The core of this ecosystem works. Conditional markets open, trade, and settle onchain; a program enforces the threshold; the treasury moves or it does not. What is missing is everything that normally grows around a working core: the indexing, the tooling, the standing checks, the data contracts. This post is a map of that gap — written from the experience of building a desk on top of the layer that exists, and feeling exactly where it ends.
None of it is inside information. Every input is cited at the foot of the page, and most of the gaps are visible to anyone who tries to answer a simple question about a market that resolved last month.
The shape of the gap is worth naming up front, because it is not the usual crypto complaint that the base layer is broken. The base layer is fine. What is missing is the second layer — the one that in a mature market is so ordinary nobody writes essays about it. Equities have exchanges and they have tick databases, corporate-actions feeds, surveillance desks, and vendors whose entire business is making the tape queryable. Decision markets have the exchange and almost none of the rest. Everything below is one item on the invoice the ecosystem has not yet paid itself.
What the ecosystem needs first
Research that compiles into standing checks. The best forensic write-ups in this ecosystem already read like detectors in prose. The Umbra treasury-attack analysis did not just narrate an incident — it enumerated evidence: a fresh wallet, bridge funding, accumulation trimmed to the stake threshold, stake recovered and reused in seconds, a metronome of fixed-size trades. That list is a check waiting to be compiled, which is exactly what this desk did with it. The opportunity generalises: every published investigation should ship a versioned, machine-readable indicator set alongside the prose, so any desk, DAO, or agent can run the check on the next proposal instead of re-reading the article. Our vector library — including its saved JSON form — is a working sketch of the shape: each mechanism gets a stable name, a version, ordered indicators, and the incident that produced it attached as provenance rather than as the name.
Alerts that carry their evidence. Analytics pushed to where holders already are — 01Resolved's Telegram Signals launch points that direction — is the right delivery. The missing half is alerts that state their own evidentiary basis: what was known at send time, what was not yet indexed, and which check fired. An alert that says "three of five indicators matched, wallet history incomplete" is worth more than one that says "risk detected," because the reader knows what would change the call.
APIs as the product, programs as the customer. The figures this category computes — pass edges, volume alignment, NAV against adjusted supply — have few human readers today. Their real buyers are programs: DAO tooling deciding whether to escalate a proposal, market makers pricing conditional liquidity, research desks, and increasingly agents. What programs need is not more coverage but a contract: stable identifiers, versioned schemas, dated figures, and explicit uncertainty states. Which leads to the uncomfortable one.
Correctness as the moat. Decision-market analytics is a small field, and the durable position in it is not being first or loudest — it is being the desk whose numbers are safe to build on without re-checking. That means published data lineage, dated vintages, restatements that announce themselves, and a public record of what changed between two issues of the same series. Every restatement handled loudly is an advertisement; every one handled quietly is a future dispute. We hold ourselves to that rule and publish how, and we think it is the correct competitive strategy for anyone in this business.
The engineering underneath
The event indexer is the keystone. Current market state is readable straight from confirmed program accounts — this site does it live. But the questions that matter most are historical: who traded, in what sizes, on which side, funded from where, and what the completed TWAPs printed at finalization. Those live in transaction history and finalization events, and they are only tractable with a purpose-built indexer. Almost everything else on this list — bot stripping, wallet attribution, cadence forensics, alerts with evidence — is downstream of that one system. It is unglamorous, and it is the platform. Whoever builds it well becomes load-bearing for the whole category.
The reason it does not exist yet is that it looks optional right up until the moment it is not. Live state answers "what is the market saying now," and that is enough to run a dashboard. The forensic questions only surface after something has already resolved — after a treasury has moved, after a wallet has been paid, after the window a defender needed to be watching has closed. An indexer is insurance against a class of question you can only ask too late, which is exactly the kind of infrastructure that a young market under-buys. The market that finally buys it will be the one that got burned, or the one that decided not to wait to be.
Wallet clustering as a first-class capability. MetaDAO's Ownership Score experiment surfaced the hard version of this problem: one holder, many wallets, and a score that depends on getting attribution right. The same capability that resolves a holder's score also detects coordination — stake assembled across addresses so that no single wallet crosses the thresholds a reviewer watches. One investment, two products, and no venue or desk currently ships it.
Microstructure as computed series, not narrative. The field already names the right concepts: depth thin enough for a cheap trade to move the print, outcomes that hinge on one marginal trade, prices that snap back the moment TWAP weight stops accruing. Today those live as analytical judgment inside write-ups. As stored per-market series, they become queryable, alertable, and testable against every historical market at once.
Uncertainty encoded in the data contract. The rule this site is built around — a missing input never becomes a favourable rating — belongs in API surfaces, not just in analyst discipline. A response schema with explicit withheld states is a small engineering decision that compounds: it keeps every downstream consumer honest by default.
The gap, measured
Our own vector library makes the dependency concrete. It carries four monitored vectors — recycled-stake treasury extraction, late-window TWAP outcome steering, wash-volume alignment inflation, and coordinated multi-wallet stake assembly — each with a shipped detector. The last three still withhold on every catalog and live row, because the evidence those detectors consume does not exist in indexed form anywhere in the ecosystem. That withhold is the contract working: missing inputs never become a favourable rating. Read the evidence requirements in the library JSON and you are reading an indexer backlog: per-trade history with in-window timestamps, wallet-attributed flow across both conditional markets, cross-wallet funding graphs, post-finalization price series.
That division of labour is the healthy shape of the stack. Intelligence platforms publish the picture and the primitives; independent desks turn them into adversarial checks and hold the results to their stated scopes; venues keep the settlement layer legible enough for both. Each monitored vector we publish is simultaneously a check we can run and a precise statement of what we hope someone indexes next.
The pressure to build all of this only grows as the pattern spreads. The design being copied here — a treasury spent by a market, a threshold enforced by a program, a verdict averaged over a window — is no longer a single venue's private mechanism, and every place it lands inherits the same missing layer: the history nobody indexed, the coordination nobody clustered, the parameters nobody priced before the market opened. The stack described here is not a wish list for one protocol. It is the standing infrastructure a category acquires once it decides it intends to last, and the first desk or DAO to treat it as infrastructure rather than as a research hobby is the one the rest will end up building on.