouteHowNotes
Arbitrum bridge (canonical)Send USDC on Arbitrum → deposit via the appThe original route; cheap and fast
CEX transfersWithdraw USDC from an exchange directly toward Hyperliquid via supported routesCheck the current supported list in the app, routes expanded through 2025–26
Cross-chain routersThird-party bridges/routers from other chainsExtra smart-contract risk; use majors only

Understand your account structure

First-week safety checklist
  • Bookmark app.hyperliquid.xyz and only ever enter via the bookmark. Fake frontends are the #1 practical way people lose funds.
  • Never sign a transaction from a link in a DM, reply, or "airdrop eligibility checker."
  • Test the full loop small first: deposit $20, place a tiny order, withdraw. Verify the door works both ways.
  • If using API keys/bots later: minimal permissions, separate wallet.
  • On HyperEVM: periodically revoke unused token approvals.
At small size vs. large size Small accounts: software wallet is acceptable, fees matter more than security theater. Once your on-chain net worth would hurt to lose: hardware wallet, approval hygiene, and a separate "hot" wallet for experiments. The transition point is personal, but most people make it too late.

Checkpoint

1. The most common practical way users lose funds is…
Phishing: fake frontends and malicious signature requests Exchange hacks Bad trades
Bad trades lose money slowly; one malicious signature loses everything at once. Bookmark the app, distrust links.
2. In unified account mode…
You must transfer USDC from spot to perps before trading Spot USDC directly collateralizes perp positions Perps and spot use different currencies
No internal transfer needed, spot USDC backs your perps automatically. (Tools reading the API should sum spot + perps state to see the real balance.)
Track A · Foundations · 3 of 4

Perps From Zero

Perps are simple: bet up or down, with borrowed size, forever. The details are where accounts go to die, so we're doing the details properly.

Figures verified: July 2026

The core idea

You post USDC collateral, choose a direction (long = up, short = down) and a leverage multiple. Position size = collateral × leverage: $100 at 10x controls a $1,000 position. Your profit and loss track the full $1,000, which is the entire appeal and the entire danger. A 5% move in your favor is +50% on your collateral; a 10% move against you is roughly all of it.

What keeps a perp glued to the real price: funding

Perps have no expiry, so something must tether the perp price to spot. That something is the funding rate, an hourly, peer-to-peer payment between traders (the exchange takes no cut):

Expert habit Funding is also a sentiment gauge. Persistently high positive funding = crowded longs = fragile market; extremes often precede squeezes. Professionals read funding the way meteorologists read pressure, the live panel on the home page shows it for the majors right now.

⚙ Funding bleed simulator

Margin modes: your blast-radius setting

Cross marginIsolated margin
CollateralAll cross positions share one poolFenced to that single position
If liquidatedCan take the whole account downOnly that position's margin is lost
Best forExperienced, hedged booksLearning, and single directional bets

Default recommendation for anyone not running a hedged book: isolated. Cap the blast radius of every mistake.

Order types worth knowing

Checkpoint

1. Funding exists to…
Pay the exchange Punish short sellers Tether the perp price to the spot price
It's the substitute for an expiry date: whichever side pushes the perp away from spot pays the other side to bring it back.
2. You hold a $2,000 notional position at 20x. Funding is charged on…
Your $100 collateral The full $2,000 notional Nothing, funding only applies to shorts
Notional, always. That's why high leverage + high funding quietly bleeds accounts that "aren't even moving."
3. Very high positive funding usually signals…
Crowded longs and elevated squeeze risk Guaranteed continuation up A data error
The market is paying heavily to be long, a crowded side. Crowded trades unwind violently.
Track A · Foundations · 4 of 4

Liquidation & Fees: the two silent killers

Nobody quits perps because they can't read charts. They quit because two invisible forces, liquidation math and fees, were working against them from trade one. Master both and you're ahead of most of the market before you ever open a position.

Figures verified: July 2026

How liquidation works

You're liquidated when account equity (including unrealized PnL) drops below the maintenance margin, defined as half the initial margin at the asset's maximum leverage (not the leverage you chose). For a 40x-max asset like BTC that's ≈1.25% of notional; for a 3x-max asset, ≈16.7%. Liquidation triggers on the mark price (a blend of CEX prices and the HL book), so a single manipulated wick won't save or kill you.

