Matching Markets Across Venues Without Guessing
Before you can compare two prices you must establish that the two contracts are about the same thing. Structural matching, and why title similarity fails.
Comparison is the easy half. The hard half is establishing that two contracts are about the same thing, and it is where every cross-venue system either earns its credibility or loses it.
Why title matching fails
Titles are written for humans. Two questions can read almost identically and differ in the source, the vintage, the threshold direction or the deadline. Two questions can read completely differently and be exactly equivalent.
Text similarity is useful for generating candidates. Used as a decision rule it produces confident, well-presented nonsense, which is worse than producing nothing at all.
What structural matching looks at
Subject identity. The same measured quantity from the same publisher. Not "inflation" but a specific named series from a specific agency.
Threshold and direction. The same number, and the same comparison, with strict versus inclusive noted separately because it only matters on ties and matters absolutely then.
Observation moment. The same time, in the same zone, with the same rule for late publication.
Outcome structure. Binary against binary, or a defined subset of a multi-outcome set against a binary. Anything else needs explicit handling. Multi-outcome markets.
Resolution mechanism. Not required to be identical, but recorded, because it feeds the grade. Resolution differences.
The pipeline
Generate candidates. Text similarity, shared entities, overlapping date windows. Cast wide; this stage is allowed to be noisy.
Normalise. Convert both sides into a common representation: publisher, series, threshold, direction, timestamp, structure. This is where most of the real work is, and it is worth writing as tested code rather than as assumptions. Normalising across venues.
Compare structurally. Field by field on the normalised form.
Grade. Produce a comparability score with the reasons attached, rather than a boolean. Equivalence grades.
Refuse when uncertain. If structure cannot be established, do not pair. A missing row costs a user nothing. A wrong row costs them their trust in every other row.
Why the output is small
A naive matcher on two venues will happily produce thousands of pairs. A structural one produces dozens. That ratio is not a failure of the matcher; it is an accurate description of how little genuine overlap there is between two venues that list different things for different audiences under different rules.
Then apply liquidity filters and the list shrinks again. Gap size versus liquidity.
Showing the reasoning
A pairing should be able to explain itself: here is the publisher on each side, here is the threshold on each side, here is where they differ, here is why the grade is what it is.
Without that, a user has to trust the matcher completely or ignore it completely, and neither is a good outcome. This is why the grade carries its reasons and why the methodology is written down rather than described as proprietary.
The limit worth stating
Automated matching is a filter, not an authority. It narrows thousands of markets to a handful worth reading the rules on. For anything you would act on, read both sets of rules. The matcher exists to tell you which two documents to open.
Next
How to read a gap table without being misled by it.
Educational content only. Nothing here is financial advice, a solicitation, or an offer to trade. Prediction markets can move against you and a position can lose its entire value. Where this guide describes a difference between venues it describes a theoretical cross-market edge in observed prices, which is not a profit: fees, spreads, capital lockup, transfer time and differing settlement rules all come out of it first.
Keep reading
Cross-Market Gaps · 3 min readStale Quotes and Why Timing Ruins ComparisonsThe largest source of phantom cross-market opportunity is not disagreement. It is one price being older than the other. How to measure and exclude it.
Cross-Market Gaps · 3 min readHow to Read a Cross-Market Gap TableA gap table sorted by size is a list of artefacts. Here is the order to read one in, what each column means, and which rows to ignore first.
Cross-Market Gaps · 3 min readTen Ways Cross-Market Analysis Goes WrongThe recurring errors in comparing prediction market prices across venues, from stale quotes to mismatched thresholds, and how to catch each one.
Order Books and Liquidity · 3 min readDepth, Slippage, and the Price You Actually GetThe quote is good for the size at the top of the book. Beyond that you walk the book, and the average price you pay is not the one you saw.