FAQ

How to read this terminal

Definitions, math, provenance, and the things you should not trust. Hover any tagged term in the app to see the same answer inline; the link from a tooltip drops you on the right entry here.

Markets & contracts

What is HIP-4?#

HIP-4 is Hyperliquid's on-chain framework for short-dated event markets. Each market is binary: contracts settle to $1 if the named condition is true at expiry, $0 if not. Markets live natively on HyperCore alongside perps and spot, with their own order books.

Why is there a separate "Yes" side and "No" side?#

Each outcome lists two markets — one to take the YES position, one to take NO. They have independent books, prices, and OI. In an efficient market, Yes price + No price ≈ $1; deviations are the basis of the completeness arb signal.

What are multi-categorical events?#

Some questions have more than two possible outcomes (e.g. "Which token will trade most volume in May?"). HL groups these as questions[] in outcomeMeta. We render them as Event cards on the home page so you see the grouping; individual Yes/No rows for each candidate also appear in the markets table.

Why is there no "Execute" or "Place order" button?#

This site is analytics-only. Orders happen on Hyperliquid; the "Trade on Hyperliquid" link on each market opens the HL app for the same coin. Keeping execution off this surface is a deliberate legal choice and is not changing.

Signals

Four kinds of signals are derived on every market tick. Each appears as a row on /signals with a one-line take, capacity, and edge after spread cost.

What is a DIV (divergence) signal?#

The perp-implied probability of an event differs from the outcome market mid by ≥5 percentage points. We compute the perp-implied probability from the underlying price, the strike, time to expiry, and 24h-realized vol (log-normal terminal spot model). Direction tells you which side is "richer" — long the cheaper side, short the richer side via the perp.

What is a COMP (completeness) signal?#

Yes-side price + No-side price deviates from $1. Sub-$1 is a buy-both-sides arb that pays $1 at settle (less fees and slippage). Over-$1 is short-both-sides if your inventory allows. HIP-4 books are thin; packages are not atomic, so flag with caution.

What is an EXP (expiring) signal?#

Market expires within 24 hours and has a material divergence. Time risk dominates here — gamma is high near ATM near expiry, so a small spot move flips the outcome. Re-hedge frequency and funding cadence matter more than usual.

What is a QUAL (quality / liquidity) signal?#

Liquidity health flag. DEAD = no volume, no OI; ignore. STALE = OI sitting, no trades in 24h; quotes may not be tradable. WIDE= spread > 10pp on a market we have BBO data for. These rows are not opportunities — they explain why a market is hidden by the Tradable filter.

What is "edge"?#

The modeled gross profit, in USDH, from taking the divergence trade at a $5,000 reference notional. It assumes you can fill at the current mid on both legs — which is rarely true on thin books, hence the spread tax adjustment below.

What is "net edge" vs "raw edge"?#

Raw edge is the modeled gross at $5k size. Net edge caps the size at the available OI (10% of standing OI by default), then subtracts the round-trip spread cost on that capped notional. Net edge is what the scanner ranks by — a +$500 raw edge sitting on $200 of OI ranks below a +$120 raw edge with $5k of capacity.

What is "capacity"?#

The maximum notional you could realistically trade given current open interest. Defaults to the lesser of $5k reference size or 10% of standing OI. Used to size net edge.

What is "spread tax"?#

Round-trip cost of crossing the spread on both legs at the chosen capacity notional. nullwhen BBO data is missing — in that case we treat spread tax as "unknown" rather than zero, so the net edge tooltip shows "unknown".

Greeks & math

Binary options have surprising Greeks near ATM near expiry. The numbers are correct; the magnitudes can look extreme.

What is delta (Δ)?#

We display N(d₁) — the standard log-normal CDF, useful as a moneyness reference. It is NOT the right sensitivity for a cash-or-nothing binary; that is ∂P/∂S = n(d₂)/(S·σ·√T). Don't use this number for perp sizing — use the hedge multiplier instead.

What is gamma (Γ)?#

Rate of change of delta. Near ATM near expiry, gamma on a binary spikes — small spot moves rapidly flip the implied probability. That's why we flag "gamma is high — re-hedge frequently" on EXP signals and on hedge calc setups where the hedge multiplier exceeds 8.

What is sigma (σ)?#

Annualized volatility used in the binary pricing model. We estimate it from a single close-to-close 24h log return on the underlying perp (markPx vs prevDayPx), annualized via √365, with a 10% floor. That's a coarse proxy — not a windowed realized-vol estimate and not implied-vol — but it's the best signal exposed by the HL Info API today. Divergence, hedge multiplier, and Greeks all depend on it, so treat the model output as directional, not calibrated.