Equity < maintenance liquidation triggered 1. Order-book close position sold on the book 2. Backstop if equity < ⅔ maintenance Liquidator vault (HLP) absorbs position + your margin In a backstop liquidation you forfeit the maintenance margin. A stop-loss placed earlier keeps that buffer in your pocket.
The liquidation waterfall. Remember who's at the end of it. HLP depositors. You'll meet them again in Track B.

⚙ Liquidation buffer: how far can price move against you?

Approximation for a 40x-max asset (maintenance ≈1.25%), isolated margin: buffer ≈ (1 ÷ leverage) − 1.25%. At high leverage, ordinary intraday noise is fatal before any thesis can play out.

Fees: small numbers, huge consequences

Base tier: 0.045% taker / 0.015% maker (14-day rolling volume tiers reduce these; staking HYPE discounts them further, see Track B). Sounds tiny. Now multiply by leverage:

⚙ Round-trip fee calculator

The lesson most people learn too late Fees are charged on notional. At 10x, a full taker round trip costs ~0.9% of your collateral, before price moves a cent. Trade 20 times a month like that and you need a ~20% monthly edge just to break even. Maker orders and lower trade frequency are not style choices; they're survival math.

Checkpoint

1. Maintenance margin depends on…
The leverage you selected The asset's maximum allowed leverage Your account age
Half the initial margin at the asset's max leverage. Your chosen leverage sets your buffer; the asset's max sets the floor.
2. Why does a stop-loss beat "riding it to liquidation"?
It doesn't; same outcome Stops are free to place You exit before forfeiting the maintenance-margin buffer to the backstop
Liquidation, especially backstop, costs you the buffer. A stop returns it to you. Same wrong trade, meaningfully different bill.
3. Approximate buffer to liquidation at 20x on BTC-type assets?
~3.75% ~20% ~10%
(1/20) − 1.25% ≈ 3.75%. BTC moves that much on a boring Tuesday. Leverage mostly buys you an earlier funeral.
Track B · Token & Native Yield · 1 of 3

HYPE & Staking: the quiet backbone

The token that pays for everything, secures everything, and gets bought back with real revenue every single day. The yield is boring. The asset is not.

Figures verified: July 2026

What the token actually does

Genesis fact worth knowing: 31% of supply was airdropped to users in November 2024 with zero VC allocation, the most community-weighted major token distribution in DeFi history, and the root of the ecosystem's loyal user base.

Staking mechanics

  1. Move HYPE: spot balance → staking balance (instant).
  2. Delegate to a validator (1-day lockup per delegation). Rewards accrue every minute and auto-compound.
  3. Exit: undelegate (after lockup), then a 7-day queue from staking back to spot.

Base reward rate floats inversely with total stake, roughly 2.2–2.4% APY at current levels, minus a small validator commission (~2–3%). Sites advertising double-digit "HYPE staking APY" are marketing a wrapped product, a promo, or extra risk. The native number is the native number.

Fee-discount tiers (staked HYPE)

TierStaked HYPEFee discount

Picking a validator

The exit-speed trade-off Staked HYPE takes ~8 days to become sellable. In a black-swan event, that's 8 days of forced holding. Liquid staking tokens (Track C) remove the queue but add smart-contract and depeg risk. There is no free option, only a choice of which risk you prefer.
At small size vs. large size Staking is an investment in HYPE first, yield second. 2.3% won't change anyone's life; the token's trajectory will dominate the outcome either way. The exception is the Wood tier: staking just 10 HYPE gives a 5% fee discount, effectively free for anyone who holds any HYPE and trades at all. At large size, the real decision is portfolio-level: how much HYPE price exposure do you want as a bet on the fee machine from Module A1?

Checkpoint

1. Total time from staked to sellable?
Instant ~8 days (1-day delegation lockup + 7-day queue) 30 days
Plan exits before you need them. This delay is precisely what liquid staking tokens monetize.
2. Realistic native staking yield right now?
~18% ~8% ~2.2–2.4%
The emission formula floats with total stake. Anything advertising much more has added risk, a wrapper, or fiction.
3. A bad validator choice can cost you…
Missed rewards only, delegation is non-custodial Your entire stake A slashing penalty on your principal
Delegators keep custody. Downtime costs rewards; your principal stays yours.
Track B · Token & Native Yield · 2 of 3

HLP: being the house instead of the gambler

Every casino has a house, and the house always wins, on average, over time, with occasional terrible nights. On Hyperliquid, you can be the house. Here's exactly what that seat costs.

