The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Two Kalshi and Polymarket markets can have nearly identical headlines without settling on the same event. To match them reliably, first find plausible candidates, then compare their full rules and outcome structures, assign an explicit confidence grade, and test the executable order books at the size you intend to trade. A headline match is a lead—not proof of equivalent contracts or a fillable price advantage.
Three different things can look like a match
It helps to label a pair before deciding what to do with it:
As an Amazon Associate I earn from qualifying purchases.
- Same headline: The titles use similar words. That is enough to search further, but not enough to infer equivalent settlement.
- Related exposure: The markets respond to the same person, event, or underlying news, but their thresholds, dates, or rules differ. They may be useful to monitor together, but one is not a substitute for the other.
- Same settlement: The contracts resolve to the same proposition under sufficiently aligned rules, including outcome definitions and edge cases. Only this category supports treating the pair as interchangeable for a cross-venue comparison.
There is no shared canonical market identifier between the venues. Their market structures can differ, too: one venue may offer a single multi-outcome event while the other offers separate binary contracts or a ladder of thresholds. A correct mapping may therefore be one-to-many or many-to-many rather than a neat pair of IDs.
Find candidates without mistaking similarity for equivalence
A practical matching pipeline reduces the search space in stages. The methodology documented in the nanare-sudo/kalshi-polymarket-spreads repository starts by blocking markets by category and close date, then scores token overlap with extra weight for rare terms. It checks numbers, dates, and named entities before optional rule verification. It retains confidence and notes, and discards low-confidence matches.
#1 Best Overall
- Collect and normalize market records. Keep the venue, market and event identifiers, title, category, close or settlement timing, outcome labels, rules text, and available book data. Preserve the original text so a normalized title does not conceal a meaningful difference.
- Block to plausible candidate groups. Compare markets in compatible categories and with reasonably close dates. This limits the search without asserting that the remaining markets match.
- Score distinctive terms and verify structured details. Rare names can be useful signals, but check thresholds, dates, and entities explicitly. A text model or similarity score can prioritize review; it cannot decide what a contract means.
- Read the rules for each shortlisted pair. Record the evidence for a match, a mismatch, or an unresolved point. Keep the candidate list separate from the verified mapping.
Simple string similarity can produce confident nonsense: as the repository’s methodology puts it, “Trump” is not “Trump Jr.,” and “nominee” is not “wins election.” A title can omit a crucial threshold or describe a related event rather than the actual settlement proposition.
Collection quirks need their own checks
One author’s indexed account of building a matcher reports seeing nested Kalshi markets marked status: "active" rather than "open", and seeing liquidity_dollars as zero. The author says they gated on volume_fp plus a two-sided book, and skipped event tickers beginning KXMVE, which they identified as multivariate parlay shards. These are reported observations from that author’s collection process, not independently confirmed or guaranteed current API behavior. Treat fields and market types as things to validate against the current data you actually receive.
Verify the settlement rules before comparing prices
The rules determine what settles. Mithril’s vendor-authored technical article summarizes this as: “The resolution rules are the contract.” For every candidate, compare these terms side by side:
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minute- Exact proposition and threshold: What must happen, and what numerical cutoff applies? A forecast about reaching a value is not the same as one about exceeding it.
- Resolution source: Which source or authority determines the outcome? Different sources can report different values or interpret an event differently.
- Settlement clock: What date and time control, and in which time zone? A shared calendar date is not enough if the cutoff differs.
- Outcome definition: Map each outcome label to its precise meaning. Check whether one market’s “Yes” corresponds to the other’s “Yes,” or instead to a different outcome in a multi-outcome event.
- Edge cases: Compare treatment of ties, postponements, cancellations, and other contingencies stated in the rules.
A comparison workbench described by its author warns that title similarity plus nearby settlement time should produce a reading list, not a verdict. A threshold mismatch, different resolution source, or different clock can leave two markets correlated but non-identical. If a material rule cannot be established, mark the mapping unresolved rather than filling the gap with an assumption.
Grade the mapping and keep the reason with it
A useful mapping is not just two IDs in the same row. Store a status or grade, the reason for it, the rule evidence checked, and the date of the last review. For example, a workflow might distinguish “verified equivalent,” “related exposure,” and “unresolved”; those labels should reflect your own explicit criteria, not just a similarity score.
Keep the grade visible wherever prices or alerts are shown. Do not route an order across a pair marked related or unresolved as though the contracts shared settlement. This is a prudent safeguard against the documented risks of rule mismatches, not a claim that any particular grading scheme is an industry standard.
Compare executable books, not displayed midpoints
A quoted midpoint is not a price you can necessarily trade. Normalize the books and simulate a fill by walking the visible levels on both venues at the intended size. Compare equivalent sides of the verified contracts, include each venue’s applicable fees, and show how much of the target order the visible depth could absorb.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
- Choose the trade size and direction. A comparison is meaningful only for a specified quantity and the relevant buy or sell side.
- Use available asks for buys and bids for sells. Walk each book level by level to estimate the average execution price for that quantity, rather than relying on the midpoint or top quote alone.
- Include costs and partial depth. Account for venue-specific fees and report when displayed liquidity cannot fill the target size. Slippage can make the average fill worse than the best visible price.
- State what the estimate excludes. A snapshot cannot guarantee that displayed liquidity will remain available, and a price comparison does not by itself account for the time and operational constraints involved in moving funds.
The resulting comparison is an estimate from observed books, not a guaranteed opportunity. A gross gap can shrink or turn negative after spread, depth, slippage, and fees. Fee schedules are time-sensitive; check the venues’ current official documentation before using a fee figure in a live calculation.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Revalidate mappings as markets change
A match can go stale if a market is amended, clarified, or relisted. Recheck both the rules and the market identifiers throughout the life of a pair; do not assume an old match remains valid because the titles still look alike. Keep dated rule snapshots or auditable notes so a later reviewer can see what supported the mapping at the time. If a material term changes, pause price comparison until the pair has been reviewed again.
Build a matcher or use a unified-data service?
Building your own pipeline gives you direct control over candidate filters, rule-review records, confidence grades, and revalidation. It also means you must maintain collection, matching logic, rule checks, order-book coverage, and fee-aware fill estimates. A third-party unified-ID or market-data service may reduce some integration work, but a shared identifier alone does not prove equivalent settlement.
Evaluate either approach on whether you can audit the mapping, inspect how rules are verified, learn when records are updated, compare sufficiently complete order books, estimate fills net of fees, meet access requirements, and understand total operating cost. Vendor descriptions of products and API capabilities are claims by their providers; assess them against those criteria rather than assuming independent validation.
What one dated scan can—and cannot—tell you
The repository’s first public scan, dated August 19, 2026, reports that it loaded 1,600 markets—800 per venue—narrowed these to 312 candidate pairs, and retained 17 verified matches at its confidence threshold, with median confidence of 0.83. For that sample and method, it reports a median gross midpoint spread of $0.0045, a median executable spread of $0.0006 at a $100 size, and a negative executable spread for 47% of pairs at that size.
Those figures describe one repository’s dated scan, not the platforms generally or the opportunities available now. The project notes that its approach can miss equivalent pairs, title-based checks can miss rule differences, and order books move. Its results are useful as a concrete illustration of why filtering candidates and simulating fills matter; they are not an independent benchmark of matching accuracy or a forecast of future spreads.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
