Was the weekend right?
7 of 7 Moves of 1% or more: our post was closer than Friday's close on 7 of 7. All markets: 10 of 18 (Weekend of 14–16 Nov 2026; our last weekend post before each market’s first real print, graded Mon 16 Nov 09:45 ET)
Method, in one paragraph
For the 20 largest Hyperliquid stock, index and commodity markets we compare two numbers with the reopen: Friday’s close (Hyperliquid’s oracle at the close, or its last trade when we missed the oracle, marked e) and Hyperliquid’s weekend price. The reopen is the median price 5–10 minutes after the market’s first real print: Sunday 18:00 ET for CME index and commodity futures, the Sunday 20:05 ET overnight session for US stocks, the Monday cash open for Korea, Japan and Hong Kong (and separately the Monday 09:30 ET cash open for US stocks). The weekly grade scores the last number we actually published before each reopen. The running rate below is a separate measure, the one the published studies use: the final weekend print 5 minutes before the reopen, over every market and weekend we measured. Intervals are 95% and clustered by weekend. Statistics, not a forecast.
61%
of market-weekends so far had a final weekend print (5 min before the reopen) closer to the reopen than Friday’s close (the published studies’ measure, separate from the weekly grade above): 145 of 236 market-weekends over 2 complete weekends, 95% interval 13%–100% (clustered by weekend: markets in one weekend move together). The weekend move had the reopen’s direction 72% of the time. Small sample: read it as indicative.
Context: the published studies
Context, not a contest. Each study uses its own definition, markets and period, so each sits next to our closest measure; none of these is a replication, and a gap between them doesn’t make either wrong. The dashed line is 50%: no better than Friday’s close.
-
Blockworks 0xResearch 2026-02-09 · 23 HIP-3 equity markets, 146 weekend samples
-
HPC / Messari 2026-08-03, commissioned by the Hyperliquid Policy Center · 614 market-weekends, 68 markets, 17 closures, Mar-Jul 2026, >=$1M weekend volume
-
4Pillars (Korea) 2026-06-09 · 62 weekends of Samsung, SK hynix, Hyundai, EWY
By size of the move
| Market-weekends | Closer | 95% interval | Same direction | n |
|---|---|---|---|---|
| Reopen move under 0.25% | 14% | 0–77% | 53% | 43 |
| 0.25% to 1% | 55% | 4–100% | 64% | 75 |
| 1% or more | 83% | 15–100% | 85% | 118 |
| Reopen move 0.25% or more | 72% | 31–100% | 77% | 193 |
| Weekend move 1% or more | 75% | 19–100% | 91% | 68 |
By asset class
| Market-weekends | Closer | 95% interval | Same direction | n |
|---|---|---|---|---|
| US stocks | 64.3% | 28%–100% | 72.1% | 140 |
| ETFs | 67.7% | 30%–100% | 76.5% | 34 |
| Commodities | 22.2% | 0%–100% | 50.0% | 18 |
| China & Hong Kong | 85.7% | 60%–96% | 85.7% | 14 |
| Korea | 50.0% | 0%–100% | 62.5% | 8 |
| FX | 50.0% | 0%–100% | 83.3% | 6 |
| India | 100.0% | 61%–100% | 100.0% | 6 |
| Japan | 50.0% | 0%–100% | 100.0% | 6 |
| Indices | 0.0% | 0%–49% | 25.0% | 4 |
By first real print
When real pricing comes back: US stocks at Sunday 20:05 ET (overnight session), CME indices and commodities at 18:00 ET, Korea, Japan and Hong Kong at their Monday cash open.
| Market-weekends | Closer | 95% interval | Same direction | n |
|---|---|---|---|---|
| US stocks (NYSE/Nasdaq) | 63.5% | 28%–99% | 71.9% | 178 |
| CME indices & commodities39% of these reopens moved under 0.25%, where the weekend price was closer only 14% of the time; on moves of 0.25% or more it was closer 27% (n=11). 2 weekends: a small sample. | 22.2% | 0%–100% | 50.0% | 18 |
| Korea (KRX) | 50.0% | 0%–100% | 62.5% | 8 |
| Japan (TSE) | 50.0% | 0%–100% | 100.0% | 6 |
| Hong Kong (HKEX) | 83.3% | 0%–100% | 83.3% | 6 |
| Shanghai STAR | 87.5% | 0%–100% | 87.5% | 8 |
| India (NSE) | 100.0% | 61%–100% | 100.0% | 6 |
| US stocks vs the 09:30 ET cash open | 65.2% | 55%–74% | 67.4% | 89 |
Weekend by weekend
- 2026-11-14 58%68/118
- 2026-11-07 65%77/118
Each weekend links to its posts on the weekend tape.
IPO day
When a company with a pre-IPO market on Hyperliquid lists, we compare the last Hyperliquid mark with the IPO open. No pre-IPO company has listed yet. IPO-day scorecard
Cite or check these numbers
- Frozen values: every Monday grade has its own page with the numbers as they were posted, which never change afterwards (latest: 2026-46).
- Rows as CSV: every graded row of our weekend posts, all weekends · 2026-46. Columns: Friday’s close and how it was taken, our posted weekend price and post, the first real print, the move, and the grade. The market-weekend rows behind the running rate: Friday’s close, the final and T−15 weekend prices, the reopen, the move and whether the weekend price was closer, one row per market and weekend.
- Data as JSON:
/api/tape/scorecard.json - How to cite: “KerbTape weekend scorecard, weekend 2026-46, <page address>, retrieved <date>.” Source data: public Hyperliquid markets.
Press contact: [email protected].
Definitions and caveats
- C
- Friday's close: Hyperliquid's own price at the venue's close time (oracle when captured, else its last trade); the study measures every market the same way
- WF
- final weekend price: the weekend print nearest 5 min before real pricing returns
- W15
- weekend price 15 min before real pricing returns
- R
- median price 5–10 min after real pricing returns (the reopen)
- closer_final
- |WF − R| < |C − R| (ties not counted)
- closer_15
- |W15 − R| < |C − R|
- dir
- sign(WF − C) = sign(R − C) (zero moves not counted)
- reopen
- US stocks: Sunday 20:05 ET overnight session; CME indices/commodities: Sunday 18:00 ET; Korea/Japan/Hong Kong: the Monday cash open (or Hyperliquid's first external session)
- cash_open
- US stocks only: median 5–10 min after the Monday 09:30 ET cash open
- Statistics, not a forecast. Markets within one weekend move together, so the effective sample is closer to the number of weekends than to the number of market-weekends.
- Intervals are 95% and clustered by weekend (markets in one weekend move together), so they are wider than a per-market Wilson interval.
- Rows measured from candles coarser than 15 min are left out (0 rows).
- Weekends with fewer than 10 scored markets are listed and left out of every headline number.
- External studies are context, not a contest: each uses its own definition, sample and period, and sits next to our closest metric, not as a like-for-like replication. The HPC/Messari study was commissioned by the Hyperliquid Policy Center.