Figures verified: July 2026
Traders longs, shorts, liquidations HLP quotes both sides absorbs backstops collects funding You pro-rata share of PnL spread payouts When traders lose collectively, HLP wins. When they win collectively, HLP pays. You sit on HLP's side.
The seat you're buying: the other side of the venue's order flow.

The three profit engines

  1. Market-making spread: quoting both sides of every book, all day.
  2. Backstop liquidations: absorbing busted positions, usually inheriting the maintenance margin at a profit.
  3. Funding capture: collecting funding on inventory it holds.

Zero fees: no management fee, no performance cut. Deposits have a 4-day lockup. Every cent of PnL is on-chain and public.

The honest return profile

The mental model HLP is a short-volatility, long-chaos strategy: it collects small profits most days and occasionally eats, or feasts on, a tail event. Judge it only by its full on-chain PnL curve, never a headline APY snapshot.

The incident history: required reading

EventWhat happenedLesson
JELLYJELLY
Mar 2025
Coordinated manipulation of an illiquid token forced HLP to absorb a toxic short (~$4M temporary hole); emergency delisting via validator voteThe backstop role makes HLP a target; governance acted fast, and centrally
POPCAT
Nov 2025
Similar coordinated attack on another thin marketThe vector repeats; listing standards and caps tightened again
Fartcoin
Apr 2026
Another manipulation attempt involving HLP exposureTreat this attack class as permanent background radiation

Every incident so far was absorbed and the PnL curve recovered. That's a good record, not a guarantee, a large enough attack or bug could impose permanent loss. Size deposits accordingly.

At small size vs. large size HLP is the strongest risk-adjusted native yield on the chain: real cash flow, zero fees, radical transparency. Sensible use at any size: treat it as the core of a Hyperliquid yield allocation, sized so a 15% drawdown is annoying rather than devastating, checked monthly rather than daily. Money you might need within weeks doesn't belong here (4-day lockup + drawdown timing risk).

Checkpoint

1. HLP's profit engines are…
Trading fees, gas, emissions Lending interest and staking Spread, backstop liquidations, funding capture
Pure trading-desk economics, shared with depositors at zero fees. No fee revenue, no token printing.
2. HLP's worst environment is…
High-volatility liquidation cascades Sustained trends where traders collectively win Quiet, choppy markets
Chop pays spread; cascades pay windfalls; trends where the crowd is right hand HLP the losing side. Short-vol in one sentence.
3. Why does the incident history matter if every attack was absorbed?
It proves the attack class is recurring, a permanent input to position sizing It doesn't; the past is the past It proves HLP can never lose
Three coordinated attacks in ~13 months is a pattern, not trivia. Recovery each time is encouraging; assuming it's guaranteed is how tail risk gets you.
Track B · Token & Native Yield · 3 of 3

User Vaults: hiring a trader with one click

A hedge-fund marketplace with no regulator and perfect receipts. One click to hire a trader, one checklist to avoid hiring the wrong one.

Figures verified: July 2026
Vault leader must keep ≥5% ownership Depositors one-click subscribe Pooled capital traded by the leader Profits 90% depositors 10% leader fee Every fill the vault ever takes is public. The track record cannot be faked, only misread.
Vault anatomy: aligned-ish incentives, radical transparency, zero regulation.

The mechanics

The six-point vetting framework

Treat every vault as an actively managed fund with no regulator, because that's exactly what it is:

  1. Age. A 90-day track record is a marketing pitch. Demand ≥6 months of live history including at least one genuinely ugly market week.
  2. Max drawdown before APR. 200% APR with −60% drawdown is strictly worse than 80% with −15%. Drawdown reveals what the leader does when wrong, the only thing that matters long-term.
  3. Leader ownership trend. 5% is the floor; the direction is the signal. A leader steadily reducing their own stake is telling you something. Listen.
  4. TVL vs. track record. A vault that ballooned from $100K to $5M now runs a different, harder strategy than the one that produced the numbers you're admiring.
  5. Strategy legibility. Read the actual fills. Can you name the edge? If returns look too smooth, ask what blows it up. Something does.
  6. Leverage habits. Check typical position size vs. vault equity in the history. Quiet 20x users eventually donate the vault to the market.
The base rate nobody advertises Most user vaults underperform or die. The leaderboard is survivorship bias rendered in real time, collapsed vaults quietly vanish from the top lists. Assume the median vault loses to simply parking in HLP, and demand strong evidence that any specific vault is the exception.
Two honest uses (1) As education: the trade history of a proven leader is a free masterclass, reverse-engineer entries and exits against your own analysis. (2) As a satellite allocation: small, only after the full six-point check, and with the expectation of sometimes being wrong anyway. Never let a hot quarterly APR set your deposit size.

