When are prediction-market contracts genuinely comparable?

Which contract and data fields must agree before two venue observations can be treated as directly comparable?

Shared topic labels are insufficient; wording, outcome structure, expiration, resolution authority, and resolution criteria must materially align.

Price disagreement can reflect different contract definitions rather than different beliefs. A controlled comparability state prevents false equivalence.

• Inspect the production adapters for Polymarket, Kalshi, and PredictIt.
• Compare the fields available for probability, bid, ask, activity, expiration, and source identity.
• Classify pairs as directly comparable, related but not equivalent, or insufficient information.

• All three adapters expose an observable probability and original-source link.
• Depth, activity units, and update cadence are not uniform across venues.
• The current production data is sufficient for related context but not for a universal claim of direct equivalence.

• A looser topic-based comparison may still be useful when clearly labeled as directional context.
• Some contracts can be equivalent even when provider field coverage differs.

• This release does not yet persist a complete resolution-rule corpus.
• PredictIt supplies fewer depth and movement fields than the other current adapters.
• No claim about market accuracy or venue superiority is made.

• A provider changes contract wording or resolution rules.
• A normalized pair is found to have different expirations or outcome structures.
• Missing source fields prevent a reviewer from confirming equivalence.