Prediction markets are exchanges where people trade contracts on future events, and rising trading volumes alone do not prove that their forecasts are reliable or their growth is sustainable.

The reported boom involves Polymarket, Kalshi and growing interest from established betting and financial brands. For you as a reader, the important distinction is between more trading, more participants and better predictions. Each requires different evidence.

The figures below preserve the estimates discussed in the original article, while separating reported claims from verified conclusions. That article attributed its data to Keyrock and Dune, with additional Dune Analytics dashboard insights maintained by Gate Research, but supplied no report links, observation dates or calculation methods. Treat those figures as claims requiring source checks, rather than independently verified market statistics.

prediction markets

What is driving the growth in prediction markets?

Prediction markets attract activity around politics, sports, economic announcements and technology milestones because these events produce questions with tradable outcomes. A sustained expansion would require participation and liquidity to persist between major events, rather than appearing only during headline-driven spikes.

A contract typically specifies an outcome, a deadline and a settlement rule. Traders buy or sell according to their assessment of the outcome’s probability. New information can change prices throughout the contract’s life, giving observers a continuously updated forecast.

Politics can create sharp bursts of attention. Sports provide a recurring calendar, while inflation releases, policy decisions and technology announcements offer further opportunities for event contracts. These categories have different settlement requirements and should not automatically be combined when assessing forecasting quality.

The original article associated an October 2026 surge with US presidential elections. That explanation is incorrect because no regular US presidential election is scheduled for 2026. Its references to late 2026 followed by early 2026 also fail to establish a usable chronological sequence. Those issues prevent a confident account of when the claimed acceleration occurred.

For prediction markets, the useful question is therefore narrower than whether a boom is happening. You need to establish which activity increased, over what dates, on which platforms and under which measurement rules.

Reading the reported market figures correctly

The original article reported more than $44 billion in total notional volume and an 85% to 90% concentration on Polymarket and Kalshi. These estimates describe claimed trading activity, but their timing and methodology need confirmation before they can support a market-wide growth conclusion.

This table keeps the main reported numbers together and explains what you should check. It is a record of the article’s estimates, not a fresh data release or a verified ranking of prediction markets.

IndicatorReported estimateEssential qualification
Total 2026 notional volumeMore than $44 billionConfirm the reporting period and included platforms.
Combined platform concentration85% to 90%, or about $38 billion to $39 billionUse the same volume definition for both platforms.
Polymarket volumeAbout $21.5 billion, January to November 2026A partial-year estimate is not an annual total.
Kalshi volumeAbout $17.1 billion, January to November 2026Check comparability with Polymarket’s measurement.
Starting monthly volumeAbout $100 millionThe original also described this as below $100 million.
Election-related monthly spikeMore than $4.5 billionThe original election explanation contains a calendar error.
Reported quieter monthly base$1.5 billion to $2 billionThe stated post-election period needs identification.
November 2026 monthly volumeMore than $13 billionOne month cannot establish a lasting baseline.
Daily and weekly activityAbout $701.7 million daily, with weekly totals exceeding $5 billionRecord claims need dated source observations.
TransactionsAbout 240,000 to more than 43 millionConfirm whether both observations cover equal periods.
Monthly active usersAbout 4,000 to more than 600,000Accounts, wallets and people are different measures.
Average Brier scoreAbout 0.09Interpret against the sample and forecast timing.
Combined open interestAbout $3.3 billion to nearly $13 billionOutstanding exposure is different from turnover or deposits.

The original described monthly volume growth as approximately 130 times. Moving from $100 million to $13 billion is consistent with that rough ratio, but arithmetic consistency does not validate the underlying observations. Both endpoints still need comparable dates and coverage.

Volume is not the amount of new money deposited

Trading volume counts activity during a period. The same contract can change hands repeatedly, so substantial turnover can occur without an equivalent increase in customer deposits. Prediction markets with active short-term traders can therefore show large volumes relative to the capital committed.