Checkpoint

1. Which vault is more attractive?
80% APR, −15% max drawdown, 14 months old 200% APR, −60% max drawdown, 3 months old The one with the most TVL
Drawdown-adjusted, aged track records win. The 200%/−60% vault is one bad month from halving.
2. Why is the vault leaderboard misleading?
The numbers are fake It updates too slowly Survivorship bias, dead vaults disappear from it
On-chain numbers are real; the selection is biased. You see the survivors, not the graveyard that funded them.
Track C · The Ecosystem · 1 of 3

Spot, Listings & Stablecoins

Where tokens are born by auction, real bitcoin lives on-chain, and billion-dollar stablecoin politics get settled by validator vote. Smaller than perps. Much better chess.

Figures verified: July 2026

HIP-1: tokens are born by auction

Nobody at Hyperliquid "decides" listings. Ticker slots are sold in a continuous Dutch auction paid in HYPE, which gets burned. Projects bid for the right to deploy a token on HyperCore. Result: spot listings are scarce and expensive, a deliberate quality filter versus the memecoin free-for-all on most chains.

HIP-2: liquidity from birth

New tokens don't launch into empty books. Hyperliquidity is a protocol-native market-making mechanism that automatically seeds a two-sided order book for every new token from day one. No paying market makers, no ghost-town charts.

Project bids HYPE Dutch auction Ticker deployed HYPE burned 🔥 HIP-2 seeds book liquidity on day one Live
HIP-1 + HIP-2: scarce listings, instant liquidity, burned HYPE.

Unit: real majors on the chain

Unit is the tokenization layer that brought native-asset spot to Hyperliquid. UBTC, UETH, USOL and more. You can hold and trade the majors on-chain, which unlocks the strategies in Track D: hold spot BTC, short the BTC perp, collect funding, all on one venue.

The stablecoin saga (a governance case study)

Why you should care Stablecoin plumbing here is a moving target decided by validator votes and business deals. Before parking size in any yield product, check which stablecoin it actually holds and who stands behind it. "USDC" in a product name is a claim to verify, not a fact to assume.

Checkpoint

1. How do spot tokens get listed?
Team decision Dutch auction for ticker slots, paid in HYPE that gets burned Free permissionless listing
HIP-1 auctions: scarcity as a quality filter, plus a structural HYPE burn.
2. Unit's role in the ecosystem is…
A stablecoin issuer A copy-trading platform Bringing native BTC/ETH/SOL spot assets on-chain
UBTC, UETH, USOL, the building blocks for real portfolio holding and basis trades on the venue.
Track C · The Ecosystem · 2 of 3

HIP-3: anyone can run an exchange

Stake half a million HYPE, get your own exchange. Tesla perps at 3am, gold on a Sunday, and three brand-new ways to lose money that your crypto playbook doesn't cover.

Figures verified: July 2026
Builder stakes 500,000 HYPE (slashable) Their own perp DEX chooses assets, oracle, collateral, fees HL engine matching, margining, liquidations inherited User fees ≈ 2× normal 50% to the deployer
HIP-3: the bond IS the trust model. Malicious operation gets the stake slashed.

Why it matters

Trading HIP-3 markets, the extra risk list
  • 2× fees: round-trip math from Track A gets twice as bad.
  • Oracle quality varies by deployer: stock perps depend on the deployer's feed: market close handling, halts, splits.
  • Weekend gap risk: the underlying stock market closes; the perp doesn't. Friday-to-Monday gaps blow through stops.
  • Thinner books: wider spreads, worse slippage than the main markets.
Practical takeaway Understand HIP-3 as part of the HYPE investment thesis (fee growth + staking sink) and as optionality for exposure you can't get elsewhere. For active trading, the main validator-operated markets remain deeper and half the price. If you trade a stock perp anyway, gaps, funding and oracle rules are three new ways to lose that a crypto playbook doesn't cover.

