Updated 2026-09-06
Overview
Nobid is a proposed dealing desk for bonding-curve positions that have no venue. This page states what the desk would do, the one measurement the rest of the site rests on, and the claims it deliberately does not make.
What the desk is
Nobid publishes a schedule and proposes to stand on the other side of it. The population the schedule is written for is narrow and it is measured: 115,094 live bonding curves on Robinhood Chain are denominated in tokenised Robinhood equity tokens that hold no pair against WETH or USDG on the canonical factory. Their holders cannot reach ether at any price — not at a bad price, at none, because there is no venue at which to ask for one.
What Nobid proposes is ordinary in every other market and absent in this one: a published schedule, a bid denominated in ether, settled on chain in the transaction that moves the position. The terms are named, numbered and additive, so a counterparty can compute the quote before requesting it and check afterwards that the desk honoured its own book. All-in for a standard-size stranded position quoted while the underlying equity is shut comes to 682 basis points, of which 420 is the charge for warehousing an asset nothing offsets. Every term is argued in How a bid is formed.
The desk does not try to beat a market that already exists. Positions in the ether-quoted tier are declined by the schedule outright: it returns no number for them and points at the pair instead. A desk that quotes everything is a desk that is wrong about something.
The measurement it rests on
The proposition reduces to one enumeration of the Pons launchpad, taken at block 55,223,440 on 2026-09-06:
- 244,487 launches recorded by the factory at
0x7eD598BcEf8bd9Edd8C97A195C6d13f40801EC7e. - 240,350 of those curves still hold a reserve. The other 4,137 have graduated or been drained and are outside the scope of everything here.
- 125,256 of the live curves are quoted in ether and have a route out.
- 115,094 are not, which is 47.89% of the live chain.
That split is not read off a screener, because the screeners are the reason it is worth stating. The deepest curve in the survey is denominated in AMC Entertainment · Robinhood Token and reports a reserve of 1625. Read as ether — which is what every index and wallet pointed at this chain currently does — that is $3,996,184 of implied capitalisation at 2459.19 USD per ether. It is 1625 AMC: not slightly wrong, but denominated in an asset that has no bid.
Limitation
One block, one factory, one snapshot
The second fact, which is about time
A curve denominated in an equity token trades every hour of the week, because the chain does not close. The equity it is priced against trades 32.5 hours out of 168. For 80.65% of the week the reference is a price at which nobody can transact, and a desk quoting into that window carries the whole distance to the next open. That is most of the risk in the business rather than a detail of it, which is why the quote takes a named out-of-hours term instead of a quietly wider spread.
What this is not
The book is published. The desk is not open: Nobid is not quoting, holding client assets, or accepting orders. Nothing on this site is an offer, a solicitation, or investment advice.
That sentence stands at the head of every page in this documentation, in those words, on purpose. Beyond it, the specific things this site does not have:
- No track record. No trade has been quoted or settled under this schedule, so there is no fill history to show and nothing to backtest.
- No custody. The desk holds no client assets and has no mechanism to.
- No price feed. The quote takes its reference price from the caller. Nobid runs no oracle and publishes no marks.
- No per-asset census. The aggregate split is measured; which equity token each stranded curve is denominated in is not. One quote asset has been confirmed by direct call, and it is the only one this site names.
- No third parties. No audits, counterparties, partners or investors are named, because none exist.
- No volume, user or latency figures. Fabricating one would discredit the measured numbers standing next to it.
Limitation
The schedule is a position, not a backtest
How to read the rest of this
There are 8 pages, in three groups. The first two establish that the problem is real, the middle three describe the mechanism, and the reference group exists so somebody can check the claims without asking.
| Page | What it covers |
|---|---|
| Start here | |
| Overview — this page | What the desk is, what it is not, and the one measurement it rests on. |
| The stranded population | How 115,094 curves ended up denominated in an asset with no market. |
| Mechanism | |
| How a bid is formed | The published schedule, term by term, and the arithmetic that turns it into a number. |
| Settlement | What happens in the transaction, and what the desk holds afterwards. |
| Risk | Inventory, the overnight gap, and the three ways this loses money. |
| Reference | |
| API | Two read endpoints and one quote endpoint. No key, no rate limit worth documenting yet. |
| Methodology | How the survey was run, and how to run it again against the same block. |
| Changelog | Every change to the schedule and the book, dated. |
Read in order if the question is whether the desk should exist. Read Methodology first if the question is whether the numbers are real: it gives the block and the procedure to arrive at the same answer, or at a different one, which would be more useful.
Limitation
Two files, and nothing outside them