A clean comparison of Polymarket and Kalshi trading volume needs more than two dollar figures pulled from a homepage. The two venues count volume on different units, settle in different currencies, sit under different regulators, and concentrate their liquidity in different categories. Once you normalise for those, the gap that the press loves to dramatise shrinks, and the gap that traders actually care about, which is depth on the markets you want to trade, becomes visible. This piece walks through how each side counts, where the comparable numbers genuinely diverge, and what rising or falling volume actually tells a working trader.
Why the headline number does not tell the story
Open any quarterly recap of prediction markets and you will find a chart with two bars: one for Polymarket, one for Kalshi, sometimes a third for a smaller venue. The chart is almost always misleading. Not because anyone is lying, but because the underlying volume figures are not measuring the same thing. One side reports notional dollar value of every share traded. The other side reports something closer to filled contract value with separate buckets for elections, sports, and recurring economic markets. The bars are drawn at the same scale, which implies they are denominated in the same unit. They are not.
The deeper problem is that volume on a binary outcome market behaves nothing like volume on an equity. A million dollars of notional volume in a market that resolves between 92 cents and 98 cents represents a very different amount of risk-bearing activity than a million dollars of notional in a market trading at 45 cents with two weeks until resolution. Comparing the two as if they were the same is the kind of mistake that makes for clean headlines and bad analysis.
For a working trader, the headline number is rarely the right metric anyway. What matters is depth on the specific market you intend to trade and how that depth changed in the last seventy-two hours. We covered the mechanics of measuring that on Polymarket in the volume explainer, and the same logic applies on Kalshi with adjustments for how their order book exposes resting size.
How Polymarket counts volume
Polymarket reports volume as notional dollar value of shares traded, with each filled trade counted once on the buyer side. A market in which a thousand YES shares change hands at 40 cents adds 400 dollars to the volume number. The figure is denominated in USDC, settles on Polygon, and is available per market and aggregated across the venue through the public Gamma and CLOB endpoints. There is no double-counting of the matching opposite NO shares, because on a binary outcome market the YES and NO are economic mirror images and counting both sides would inflate the figure by a factor of two without representing additional activity.
What the headline figure does include is volume from market-maker flow, from prop traders running directional positions, and, by every credible third-party analysis, a non-trivial slice of wash activity that the venue has limited ability to filter. The wash component is the topic of a separate piece on wash trading on Polymarket, and the short version is that the headline number should be discounted before any cross-venue comparison.
One quirk worth knowing: Polymarket volume spikes hard around resolution windows. The hour before a major election market closes can see ten times the average hourly volume of the prior week. Quarterly volume figures that average over a quiet stretch followed by a single resolution spike are not a faithful picture of the venues activity. They are a picture of one event.
How Kalshi counts volume
Kalshi is a CFTC-regulated designated contract market, which constrains the reporting in ways that matter. Volume is reported as filled contract count and notional dollar value, with each contract being a binary cash-settled instrument denominated in US dollars and worth one dollar at resolution. A contract that trades at 40 cents has a notional of one dollar but contributes 40 cents of premium paid, and depending on which figure a chart is using you will see very different numbers. The official CFTC designated contract market filings are the canonical reference for how Kalshi is required to report activity to the regulator.
Settlement is in fiat dollars from a regulated brokerage account, not USDC on a chain. There is no wallet, no on-chain trail, no public ledger of every fill. Kalshi has its own internal order book and trade tape, and the trade tape is published per market in roughly the same level of detail that an equities venue would publish. The participant set skews more toward retail dollars from US-based traders who cannot access the offshore venues, and there is a meaningful institutional slice that has grown since the CFTC settled the election market eligibility question.
Category mix on Kalshi is broader than Polymarket in some directions and narrower in others. Recurring economic indicator markets, where the same release prints monthly and traders can roll positions, give Kalshi a steady baseline that Polymarket lacks. Elections, when they are running, contribute heavy peak weeks but very little baseline volume between cycles. Sports remained limited on Kalshi until comparatively recently and the category is still finding its level.
Where the gap is real and where it is reporting artefact
If you take both venues quarterly notional figures at face value during a non-election quarter, Polymarket usually reports a multiple of Kalshis figure. If you do the same comparison during a US election quarter, the multiple compresses sharply because Kalshis election volume becomes a much larger share of its total. None of those raw multiples are meaningful without three adjustments.
