Week four was the first time Hypercall carried real stakes. The first Hypercall trading competition went live on Monday and has been running all week, which turned testnet from a feedback loop into something closer to a live environment. We spent the week reading what users actually did, shipping the features that the competition exposed as missing, and landing the first pass of RFQ on the instrument page.
Tl;DR:
- The first trading competition is live and ~4 days in. 30 active traders, ~$2.6M in combined volume, a top P&L of +$308k, a bottom P&L of -$122k, and a biggest-volume-trader-loses-the-most story that has been surprisingly instructive
- RFQ shipped on the instrument page: indicative bids from market makers, quote input synced to the order, wallet-based signing, and inline footer errors
- Competition P&L now snapshots off theoretical marks, not top-of-book
- Stat card sparklines and a new vol surface visualization dashboard
- Three new insights articles: HIP-4, an oil curve primer, and a macro vol shift piece on RWAs
- Docs site now ships in ten languages
- Mobile payoff flow polish, feedback buttons linking to the correct session replay
What The Competition Has Actually Shown Us
We talk about competitions mostly in marketing terms: leaderboard, prizes, engagement. The more interesting thing is what four days of real trading under a scoreboard has done for our read on the product.
Winners are concentrated, and not where we expected. The top three traders account for ~85% of all positive P&L on the board. That is a steeper curve than we would have guessed for a first competition. It tells us the product is legible enough for a handful of people to play it well, but the long tail is still struggling — and that is where the next round of UX work has to land.
Volume is not skill. The highest-volume account on the leaderboard traded ~$603k of notional — more than 3x the leader — and is down ~$122k. The leader, by contrast, did ~$197k of volume for a P&L of +$308k, an efficiency of 1.56. The single highest-efficiency trader did only $60k of volume and finished at +$147k. Traders who wait for the right setup are outperforming traders who are constantly in the book. It is also the strongest argument we have yet for making the efficiency leaderboard feel like a first-class citizen, not a third toggle behind P&L and volume.
The books held up. ~$2.6M of combined notional across five underlyings, with call flow running roughly 2:1 over put flow, is well above what we had previously stress-tested organically. No leaderboard resets, no P&L freezes, no mark divergences. The theoretical-mark snapshotting we landed at the start of the week is doing exactly what it was supposed to — ranks have been stable on thin books in a way they were not during internal dry runs.
Cross-asset matters more than we assumed. BTC is ~45% of recent flow, but USOIL, ETH, HYPE, and US500 each pulled enough volume to matter. USOIL in particular is punching above its weight — the oil insight article we published this week was written partly in response to how much testnet interest the underlying was getting. Giving people real non-crypto markets to trade is not a novelty, it is the thing people want.
RFQ Is Live On The Instrument Page
RFQ has been our single largest effort for weeks, and this week it started showing up in the actual product.
On any instrument page, you can now see indicative bids from market makers who are quoting that contract. They update in real time and give you a feel for where a fill would actually land, even for strikes that look dead on the order book. The quote input stays synced with the order you are building, signing runs through your wallet rather than an agent key, and errors surface inline in the footer instead of failing silently.
This is the piece that decides how new markets launch on day one. Order books are thin at the open. RFQ is how a market goes from “listed” to “tradable” the instant we turn it on.
Competition Plumbing You Can Feel
Two changes that made the leaderboard behave like a serious scoreboard:
- Theoretical marks for competition P&L. Rankings are now snapshotted from theos instead of raw order book marks. On thin books, a single stale quote could have swung a rank by tens of thousands of dollars. Using theos means your rank is a function of your position and the market, not whoever happened to be quoting at that instant.
- Historical theo fallback. For positions that opened before a theoretical mark was available, we now fall back to a sensible prior rather than silently dropping the P&L contribution.
Stat Card Sparklines And The Vol Surface Dashboard
- Stat card sparklines now appear on every asset page. A number plus a trend in the same glance is what you actually want out of a stat card, and it meaningfully changed how fast traders were orienting on a new asset.
- A new vol surface visualization dashboard shows the full surface for an underlying in one view, rather than making you read it one strike at a time. The traders who surfaced to the top of the leaderboard this week are, unsurprisingly, the ones who were already mentally building this view themselves.
Insights Expands
Three new articles shipped this week:
- HIP-4, Explained — what it is, what it unlocks for on-chain derivatives, and how binary options fit into the picture
- Why Oil Futures Roll — a primer on the oil curve and roll mechanics, pinned to a specific snapshot time so the numbers stay honest
- Macro Vol Shift and RWAs — why the stablecoin RWA trade is a vol trade in disguise
Ten Languages Of Docs
The docs site is now fully available in ten languages. Reference pages, insights articles, the options education path, and the blog all translate end to end — not a half-localized shell with English leaking through the component chrome.
Mobile Polish
- Payoff flow got tighter spacing, refined timeline controls, and better viewport sizing
- Trade button restored on asset pages where a hardcoded
canTradecheck was quietly hiding it - Feedback buttons now generate the correct PostHog session replay URL
Looking Forward
The competition closes at the end of next week. The more interesting thing is that we now have ~4 days of real behavior data from users with skin in the game, and most of our next two weeks of product work is already downstream of it: better efficiency-leaderboard framing, first-class position risk surfacing, more non-crypto markets, and the rest of the RFQ rollout across the order management flow.
Testnet has started feeling less like a dress rehearsal and more like the real thing. That was the point.