See also  EFL Championship Round-Up: Sheffield United Open Gap at the Top, Baggies Earn Draw at Norwich City and Cardiff City Sack Morison

For cash turnover, a common calculation sums execution price multiplied by contract quantity across trades. Some reports instead use face-value notional measures. Multiplying a price by the number of trades is insufficient when trades contain different quantities. Check the dashboard definition before comparing platforms.

How to distinguish lasting adoption from temporary spikes

Lasting growth in prediction markets should appear across repeat participation, liquidity and activity outside major events. Transaction counts and downloads provide clues, but neither establishes that more distinct people are participating regularly.

The original described transactions rising from roughly 240,000 to over 43 million and monthly active users growing from around 4,000 to over 600,000. Those changes would be substantial if measured consistently. Automated trading, multiple accounts and changes in dashboard coverage can nevertheless affect the interpretation.

Application downloads for Kalshi and Polymarket were also cited as evidence of momentum. Downloads measure interest in an application, not funded accounts, completed trades or retained users. Without a dated series, the claim that downloads show no slowdown cannot be assessed.

  1. Match reporting periods. Compare full months with full months, and avoid treating a partial month as a decline.
  2. Separate unique users from activity. Establish whether the dataset counts people, accounts, wallets or trading addresses.
  3. Check repeat participation. Returning participants provide stronger evidence of retention than registrations alone.
  4. Examine quiet periods. Look for activity between elections, tournaments and major announcements.
  5. Inspect individual contracts. Check whether useful liquidity extends beyond a few popular events.
  6. Review incentives. Temporary promotions can increase activity without establishing durable demand.

Open interest adds another perspective because it measures outstanding positions rather than completed turnover. However, the reported increase from $3.3 billion to nearly $13 billion does not establish customer confidence or longer holding periods by itself. Contract duration, collateral arrangements and valuation methods matter.

A healthier assessment of prediction markets combines these indicators. High volume with thin order books or weak retention tells a different story from steady participation across many well-defined contracts.

Polymarket and Kalshi: concentration and category differences

Polymarket and Kalshi dominate the original article’s reported figures, but their category mixes differ substantially. Concentration can bring more buyers and sellers together while increasing dependence on a small number of operators and settlement systems.

The stated January-to-November estimates were approximately $21.5 billion for Polymarket and $17.1 billion for Kalshi. Together, these equal $38.6 billion, consistent with the original’s rounded $38 billion to $39 billion description. Whether they represent 85% to 90% of the sector depends on the total used and its coverage.

Reported transaction activity

  • Polymarket was credited with 95 million trades, or roughly 54% of the total. Its monthly activity reportedly rose from about 45,000 to roughly 19 million, described as a 421-fold increase.
  • Kalshi was credited with 74 million trades, or nearly 42% of the total. Monthly transactions reportedly increased from approximately 196,000 to around 21 million, described as 106-fold growth.

Those growth multiples are approximate and sensitive to rounded starting values. The cumulative trade totals also should not be compared directly with the separate 43 million transaction figure unless its measurement period is clear. For prediction markets, an executed trade, an order and an on-chain transaction may be different units.

Reported sports and political exposure

Sports reportedly accounted for about 85% of Kalshi’s notional volume. Polymarket’s reported mix was 39% sports, 34% politics and 18% cryptocurrencies, together accounting for 91%, consistent with the original’s description of more than 90%.

These shares help explain why different events might affect each platform differently. They do not establish that one has better forecasts. A sports-heavy platform and a politics-heavy platform can have very different resolution times, uncertainty levels and participant behaviour.

For UK readers familiar with betting exchanges, the shared feature is trading against other participants. However, contract design, legal classification, commission and settlement rules can differ. Similar mechanics do not imply identical consumer protections.

How accurate are prediction markets?

Prediction markets can provide useful probability estimates, but accuracy depends on which contracts are measured and when forecasts are recorded. The reported average Brier score of about 0.09 does not, by itself, prove superiority over polls or sports betting markets.