Checkpoint

1. What secures an HIP-3 market against a malicious operator?
The deployer's 500,000 HYPE stake is slashable A government license Community moderation
Economic security: misbehave, lose a multi-million-dollar bond. No permission needed, but skin in the game required.
2. Why is trading a Tesla perp on Sunday different from trading BTC?
It isn't The underlying market is closed. Monday's open can gap through your stop Funding doesn't apply to stock perps
The perp trades 24/7 against a reference that only updates on market days. Gap risk is structural, not incidental.
Track C · The Ecosystem · 3 of 3

HyperEVM DeFi: the yield playground

A billion and a half dollars sits on the smart-contract layer, earning more than the base chain pays. The yields are real. So is every extra way to lose the principal. Both compound.

Figures verified: July 2026

The map

CategoryExamplesWhat you earn
Liquid stakingKinetiq (kHYPE), Looped HYPE (LHYPE)Staking yield (~2.2%) with the token staying liquid & usable as collateral
LendingHyperLend, FelixUSDC ~3–8% APY; HYPE 6–12% when borrow demand spikes
CDP / stablecoinsFelix (feUSD)Mint against collateral; stability-pool yields
DEXes / LPsKittenSwap, HyperSwap, LaminarSwap fees + incentives (minus impermanent loss)
AggregatorsHyperbeat & peersAuto-compounded strategies, typically 1–6% on stables

The signature move: yield stacking

1 · Stake HYPE → kHYPE ~2.2% staking yield · +contract risk 2 · kHYPE as lending collateral +lending yield · +another contract · +depeg risk 3 · Borrow → buy more → loop amplified yield · +liquidation risk · leveraged long floor 1: fine floor 2: acceptable, eyes open floor 3: not passive income
The yield tower. Every floor adds income AND a new way to lose the principal.
The risk stack, read twice
  • Smart-contract risk × N: your worst case is the weakest contract you touch, and stacking multiplies N.
  • Depeg risk: kHYPE/LHYPE can trade below HYPE precisely when you most want out. Liquid staking trades the 7-day queue for this.
  • Liquidation risk: a looped position is a leveraged long wearing a yield costume. Collateral drops → cascade.
  • Incentive decay: headline APYs usually include temporary emissions. Durable HyperEVM stable yields cluster around 3–8%; a 40% farm is emissions, extra risk, or both.

The 10-minute protocol vetting routine

  1. Audits: at least one reputable firm; check the audit date against the current code.
  2. TVL trend and age. 6+ months of TVL that survived a drawdown beats a shiny launch.
  3. Name the payer: who pays this yield? Borrowers? Traders? Emissions? If you can't name the payer, the payer is you.
  4. Exit test: deposit small, withdraw immediately. Verify the door before bringing furniture.
  5. Admin keys: can a multisig upgrade contracts and touch funds? (Usually yes on young chains, that's a trust assumption to price, not ignore.)

Checkpoint

1. kHYPE's advantage over native staking is…
Much higher yield Liquidity and composability, no 7-day queue, usable as collateral Zero risk
Same underlying ~2.2%, the win is liquidity. The price: contract and depeg risk. No free option, ever.
2. A 3x looped kHYPE position is best described as…
Triple passive income A delta-neutral strategy A leveraged long on HYPE with extra yield and liquidation risk
Borrowing to re-buy the same asset amplifies yield, price exposure and liquidation risk together. Floor 3 of the tower is a trade, not income.
3. "Who pays this yield?" is the right first question because…
If you can't identify the payer, you're probably the payer It's polite Yields are always paid by exchanges
Sustainable yield has an identifiable source: borrowers' interest, traders' fees, funding. "Points" and emissions are you being paid with dilution, sometimes worth it, never confuse the two.
Track D · Mastery · 1 of 4

The Passive-Income Playbook

Eight ways to earn on this chain. One honest table. Zero hype. This is the page to reopen the night before you move real money, and the morning after you get excited about something.

Figures verified: July 2026
RISK → REALISTIC YIELD → HYPE staking ~2.3% kHYPE ~2.2% USDC lending 3–8% HLP 15–30% (lumpy) Basis trade ≈ funding Vetted user vault ??? Looping / farms (and down to zero) Everything on this chart also shares ONE venue risk (see Risk Map), that's the invisible axis.
The menu, plotted honestly. Up and to the left is where sleep lives.

