abadi

Market making for
DreamDEX Event Contracts
Shannon 50312
since 26 Aug 2026

The markets expire every window. The liquidity does not.

Abadi rests a two-sided quote inside the incumbent spread and holds no view on the outcome. Every number below was read off the chain, not modelled.

Live from the chain read in your browser, no server of ours

NAV
tUSDC
Idle
tUSDC
Resting
tUSDC
Per share

    block ·vault on the explorer · refreshes every minute

    Where the market sits ETH-0-27AUG26 · 24h tier

    0.742 — 0.772
    abadi 0.744 — 0.770
    0.000 — certain downcertain up — 1.000

    A prediction market has a bounded axis: every price is a probability between 0 and 1. At this width the entire tradable spread is three percent of the rail. Abadi's quote sits inside it — the narrow band, not the wide one.

    The book, live ours marked by owner address

    Reading the book…

    Before Abadi quoted, this market was 0.742 / 0.772 — a spread of0.030. After, it reads 0.744 / 0.769, a spread of0.025. The incumbent quoter tightened in response.

    One vault quoting inside made prices better for everyone else trading this market.

    Track record, live from the chain every vault, every episode

    Per share
    Depositors, vs. par
    tUSDC
    Realised, every closed episode
    tUSDC
    Episodes
    Complete sets
    One-sided
    Realised, cumulative, every closed episode, tUSDC

    The equity curve is not drawn, because the read behind it failed. An empty frame is not a flat month.

    vaultmarketquotedbid / askbasisbackclosed byresult

    Per share is what a depositor’s claim is worth right now, and it carries every loss the vault has taken. Realised prices every closed episode against its full basis — the escrow the pool released on a leg that never filled is read from the collateral transfers, so a one-sided fill is worth what it actually returned rather than being left out. Read from the explorer’s log API in your browser and decoded with the vault’s ABI ·

    What filled both legs taken

    Paid at quote time
    97.40 tUSDC
    Holds — up contracts
    100.000000
    Holds — down contracts
    100.000000
    Worth at settlement, either way
    100.00 tUSDC
    Locked in
    +2.60 tUSDC · 2.67%

    Both legs were buys. Neither could have filled against a seller, because no seller was involved: two opposite-side buyers crossed and the pool minted a fresh up/down pair from their combined collateral.

    That is why the position cost 97.40 rather than 100, and why it needed no inventory to begin with. Holding one of each side, it redeems to exactly 100 whichever way the market resolves.

    That last sentence stopped being a design claim on the 27th. A later window ran the same shape end to end — quoted, both legs filled, resolved, andsettle() redeemed 100.00 against a 97.60 basis. Net asset value did not move by a single unit across settlement, because a complete set was already marked at exactly what it redeems for.

    0.00
    directional exposure

    The number a market maker is judged on. Abadi never took a view, and the 2.60 is not a bet that paid — it is the spread, collected.

    Why quote instead of predict 2,422 settled markets

    49.6
    60s
    48.6
    300s
    48.2
    900s
    51.0
    1h
    51.9
    4h
    58.6
    24h
    deviation from a fair coin · dashed line is 50%pooled 49.96% · z = −0.04

    Across every tier, up won as often as a coin. Pooled over 2,422 settled markets:49.96%, four hundredths of a standard error from 50. The 24-hour tier looks tempting at 58.6% — it has 58 samples and a z of 1.31. That is noise, and reading it as signal is the mistake to avoid.

    With a coin flip and a three percent spread, crossing it costs−1.45% a contract and collecting it earns +1.45%. So Abadi collects. The strategy was chosen by the measurement, not the other way round.