# How Hyperliquid Stock Perps Price the Weekend: Friday Close vs Monday Open for Every HIP-3 Stock Market

> We measured every trade.xyz stock perp on Hyperliquid from Friday close to Monday open: how far each moves and how close it lands to the real opening price.

*Source: https://hyperliquidguide.com/guides/trading/hyperliquid-stock-perps-weekend-gaps*

When Nasdaq closes at 4 p.m. ET on a Friday, the stock perps on Hyperliquid keep trading. Across the stock-weekends we measured, the typical perp moved by 9 a.m. Monday, pointed the same way as the stock's opening gap of the time, and finished a median from the real open.

Those markets are deployed by [trade.xyz](/guides/trading/hyperliquid-xyz-explained) under HIP-3, so the market rules and the responsibility for them sit with trade.xyz, not with Hyperliquid. This page measures every one of them that references a US-listed stock, weekend by weekend, using the perp's own hourly prices and the stock's real Friday close and Monday open. The data was collected on . Each time we rerun the collection script, the new weekends are added to the stored ones.

**Key takeaways**

- The perp's weekend move and the stock's Monday gap were about the same size at the median: for the perp against for the stock.
- Direction is where the weekend price helps most. In of the weekends where both moved, the perp moved the same way as the stock's opening gap.
- Level is harder. At 8 p.m. ET Sunday, the end of the purely internal weekend session, the perp was closer to the real open than Friday's close was in only of stock-weekends. By 9 a.m. Monday that rose to .
- The biggest single move in the data was at over the weekend of .

> **Key takeaway:** Treat the weekend price as what traders on one order book will pay right now. It usually gets the direction of Monday's gap right. It is often off on the size until overnight pricing starts on Sunday evening, and funding accrues every hour while you wait.

[Get started](https://app.hyperliquid.xyz/join/Concept211)

---

## Weekend moves for every trade.xyz stock perp

One row per stock, sorted by how far the perp typically travels over a weekend. Each ticker links to its market page, and each row has its own anchor (for example `#weekend-gap-nvda`) so you can link straight to a single stock. "Same direction" counts only weekends where both the perp and the stock moved.

Read the last three numeric columns together. The two "vs real open" columns show how close the perp came to Monday's opening print at two different moments, and the Sunday column is almost always the wider of the two. That difference is the subject of the next section. For a longer view of how closely these same markets follow their reference during the week, see the [30-day tracking error study](/guides/trading/stock-perps-tracking-error).

---

## How the weekend price is set

trade.xyz documents three pieces of machinery that decide what a stock perp does while the stock market is closed. Everything in this section comes from its documentation, read on November 17, 2026. Where the docs are silent, this page says so rather than filling the gap.

Start with the clock. trade.xyz's [specification index](https://docs.trade.xyz/perpetuals/specifications-and-schedules/specification-index) gives each market an "External Session" and an "Internal Session". For most US single stocks, external pricing runs "24/5, from Sunday 8:05 PM ET to Friday 8:00 PM ET", and the internal session covers "Friday 8:00 PM to Sunday 8:05 PM ET". A few names, DKNG and BX among them, have external pricing only from Monday 9:30 AM to Friday 4:00 PM ET, so their internal session lasts all weekend.

Next, the oracle. According to trade.xyz's [oracle documentation](https://docs.trade.xyz/perpetuals/mechanics/oracle-price), "when external markets are open, the externally-derived fair price is transmitted as the oracle price." When external inputs are unavailable, "the oracle advances via a continuous-time exponentially weighted moving average that incrementally adjusts the previous oracle price by a fraction of the impact price difference", with a time constant of 30 minutes. So over the weekend the oracle follows the perp's own order book, slowly, and when external prices come back it "reverts to the externally derived price on the next tick." The [external price page](https://docs.trade.xyz/perpetuals/mechanics/external-price) adds that while the external market is closed, "the external price remains fixed at the external close price."