Headline accuracy asks whether the most likely outcome occurred. The original cited overall rates of 90% to 95%. That measure can be misleading when a sample contains many easy questions, such as events whose outcomes are almost certain shortly before settlement.

See also  Top Picks for Fast Withdrawal Casinos in 2026

A forecast can also be reasonable when the favoured outcome loses. If an event has a genuine 70% chance, the alternative should still occur about 30% of the time across comparable forecasts. A single losing favourite does not establish that the forecast was poor.

What a Brier score measures

For a binary event, the Brier score is the squared difference between the forecast probability and the outcome, recorded as one for occurrence or zero for non-occurrence. Lower scores indicate smaller forecasting errors. A perfect set of predictions scores zero.

A constant 50% forecast produces a score of 0.25 under this binary convention. Calling that the score of any random forecast is too broad, because randomly generated probabilities can produce different results. The original’s 0.09 estimate needs a comparable benchmark and a defined sample.

To illustrate the calculation, a 70% forecast scores 0.09 if the event happens, because the difference is 0.30 and its square is 0.09. If the event does not happen, the same forecast scores 0.49. These are mathematical examples, not observed platform results.

Reported Polymarket results

The original attributed the following results to Polymarket. Without a linked dataset, they should be read as reported sample findings rather than guarantees about future prediction markets.

  • Approximately 90% headline accuracy one month before resolution.
  • Around 89% accuracy one day before resolution.
  • Approximately 94% accuracy in the final four hours.
  • Brier scores clustering around 0.09 across resolved markets, with lower scores reported for higher-volume contracts.

The small decline between one month and one day before settlement does not establish a cause. Late information is one possible explanation, but changing sample composition or measurement choices could also matter. A dataset is needed before attributing that movement to market behaviour.

Reported Kalshi results

Kalshi’s reported Brier scores were routinely below 0.10, approached 0.00 to 0.01 in the final days, and ranged from 0.05 to 0.06 at 200 days before resolution. The original also described a narrower distribution than Polymarket’s, with fewer outcomes above 0.10.

These claims require particular care because distant forecasts can still involve highly predictable events. Standardised contracts may improve comparability, but they do not independently explain a lower score. Prediction markets should be compared on matched questions, timing and scoring conventions before declaring a winner.

Why market prices and opinion polls answer different questions

Opinion polls generally measure respondents’ preferences or intentions, while prediction markets price contracts on specified outcomes. Comparing them fairly requires converting the polling evidence into a forecast for the same event and evaluating both at the same time.

A poll showing a candidate’s vote share is not automatically a probability that the candidate will win. Electoral rules, turnout uncertainty and polling error affect that conversion. Treating vote share and win probability as interchangeable creates a false accuracy comparison.

Prices are also imperfect probability estimates. Fees, limited access, risk preferences, position limits and thin liquidity can affect them. A displayed last-traded price may be stale, particularly in a quiet contract where the available buying and selling prices are far apart.

Before using prediction markets as evidence, check the bid-ask spread, recent trades and order-book depth. A large headline volume for the platform says little about whether you can trade a particular contract at the displayed price.

Claims of statistically significant superiority need more than a lower average score. You need sample size, uncertainty estimates and a fair comparison. The original supplied none of those details, so its stronger claims about outperforming traditional polling are not established here.

What UK readers should check before participating

UK readers should confirm whether a specific service can legally accept them and what protections apply before depositing money. Prediction markets are not automatically authorised for British customers because a website is accessible or a platform is well known overseas.

Where a product constitutes gambling offered to consumers in Great Britain, check the relevant operator and licence details through the UK Gambling Commission. A foreign authorisation does not substitute for the permissions required to serve customers locally. Northern Ireland has a separate gambling framework.

