Recurring Prediction Market Contracts: How to Read a Chain, Not a Single Bet

TL;DR

• A recurring prediction market contract (a weekly approval tracker, a monthly CPI print, a per-meeting Fed rate contract) resolves independently each cycle, then reopens at a fresh price that doesn't inherit the prior close.
• Reading one instance in isolation tells you what the market thought about that specific cycle, not where sentiment is headed. The trend only shows up across a full chain of instances.
• A clean chain requires the same question and threshold, comparable liquidity, and the same resolution source across every instance. Without that, cross-instance or cross-platform comparisons are misleading.
• Polymarket, Kalshi, Limitless, Predict.Fun, and Opinion each run their own version of recurring contracts, so aggregating the same question across all five is what turns single snapshots into an actual time series.

Ever watch a Fed rate contract close at 70% and assume the market's turned hawkish, only to see next month's contract open cold at 30%? You're not misreading the data. You're reading one link in a chain and mistaking it for the whole picture.

A recurring contract is not one market. It's a series. Each week's Fed rate contract, each month's inflation print, each cycle's approval-rating tracker resolves on its own, then a new instance opens at a fresh price. Treat any one of them as a standalone number and you'll miss the story. Treat them as a chain and you get a time series that shows whether sentiment is building, fading, or stuck.

Why Do Some Contracts Reset Instead of Running Once?

Most prediction market contracts are one-shot. Will a specific bill pass. Will a named candidate win a named election. The question has one answer and the contract dies the moment it does.

Recurring contracts exist because some questions repeat on a fixed calendar. The Fed meets eight times a year and each meeting gets its own rate-decision contract. The Bureau of Labor Statistics publishes CPI monthly, and each print gets its own inflation contract.

A president's approval rating gets measured every week. Some platforms run a fresh contract for each measurement window rather than one contract that drifts forever.

The reset is structural, not arbitrary. A single Yes contract can't answer "will the Fed cut in September" and "will the Fed cut in December" at the same time, because those are different events with different information sets. Bundling them into one open-ended contract would blur exactly the thing traders want to isolate: what does the market think about this specific meeting, this specific print, this specific week.

How Does Pricing Reset Each Cycle?

When a recurring contract resolves, the platform closes the books on that instance and opens a new one at a new starting price. The new price is not inherited from the old one. It reflects whatever information exists at that moment: the last print, the latest Fed commentary, the newest polling.

Say a monthly inflation contract resolves Yes at $0.71 (a 71% implied probability that CPI came in above a given threshold). The next month's contract on the following threshold doesn't open at $0.71. It opens wherever traders price the new question, which might be $0.30 if the prior print already cooled expectations, or $0.85 if inflation has been running hot for three straight cycles.

Each instance prices its own event. The prior result is an input to that pricing, not a starting point for it.

This is the part researchers most often get wrong. They watch one instance close high and assume the "market" is bullish going forward. What actually happened is that one specific cycle resolved. The next cycle is a new market with its own supply of contracts, its own liquidity, and its own information set.

How Do You Read a Chain as a Time Series?

The value in recurring contracts isn't any single price. It's the sequence. Pull the last six or eight instances of the same recurring contract and you're looking at how consensus probability moved across real time, cycle by cycle.

A tracker that's opened each cycle at 40%, 45%, 52%, 58% is telling you something a single snapshot never could: momentum is building in one direction, cycle over cycle.

Three things to check before you treat a sequence as a real trend:

• Same question, same threshold. A rate-cut contract for a 25 basis point cut and one for a 50 basis point cut are not the same series, even if they run on the same Fed meeting calendar. Compare like to like or the chain is meaningless.
• Comparable liquidity across instances. A cycle with a few thousand dollars of open interest and a cycle with six figures don't carry the same signal weight. Thin instances swing on small positions and can look more decisive than they are.
• Consistent resolution source. If the underlying data source changes mid-series (a different polling aggregator, a revised BLS methodology), the chain breaks even if the question text looks identical.

If you're not sure how to judge whether a given instance's liquidity is thin or deep, our explainer on market volume walks through reading open interest correctly before you trust a single print.

Once you've confirmed the series is clean, the read is straightforward. Rising consensus probability across consecutive cycles means the market is converging on an outcome. A series that oscillates without a clear direction means the market is genuinely uncertain, cycle to cycle, and any single high or low print is noise rather than signal.

Why Does Cross-Platform Comparison Get Harder With Recurring Contracts?

Comparing odds across platforms is already useful for one-shot events, and our breakdown of why odds disagree across venues covers the mechanics behind that gap. It gets more valuable, and more error-prone, with recurring contracts, because the platforms don't always structure the same recurring question the same way.

Polymarket and Kalshi might both run a monthly CPI contract, but Kalshi's version could resolve against a specific BLS release time while Polymarket's resolves against a broader "as reported" standard.

Limitless tends to run shorter-duration recurring markets, which means its weekly instances update more often but carry less liquidity per cycle than a monthly contract elsewhere. Predict.Fun and Opinion both serve crypto-native traders and often price the same recurring question a few points off the US-heavy platforms, because the trader base skews differently.

None of that makes any platform wrong. It means the instance-to-instance comparison only works if you're matching the same threshold, the same resolution window, and the same underlying data source across platforms, not just the same headline question.

Why Does This Matter for Long-Term Tracking?

A journalist tracking approval rating over a presidential term, or a researcher building a six-month view of rate-cut expectations, isn't looking for a single number. They're building a dataset. Recurring contracts are built for exactly that, provided you pull the full chain and not just the most recent instance.

The mistake to avoid is treating the latest cycle's number as "what the market thinks," full stop. It's what the market thinks about that cycle, given what was known at that moment. String enough of those together and you have something closer to the truth: not a prediction, but a documented path of how consensus moved as new information arrived.

Polymarket, Kalshi, Limitless, Predict.Fun, and Opinion each run their own version of recurring contracts, with different cadences and different resolution mechanics. Aggregating across all five is what turns a pile of individual snapshots into an actual time series.

---

PredictionHero aggregates publicly available prediction market data for informational purposes only. This is not financial advice. Prediction markets may not be available in all jurisdictions.