Then there are discovery bounds. The [discovery bounds page](https://docs.trade.xyz/perpetuals/mechanics/discovery-bounds) says the bounds "restrict the mark price within ±(1/max leverage) of a reference price during both external and internal pricing sessions." In the internal session that reference "is initially based on the last externally derived oracle price (e.g., the Friday close...)". The reference can be moved out a set number of times per market; once that limit is used, "the bound becomes a hard cap until external pricing resumes." The bound width and the number of resets differ by market and are listed in the specification index.

The mark price itself, per the [mark price page](https://docs.trade.xyz/perpetuals/mechanics/mark-price), is "the median of three components": the oracle price, the oracle plus a 150-second moving average of the gap between the perp's mid-price and the oracle, and the median of the best bid, best ask and last trade. HIP-3 as a framework leaves oracle prices, leverage limits and settlement to the deployer, per [Hyperliquid's HIP-3 documentation](https://hyperliquid.gitbook.io/hyperliquid-docs/hyperliquid-improvement-proposals-hips/hip-3-builder-deployed-perpetuals), which is why these rules are trade.xyz's rather than Hyperliquid's.

That is why the table carries two snapshots. The Sunday 8 p.m. ET price is the last hour of the internal session, a price set by the book alone. The Monday 9 a.m. ET price comes after roughly thirteen hours of external pricing, so by then the oracle has been following an outside feed again. trade.xyz's docs do not name the sources behind that overnight feed, so this page does not either. The gap between the two columns shows how much the overnight session moved the perp toward the real open: across all stocks, a median away on Sunday evening against by 9 a.m.

---

## The biggest weekend moves

The ten largest perp moves in the data, across every stock. "Fri 4 p.m. to Sun 8 p.m." is the part of the move made while the book alone set the price; the rest came after external pricing resumed.

### Worked example: Nvidia

NVDA makes a useful test because one weekend in the sample carried real news, covered below. Over weekends, the NVDA perp's typical move was , against a typical opening gap of for the stock. It pointed the same way as the gap on of the weekends where both moved, and its 9 a.m. price sat a median from the real open.

The outlier is the weekend of , when Nvidia's board announced its review of Jensen Huang on the Sunday. The perp fell from Friday's 4 p.m. price to 9 a.m. Monday, and of that came before 8 p.m. Sunday. The stock then opened below its Friday close. So the perp read the direction correctly and still undershot: at 8 p.m. Sunday it was above where the stock opened, and at 9 a.m. Monday it was still above. The hour-by-hour story of that Sunday, including the funding swing and the discovery bound, is in [how the NVDA perp priced the Jensen Huang news](/guides/trading/nvidia-perp-weekend-price-discovery).

Each pair of bars is one weekend: the perp's move from Friday 4 p.m. to Monday 9 a.m. ET next to the stock's Friday-close-to-Monday-open gap. On a quiet weekend the two bars look alike. When they differ, the stock opened somewhere the perp had not fully priced. Live NVDA prices, funding and the mark-versus-oracle gap are on the [NVDA perp market page](/markets/xyz/nvda).

---

## What this means for holding a stock perp over the weekend

None of this is advice about whether to hold anything. It describes what the position is exposed to between Friday's close and Monday's open, according to the data above and trade.xyz's own documentation.

Funding first. trade.xyz's [funding page](https://docs.trade.xyz/perpetuals/mechanics/funding) describes funding as "an hourly peer-to-peer payment between long and short positions" and lists no exception for closures. A position held from Friday 4 p.m. to Monday 9 a.m. pays or receives roughly 65 hourly prints, and weekend news can move the rate sharply, as the Nvidia weekend showed. How [funding rates work](/guides/trading/funding-rates-explained) covers the mechanics, and [what Hyperliquid perps actually cost to hold](/ecosystem/what-hyperliquid-perps-cost) measures it over 90 days.

Then liquidation. Inside the internal session the mark cannot leave its discovery bound, and the [discovery bounds page](https://docs.trade.xyz/perpetuals/mechanics/discovery-bounds) says "if a trader's liquidation price lies outside the active price bounds, their position cannot be liquidated while those bounds are in effect." That protection lasts only while the bounds do. When external pricing resumes, the oracle "reverts to the externally derived price on the next tick", so a gap the book had not priced reaches the mark all at once. The data shows how much the book leaves unpriced: the perp's Sunday-evening price was closer to the real open than Friday's close was in only of stock-weekends. [How liquidation works](/guides/trading/liquidation-explained) and the [isolated versus cross margin guide](/guides/trading/isolated-vs-cross-margin) explain where your own liquidation price sits.