What is the hedge multiplier? Why not delta?#

For a binary contract, the dimensionally-correct sizing for a delta-neutral perp hedge is n(d₂) / (σ·√T) — call this the hedge multiplier. It tells you how much perp notional to short per dollar of outcome-contract notional to neutralize spot risk. Delta (N(d₁)) gives you the wrong number; it's the answer for a different question. Multipliers above ~8 mean the position is gamma-dominated and re-hedging frequency matters as much as the initial size.

What does "moneyness" mean?#

How far the underlying spot is from the strike, expressed as a percentage. Negative = below strike, positive = above. ATM (0%) is the regime where binary Greeks are most extreme.

What does "T-minus" mean?#

Hours remaining until the market expires and settles.

What is "overround" / why does Σ Yes+No matter?#

On an event card, Σ is the sum of Yes-side probabilities across all candidates including the fallback. In an efficient market it's ~100%; deviations indicate completeness pressure across the whole event, not just one outcome.

Provenance tags

Every Take row carries a small chip telling you where the numbers come from. Hover any tag in-app to see this in context.

LIVE#

Streamed via WebSocket. Updates as quotes change.

POLLED#

Fetched on a 15-second polling interval. May lag the order book by up to 15s.

TOP-N#

Top-of-book bid/ask present from the live BBO WebSocket channel. Quality and spread reflect live data for this market — only the top 20 markets by OI subscribe to BBO.

EST#

Derived from a model (binary-option pricing, hedge sizing, edge calc). Uses a 24h-realized vol proxy and assumes log-normal terminal spot — directionally useful, not a guaranteed PnL.

STALE#

No volume in 24h or spread is wide enough that quoted prices may not be tradable.

Tools & filters

What is the Tradable filter?#

Default-on toggle on the markets table, signals scanner, and hedging page. Hides rows whose OI, 24h volume, or spread fall below the thresholds (defaults: $500 OI, $100 24h vol, ≤15pp spread). Edit thresholds via the gear icon next to the toggle — they persist per-device. Watchlist always wins; pinned markets are never hidden by this filter.

How does the watchlist work?#

Click the star next to any market to add it. Toggle "★ Watchlist" on the markets/signals/hedging pages to filter to only your starred markets. Stored in localStorage (per-device); cross-tab updates fire automatically.

What is the hedge calculator for?#

Modeling a cross-venue hedge before you put it on. Inputs: outcome side, size, entry probability, perp leg (auto/long/short/none), hedge ratio, fees, funding cadence. Outputs: prob-weighted PnL, max gain, max loss, breakeven prob, scenario ladder (5 spot moves + settlement), PnL chart, and Greeks. Funding is annualized at HL's actual hourly cadence — not a generic 8h assumption.

Settlements

How are settled outcomes determined?#

HL's Info API doesn't expose oracle resolutions, so for price-binary outcomes we infer YES/NO from the underlying perp's 1-minute close at the expiry timestamp. If HL's actual oracle uses a TWAP window or a different observation rule, our derived result will occasionally disagree with the canonical resolution. Multi-categorical resolutions are different — those come from questions[*].settledNamedOutcomes and are authoritative.

Why is "Crowd accuracy" prefixed with ≈?#

Because the underlying YES/NO call is approximated (see above). The KPI is informational; treat the trend over time as more meaningful than any single number.

Where is settled-market history stored?#

In localStorage on this device, with a 1-year retention. Clearing browser storage loses the history; there is no backend to fall back on. A shared backend is deliberately out of scope for now.

Network & status

Is mainnet supported?#

The code is mainnet-ready — set NEXT_PUBLIC_HL_NETWORK=mainnet and redeploy. As of writing, outcomeMetareturns empty on mainnet, so the app gracefully shows "no markets" until HL enables HIP-4 there.

Where does the data come from?#

All data is fetched from the public Hyperliquid Info API (REST + WebSocket). Endpoints used: spotMetaAndAssetCtxs, outcomeMeta, metaAndAssetCtxs, candleSnapshot, l2Book, recentTrades, predictedFundings, and the allMids/bbo/l2Book/trades/activeAssetCtx WebSocket channels. No backend; all enrichment runs in your browser.

For full legal terms, see Disclaimer & Terms of Use. This site is not affiliated with Hyperliquid.