The full menu

OptionRealistic yieldMain risksLiquidityEffort
USDC lending (major market)3–8%Contract, bad debtUsually instantNone
HYPE staking (native)~2.3% + fee discountsHYPE price, 8-day exit8 daysNone
kHYPE (liquid staking)~2.2%+ contract & depegInstant (market)None
HLP vault15–30% long-run, lumpyCounterparty, tail events, 5–12% drawdowns4-day lockupNone
Delta-neutral basis≈ funding rate (5–20%+ in hot markets)Funding flips, margin managementDaysMedium, monitoring
User vaults (vetted)Assume less than advertisedLeader risk, survivorship bias~1-day lockupVetting is work
LP / farming / pointsWildly variableIL, emissions decay, rugsVariesHigh, it's a job
LoopingAmplified, both directionsLiquidation cascadePoor in stressHigh + stressful

⚙ Yield explorer: filter the menu by your constraints

The one advanced strategy worth learning properly: the basis trade

LONG spot (e.g. UBTC) gains when price rises SHORT perp, 1x, same size gains when price falls = Price exposure ≈ zero You keep collecting funding on the short
Delta-neutral: direction cancels, funding stays.

⚙ Basis-trade simulator: drag the price, watch the PnL

Real-world caveats: funding can flip negative for long stretches (then you pay); a violent pump can liquidate an under-margined short leg before you rebalance (keep it at 1x, margin topped up); entry/exit fees eat weeks of funding, so this is a months-scale position, not a days-scale one.

An allocation framework (a template, not advice)

The invisible axis Everything above lives on one venue. Spreading across HLP + lending + staking is strategy diversification, not venue diversification, a chain-level failure hits it all at once. That's the argument for never putting all investable capital on any single chain, no matter how impressive.

Checkpoint

1. The basis trade makes money from…
Price appreciation Funding payments while price exposure is hedged to ~zero Leverage
Long spot + short perp cancels direction; the short collects funding while it's positive. The enemy is negative funding, not price.
2. Holding HLP + lending + staking + a user vault means…
Full diversification Guaranteed positive returns Strategy diversification, but 100% exposure to one venue
Four strategies, one chain, one bridge, one validator set. Real diversification requires capital off the venue too.
Track D · Mastery · 2 of 4

The Risk Map: what can actually break

Amateurs can tell you what might go right. Experts can tell you exactly what can go wrong, in order of blast radius. This is that list, unflinching, worst-first.

Figures verified: July 2026
Your funds self-custody wallet Bridge crypto's most-hacked layer ~21 validators ⅔ could forge state Contracts every EVM position Your money passes through every box. Each one is a separate trust assumption.
The trust chain. Rank your exposure by the weakest box you touch.

1 · Validator concentration

HyperBFT runs on a small permissioned set (~21 active validators). Two-thirds control, roughly 14 colluding or compromised nodes, could theoretically forge state, including withdrawals. The set is growing and decentralizing over time, but today this is a real, material trust assumption compared to Bitcoin or Ethereum.

2 · Bridge risk

USDC enters via a bridge contract. Bridges are the most-hacked component class in crypto history. A bridge exploit, or a Circle-level freeze on the deposit contract, would hit everything on the chain simultaneously. The 2026 Coinbase/USDC partnership deepens the plumbing; it doesn't remove the dependency.

3 · Governance centralization

The JELLY incident proved the team can act unilaterally-fast via validator vote in an emergency, delisting an asset and settling positions at chosen prices. That decisiveness protected HLP depositors and proved the venue is not neutral infrastructure yet. Model Hyperliquid as a very good, very transparent, still centrally-steerable exchange.

4 · The HLP attack surface

Three coordinated manipulation attempts in ~13 months (JELLYJELLY, POPCAT, Fartcoin), each targeting the backstop role via illiquid assets. Parameters tightened after each. Assume the cat-and-mouse is permanent and size HLP deposits with it priced in.

5 · Smart-contract risk (HyperEVM)

Young chain, young protocols, large TVL honeypot. Every EVM position adds contract risk on top of venue risk. Audits reduce it; nothing removes it.

6 · Regulatory & jurisdictional risk

Hyperliquid geo-blocks some jurisdictions and lives in gray zones elsewhere. Access rules and the tax treatment of perps, funding and vault income vary by country and change. Know your local situation before yields become meaningful money.