Last, size. The table's "Largest weekend move" column shows how far one weekend has already carried each stock in this sample. If a move of that size would reach your liquidation price, the position is only safe on a quiet weekend. For the broader picture of how these contracts work and settle, see [equity perps on Hyperliquid](/guides/trading/equity-perps-guide), and for trading around the cash session generally, the [after-hours trading guide](/guides/trading/after-hours-trading-guide).

**Trade Stock Perps With a 4% Fee Discount** — Every market in this study trades on Hyperliquid through the weekend. The referral discount applies to your first $25M of volume. [Join Hyperliquid](https://app.hyperliquid.xyz/join/Concept211)

---

## Methodology

**Which markets.** The collector starts from every live market on the trade.xyz DEX, read from the Hyperliquid API with `{"type":"meta","dex":"xyz"}`. It drops anything this site classifies as a commodity, index or currency market. It then keeps a market only if Yahoo Finance lists the same symbol as an ordinary stock on a US exchange, which removes pre-IPO names, stocks listed in Korea, Japan, China or India, and ETFs. Finally, the perp has to track that listing: across the weekends measured, its Monday price must sit within a median 2% of the stock's real open, which rules out a symbol shared by two different companies. In this run that left of live trade.xyz markets. The excluded list, with the reason for each, is in the data file.

**Prices and times.** All times are New York time, worked out per date, so daylight saving changes are handled. The perp's Friday price is the close of the hourly candle from 3 p.m. to 4 p.m. ET, the last perp price at the 4 p.m. close. Hyperliquid's candles start on the hour in UTC, and 9:30 a.m. ET always falls on the half hour, so the Monday price is the close of the 8 a.m. to 9 a.m. ET candle: the nearest hourly mark that falls entirely before the opening bell. The Sunday price is the close of the 7 p.m. to 8 p.m. ET candle, the last full hour of trade.xyz's internal session for most stocks. If one of those hours had no trades, the most recent earlier candle within three hours is used. The stock's Friday close and Monday open are Yahoo Finance's official daily values.

**Which weekends.** Only Friday-to-Monday pairs with a full session on both days count, read off each stock's own trading calendar. Long weekends such as Good Friday or a Monday holiday are skipped, as are the early-close Fridays (July 3, December 24 and the day after Thanksgiving), when the stock market shuts at 1 p.m. rather than 4 p.m. A weekend where the perp and the stock disagree by more than 20%, a sign of a stock split or a unit mismatch, keeps its perp-only move and drops its comparison. None did in this run.

**Measures.** Weekend move is the perp from Friday 4 p.m. to Monday 9 a.m. Cash gap is the stock's Monday open against its Friday close. "Vs real open" is the distance between the perp at that hour and the Monday open. Medians are used throughout so one extreme weekend cannot drag a market's figure. "Closer than Friday" counts the weekends where the perp, at that hour, sat nearer the real open than the stock's own Friday close did.

**Limits.** Hyperliquid's candle endpoint returns about 5,000 hourly candles per request, so hourly history currently reaches back to April 2026. The collector keeps every weekend it has already stored, so the history grows each time it runs. Candle prices are traded prices, not mark or oracle prices. Recently listed stocks have fewer weekends, shown in the Weekends column, and their medians rest on less data.

**Sources and run date.** Perp prices: the public Hyperliquid API at `api.hyperliquid.xyz/info`, `candleSnapshot` with a 1-hour interval. Stock prices: the Yahoo Finance chart API, daily bars. Collected , covering Fridays from to . Regenerate with `node scripts/collect-weekend-gaps.js`.

**Start Trading on Hyperliquid** — Stock perps, crypto and commodities on one order book, open through the weekend. Save 4% on fees on your first $25M of volume. [Join Hyperliquid](https://app.hyperliquid.xyz/join/Concept211)