The first adjustment is for the wash component. Independent analyses of on-chain Polymarket flow have repeatedly found that somewhere in the range of fifteen to thirty per cent of headline volume on certain categories looks more like circular self-trading than like economic activity. Kalshi, as a regulated venue with mandatory KYC, has a structurally smaller wash component, although it is not zero. Apply a haircut to Polymarkets figure and the gap closes by roughly that amount.
The second adjustment is for resolution-week concentration. Polymarket volume is more peaky than Kalshi volume because elections and binary news events dominate. A trailing-thirty-day window that includes a major resolution looks dramatically larger than one that does not. Trailing-ninety-day windows smooth this out and are the right reference period for a venue-versus-venue comparison.
The third adjustment is for what is actually tradeable. A venue can post a large notional figure and still have markets that are too thin to absorb a five-thousand-dollar order without three cents of slippage. Volume that lives in fifty markets is operationally different from the same volume spread across five hundred. Both venues skew toward concentration, but the shape of that concentration is different.
Illustrative monthly notional by category — not a real-time snapshot
Category-by-category comparison
The table below is the comparison I actually use when a friend asks which venue to trade a given thesis on. It is dimensional rather than numeric, because the dollar figures move week to week and any number I write here will be wrong by the time someone reads it. The dimensions, however, are stable.
| Dimension | Polymarket | Kalshi |
|---|---|---|
| Counting method | Notional, buyer side only, per filled trade | Filled contract count plus notional, regulator-reported |
| Settlement currency | USDC on Polygon | US dollars, regulated brokerage |
| Regulator | None in the US; offshore for US persons | CFTC designated contract market |
| Dominant categories | Elections, crypto, news, sports growing | Economic indicators, elections, sports broadening |
| Peak volume context | Resolution windows of major news and elections | Election cycles and high-impact macro releases |
| Fee model | Zero maker and taker on most markets | Per-contract fee, tiered by volume |
| Wash exposure | Higher; mitigated but visible on-chain | Lower; KYC and regulated reporting reduce surface |
| Order book transparency | Public CLOB endpoints, on-chain mirror | Internal book, trade tape published per market |
Notice that the fee row alone changes the read on volume. A venue that charges per-contract fees and still posts a given notional has a different underlying participant set than a venue that charges nothing and still posts a comparable figure. The free-trading venue is in part subsidising activity that would not happen on a fee-charging venue, and the fee-charging venue is in part filtering activity that would not survive its economics. Neither is bad. They are different products.
The wash trading wrinkle
No honest comparison can skip this. Polymarket has a structurally larger wash trading surface than Kalshi for one simple reason: anyone can spin up a new wallet without identity verification and trade against themselves. On-chain analysis tools can spot the more obvious patterns, and the venue has improved its monitoring, but a baseline level of circular flow is impossible to eliminate without breaking the permissionless design.
Kalshi, as a CFTC-regulated venue, requires KYC, ties accounts to verified identities, and reports activity in a way that subjects wash patterns to regulatory scrutiny. That does not mean wash activity is zero. It means the surface is much smaller and the cost of attempting it is much higher.
For a trader comparing the two on volume terms, the practical adjustment is to mentally discount Polymarkets headline by an amount that reflects best-estimate wash share for the category you care about. Crypto markets and certain news markets carry a heavier wash component than political markets, which are watched more closely. The full breakdown of how to spot it lives in the wash trading writeup.
What rising or falling volume actually signals
Volume going up on either venue is not automatically bullish for the venue or for the trader. The interpretation depends entirely on what is going up and why. A spike concentrated in one market the week before its resolution is a feature, not a trend. A steady rise across many markets and many counterparties over a quarter is a trend, and a more interesting one. Distinguishing the two requires looking past the homepage banner and into the per-market trade tape.
There are four patterns worth recognising. The first is event-driven peak: a single market or cluster of related markets spikes by ten or twenty times average daily volume in the seventy-two hours before resolution, then returns to baseline. This is healthy and expected on any prediction market and tells you almost nothing about secular venue health. The second is category drift: volume shifts from one category to another over a quarter, for instance from elections to crypto-news or from sports to macro. This usually reflects shifting trader interest and venue product decisions, and rewards close reading.