  1. Check eligibility. Read the current location restrictions and account conditions. Do not use technical workarounds to bypass them.
  2. Read the complete contract. Identify the exact event, deadline, time zone and official settlement source.
  3. Understand disputed outcomes. Check who decides ambiguous results, what evidence counts and whether an appeal process exists.
  4. Calculate the full cost. Include trading fees, spread, currency conversion and any applicable withdrawal or network charges.
  5. Check exit conditions. Selling before settlement requires an available counterparty and may involve a loss.
  6. Set an affordable limit. Decide what you can lose before participating and never fund positions with borrowed money.
See also  Can You Count Cards In Online Blackjack?

For gambling participation, you must be 18+. Event trading can involve the same harmful behaviours as other betting, including chasing losses, repeated deposits and compulsive checking. A financial-looking interface does not remove those risks.

If your spending or time feels difficult to control, stop and seek support. Our guide to problem gambling explains warning signs, while BeGambleAware provides a route to gambling information and support.

Do not assume that every event contract receives the same tax treatment as an ordinary bet. Product classification and personal circumstances can matter. Read the overview of gambling winnings taxation as background, and obtain qualified advice where the treatment is uncertain.

What would support further growth in 2026?

Further growth in prediction markets would be more convincing if activity broadened across categories, participants returned and trading conditions improved. Competition from established brands could increase distribution, but brand recognition alone does not establish liquid contracts or reliable forecasts.

The original reported roughly tenfold growth in economic markets and more than 17-fold growth in technology and science markets. Those figures suggest possible diversification, but they need starting values and consistent category definitions. A large percentage increase from a small base can still represent a modest share of overall activity.

Jordan Bender, identified in the original as an equity research analyst at Citizens, was cited as expecting continued adoption and little slowdown. The article connected that outlook with the Olympics, World Cup and World Baseball Classic on the 2026 sporting calendar. These events provide recurring subjects for sports betting, but they do not guarantee platform growth.

The original also discussed DraftKings through Predictions and FanDuel through Forecasts, describing ambitions for product parity and broader offerings. It characterised their initial reception as mixed without supplying supporting acquisition or retention figures. Those comments should be treated as an attributed assessment, rather than a verified comparison of operating performance.

Bender was further described as expecting DraftKings and FanDuel eventually to feature among the top three or four contenders. Robinhood was identified as another potential competitor because of its existing customer base and cross-selling opportunities. These are competitive hypotheses, not established future rankings.

For prediction markets, a more useful monitoring list includes retained users, executable liquidity, transparent resolution and comparable accuracy scores. Higher volume may accompany those improvements, but cannot replace them. Equally, a correlation between volume and lower Brier scores does not prove that additional trading caused better forecasts.

Frequently asked questions about prediction markets

Prediction markets are best understood through their contract rules, probability estimates and local eligibility requirements. These answers address common UK questions without treating reported industry figures as verified results or presenting participation as a dependable source of income.

Are prediction markets legal in the UK?

Eligibility depends on the product, operator and applicable local rules. For gambling services offered in Great Britain, check the operator’s relevant UK Gambling Commission permissions and current terms. Overseas availability does not establish UK authorisation.

Can you use Polymarket or Kalshi from the UK?

Do not assume either platform accepts UK residents. Check its current official eligibility rules and the permissions applicable to the specific service before registering or depositing. This article does not verify current access for either platform.

How accurate are prediction markets compared with polls?

There is no universal accuracy ranking. Compare forecasts for the same outcomes at the same time, using measures such as Brier scores. Polls measure opinions, so they need an appropriate forecasting model before comparison with event probabilities.

How much money is traded on prediction markets?

The original article reported more than $44 billion in 2026 notional volume, but did not provide linked evidence or a consistent timeline. Treat that as an unverified estimate, and distinguish trading turnover from deposits or outstanding positions.

Can you lose all the money spent on an event contract?

Yes. A fully paid binary contract can settle at zero if its specified outcome does not occur, meaning you lose its purchase cost plus applicable fees. Other structures may carry different exposure, so check the terms and maximum loss.

What does a 70% market probability mean?

It indicates that the market price implies roughly a seven-in-ten chance under the contract’s pricing and payout structure, before relevant distortions. It is not a promise. The outcome can fail to happen even when the estimate is well calibrated.