7 · You (the risk you control)

The quarterly re-check This threat model is a snapshot. Validator count, bridge design, and governance maturity all improve over time, re-run the list quarterly. The 2027 picture should be better than today's. Verify that it is.

Checkpoint

1. The most systemic technical risk to funds on the chain is…
A vault leader losing money Bridge compromise or validator collusion High funding rates
Vault losses are local; a bridge or validator failure hits every dollar on the chain at once.
2. The JELLY incident's governance lesson:
The team can and will intervene fast and centrally in emergencies Governance is fully decentralized Nothing happened
Central speed saved depositors money AND revealed steerability. Both halves of that sentence belong in your risk model.
3. Adding a HyperEVM lending position to your HLP deposit…
Diversifies away your risk Changes nothing Adds contract risk on top of the venue risk both share
Same chain underneath, plus a new contract layer. Count risks per layer touched, not per app used.
Track D · Mastery · 3 of 4

Trading Survival Rules

You're going to trade anyway. Fine. These ten rules decide whether the market charges you tuition or takes the whole account. Most people should trade less. Everyone should trade smaller.

Figures verified: July 2026

The ten rules

  1. Risk a fixed fraction per trade. 1–2% of the account, never more. At 1% risk you can be wrong 20 times in a row and still be standing. At 10% risk, seven bad trades end you.
  2. Isolated margin while learning. Cap the blast radius of every position to what you allocated to it.
  3. The stop-loss goes in with the order, not "later." Later means after the damage. Decide the invalidation point before entry; if you can't name it, you don't have a trade.
  4. Leverage is position-sizing math, not a conviction dial. Choose the stop distance first, then the size that makes that stop cost 1–2%, and let leverage fall out of the arithmetic. 3–5x covers almost every legitimate use.
  5. Count fees and funding before entry. Track A's calculators exist for a reason: a strategy that trades often at taker fees needs an implausible edge just to reach zero.
  6. Expectancy beats win rate. A 40%-win-rate system that wins 2R and loses 1R prints money; an 80%-win-rate system that wins 0.3R and loses 2R bleeds out. Track (win% × avg win) − (loss% × avg loss), nothing else.
  7. Journal every trade, thesis, entry, stop, exit, emotional state. Unjournaled trading is unfalsifiable, and unfalsifiable systems never improve. Your fills are all on-chain; you have no excuse for fuzzy records.
  8. Daily loss limit: two losers or a fixed %, then stop. The third loss of the day is rarely analysis; it's usually revenge. The market opens again tomorrow.
  9. A correct no-trade is a win. Missing a move you had no setup for costs zero. The scoreboard is process adherence, not action.
  10. No rule changes on small samples. Ten trades prove nothing. Change systems on evidence, quarterly, one variable at a time, or you're just wandering with extra steps.
Why these rules exist Every rule above is a defense against the same enemy: the gap between what you decided calmly and what you do with a position open. The venue is fair, fast and transparent; the counterparty that empties most accounts is in the chair, not the code.

Checkpoint

1. Risking 1% per trade means 20 straight losses cost you…
Everything, you're liquidated Roughly 18% of the account, painful, survivable Exactly 20%
Compounding down: 0.99²⁰ ≈ 0.82. Survivable. That's the entire point of fractional risk, it buys you enough attempts for skill to matter.
2. Which system makes money?
80% win rate, +0.3R wins, −2R losses Neither, win rate is all that matters 40% win rate, +2R wins, −1R losses
System A: 0.8(0.3) − 0.2(2) = −0.16R per trade. System B: 0.4(2) − 0.6(1) = +0.2R. Expectancy, not win rate.
3. Leverage should be chosen…
Last, derived from stop distance and the 1–2% risk cap First, based on conviction Always at maximum for capital efficiency
Stop distance → position size → leverage falls out. Conviction-based leverage is how liquidation tables get populated.
Track D · Mastery · 4 of 4

Operating Like a Native

The last 10%: a free Bloomberg terminal almost nobody opens, an airdrop culture full of traps, and the habit loop that keeps you sharp long after this tab closes.