The third pattern is depth widening without notional change. The book gets thicker at every price level even though aggregate notional is flat. This is the most important pattern for a working trader because it directly improves the cost of execution. The fourth is the opposite: notional spikes while depth thins, which happens when one or two large participants are crossing the spread aggressively. That pattern looks great in a press chart and terrible if you are trying to put on a five-figure position.
For cross-venue arbitrageurs, the right question is not which venue has more volume in aggregate. It is which venue has deeper books on the specific market you are trying to mirror, and how quickly the spread between the two venues prices closes when news drops. We covered the operational side of that work in the Polymarket Kalshi arbitrage bot guide, and the short version is that volume is a poor proxy for the depth and latency profile an arbitrageur actually needs.
Volume is a vanity metric until it is depth at the price you want to trade. A venue that posts a billion dollars of notional across markets you cannot execute in size on is less useful than a venue that posts a quarter of that figure with twice the depth on your three markets.
The traders honest takeaway
If you are choosing between the two venues to trade a specific thesis, the headline volume figure is not the input. The inputs are jurisdiction, settlement preference, the category mix, the depth on the specific market, and the fee model. Polymarket is the answer for permissionless access, USDC settlement, deeper books on most election and crypto news markets during peak windows, and zero per-trade fees. Kalshi is the answer for US-based participants who need regulated venue exposure, fiat settlement, steady recurring economic indicator markets, and a participant set that institutional desks can clear into.
If you are writing about the two venues, the honest framing is that they are increasingly complementary rather than directly substitutable. The overlap is real on elections and growing on sports, but each venue has a tail of markets the other does not seriously contest. A press chart that draws one bar against the other and lets the reader infer winner and loser is doing a disservice to both venues and to anyone trying to learn from the numbers.
If you are an arbitrageur, the comparison is operational. Which venue has the API throughput your bot needs. Which one settles fast enough to redeploy capital before the spread closes. Which one charges fees that your strategy can absorb. Volume figures barely enter the calculation; latency, depth, and settlement do.
And if you are a casual reader who saw the chart and wanted to know which venue is winning, the answer is the unsatisfying one: the question is wrong. Both venues are larger than they were a year ago. Both have product roadmaps that are bringing them closer to feature parity in some areas and further apart in others. The market for prediction-market venues has room for two well-run businesses with different regulatory postures, and the volume gap that draws headlines is in large part an artefact of how each side counts.
About the author
Maria Ostrowski spent eight years on a systematic equities desk before moving full time to prediction markets in 2023. She runs a small statistical arb book across Polymarket, Kalshi, and a handful of smaller venues, and writes about the parts of the data that get lost when venues are compared by headline number. Maria reviews the data-category posts on this site and answers reader questions on methodology.
Frequently asked questions
Does Polymarket or Kalshi have higher trading volume?
It depends on the period and the category. In a typical non-election quarter, Polymarket reports a multiple of Kalshis figure on a raw notional basis. Adjusted for wash trading exposure, resolution-window concentration, and what is actually tradeable in size, the gap is smaller and varies by category. In economic indicator markets, Kalshi is the deeper venue.
Why do Polymarket and Kalshi volume figures use different units?
Polymarket reports notional dollar value of filled trades in USDC on Polygon, counting the buyer side once per fill. Kalshi reports filled contract counts and notional in US dollars under CFTC designated contract market rules. The two figures are denominated differently and computed under different reporting regimes, which is why side-by-side bar charts of raw numbers are usually misleading.
Is Polymarket volume inflated by wash trading?
A meaningful share of headline volume on certain categories looks more like circular self-trading than economic activity. Independent on-chain analyses have estimated this share in the range of fifteen to thirty per cent for the more affected categories. Kalshi, as a KYC-required regulated venue, has a structurally smaller wash component.
What is the right time window to compare Polymarket and Kalshi volume?
Trailing-ninety-day windows smooth out the resolution-week spikes that distort shorter windows. Even then, splitting the comparison by category gives a much clearer picture than a single aggregate number, because the venues concentrate their liquidity in different places.
Does higher trading volume mean better fills for my orders?
Not necessarily. Aggregate venue volume is a poor proxy for depth on the specific market you want to trade. A venue can post a large notional figure and still have thin books on the markets that matter to you. Always check the order book depth on the specific market before sizing up, regardless of venue-level volume figures.