Figures verified: July 2026

The info API: everyone's Bloomberg terminal

Everything on Hyperliquid is queryable from one public endpoint, no key required: POST https://api.hyperliquid.xyz/info. The shortlist worth memorizing:

Request typeReturns
metaAndAssetCtxsEvery perp market: funding, open interest, mark/oracle prices (powers this site's live panel)
clearinghouseStateAny address's positions, margin and leverage, yes, anyone's
spotClearinghouseStateAny address's spot balances
vaultDetailsAny vault's full PnL history and depositors
delegatorSummaryStaking positions and rewards

Radical transparency is the edge: whale positions, vault performance, funding across every market, public, free, real-time. No CEX user has this. Most Hyperliquid users never use it. Be the exception.

The airdrop meta: handled like an adult

The November 2024 genesis airdrop (31% of supply, zero VC) is the most famous wealth-distribution event in DeFi, and it permanently changed behavior: a large population now farms everything, hoping for a sequel.

The sane play Use protocols you'd use anyway. Keep real balances active because the products are good. If an airdrop lands, it's a bonus on behavior you'd have done for free. The moment farming becomes the strategy rather than the bonus, you're doing unpaid labor with principal at risk.

Referral & builder codes

Referral codes give new users a fee discount (~4%) for an initial volume period and pay the referrer a share, the standard, transparent monetization for content and communities in this ecosystem (including, disclosure, this site's). Builder codes let app developers charge a small routed-order fee, the business model behind most third-party Hyperliquid frontends and bots. Neither is sinister; both should always be disclosed.

The maintenance loop: staying an expert

  1. Primary sources first: the official docs and announcements. Third-party guides (this one included) age; the docs are the ground truth.
  2. Quarterly re-check: staking APY, HLP PnL curve, fee schedule, validator count, top HyperEVM protocols by TVL.
  3. Data over narrative: when a claim sounds exciting, pull the number from the info API before believing it. You now know how.
Graduation criteria You've internalized this course when you can, from memory: explain how HLP makes and loses money; rank staking, lending, HLP and basis trades by risk; describe the JELLY incident and its implication; sketch HIP-1/2/3; and name the three biggest venue-level risks. The Final Exam checks exactly this.

Checkpoint

1. The smartest stance on airdrop farming:
Buy a guide that reveals the snapshot criteria Avoid the ecosystem entirely Use good protocols you'd use anyway; treat any airdrop as a bonus
Criteria are never pre-announced. Real usage of real products is the only farming that's never wasted effort.
2. Where should a surprising Hyperliquid claim be verified?
Official docs and the public info API A YouTube video from last year A Telegram alpha group
Primary sources and live data. The ecosystem changes monthly; checking beats memorizing.
Legal

Privacy policy

Plain language, because that's the whole brand. Last updated: July 2026.

Who runs this site

The Hyperliquid Field Guide is a free educational project by , an individual based in the EU. It is not affiliated with Hyperliquid Labs. Contact: via X .

What data this site touches

WhatWhere it livesConsent needed?
Course progress & quiz/exam scoresOnly in your browser (localStorage keys hfg_progress_v1, hfg_consent). Never transmitted anywhere.No, strictly functional, no tracking
Live market dataYour browser calls Hyperliquid's public API (api.hyperliquid.xyz) directly to show funding/OI. Like any web request, that server sees your IP address. No identifiers are sent.No, but noted here for transparency
Analytics (Google Analytics 4)Loaded only after you click Accept in the cookie banner. Sets _ga cookies, collects pseudonymous usage data (pages viewed, device type, approximate location). IPs are anonymized by Google by default.Yes, opt-in. Decline and GA never loads; the site works identically
FontsSpace Grotesk is loaded from Google Fonts (your IP is visible to Google's font CDN on load).Noted for transparency

What this site does NOT do

Your rights (GDPR)

You can withdraw analytics consent anytime via "Cookie settings" in the footer. You can wipe all local data with "Reset my progress" (module drawer) plus clearing this site in your browser settings. For anything GA-related (access, deletion), contact us via X and we'll handle it; analytics data is retained for the minimum period GA4 allows (2 months) and is not linked to your identity by us.

Reference

Glossary

Every term the course uses, in one place.

Reference

Changelog & data freshness

Hyperliquid moves fast. Every number in this course lives in one data block with a verification date. This page is the honesty ledger.

DateChange
2026-07Initial public release. All figures verified against official docs and on-chain data as of July 2026.

Figures most likely to drift: staking APY, HLP returns, lending rates, fee schedule, HIP-3 open interest, HyperEVM TVL. If you spot something stale, ping on X.

Final exam

The Field Exam

20 random questions from the whole course. 16+ means you genuinely understand this machine better than most people using it.