Prediction Markets for Beginners: 2026 Starter Guide

Key takeaways

1. A contract's price is its probability, and it settles at $1.00 or $0.00. Yes at 40 cents means the market is pricing a 40% chance, so 100 contracts cost $40.00 plus a taker fee of 96 cents.
2. Fees are a parabola, not a flat percentage. Both venues charge rate x contracts x price x (1 - price), so a 50 cent coin flip is the most expensive trade on the board and a 90 cent favourite is one of the cheapest.
3. Your breakeven is the price plus the fee, never the price alone. At 50 cents that is 51.00% on Polymarket politics, 51.25% on sports and weather, 51.75% on crypto or as a Kalshi taker, and exactly 50.00% on a fee-free geopolitics market.
4. A sportsbook pricing both sides of a coin flip at -110 needs you to hit 52.38%. That gap of roughly one point per trade is the entire financial case for trading events rather than betting them.
5. The spread usually costs more than the fee. A 2 cent spread on a 40 cent contract is 5% of your stake against 2.4% for the taker fee, so start on liquid markets and post limit orders, since makers are never charged on Polymarket.

On this page

1. What a prediction market actually is
2. One 100 contract trade, start to finish
3. The price is the probability, and that is the whole trick
4. What it costs: the fee formula and every current rate
5. Makers, rebates and the October 2026 changes
6. The win rate you need just to break even
7. How a market resolves, and who decides
8. Where prediction markets stand legally in October 2026
9. Your first month: a plan that does not blow up
10. Frequently asked questions
11. Sources and when we last checked

This is a plain list rather than jump links, because guide pages on this site render section text as paragraphs. Scroll to whichever heading you need. If you read only one, read the fee section and the breakeven table: together they decide whether a trade is worth placing.

What a prediction market actually is

A prediction market is an exchange where you trade contracts on whether something happens. Each contract is a yes or no question with a fixed payout: $1.00 if it happens, $0.00 if it does not. Will the Fed cut rates in December? Will this party hold the Senate? Each is its own market with its own order book.

The part that surprises people coming from a sportsbook is that there is no house. You buy from and sell to other traders, and the platform matches you and takes a fee. Nobody sets a line or adjusts it to balance a book. The price is whatever the last trade cleared at, so it moves as people with money form a view.

Two consequences follow. First, you can sell before resolution, so a position is not a bet you are locked into: buy at 40 cents, watch it move to 55, and you can take the 15 cents without ever learning the answer. Second, your downside is strictly what you paid, because a contract cannot go below zero. 100 contracts at 40 cents can lose you $40.00 and not a cent more.

The two venues a US beginner will actually use are Polymarket and Kalshi, both CFTC-regulated. The table above is the short version of how they differ, and our full comparison has the detail. Everything below applies to both, because their mechanics and even their fee formulas are closer than their marketing suggests.

One 100 contract trade, start to finish

1. Buy 100 Yes contracts at 40 cents: stake $40.00
2. Taker fee on entry, politics rate 0.04: 0.04 x 100 x 0.40 x 0.60 = $0.96
3. Total out of pocket: $40.96
4. If it resolves Yes: you receive $100.00, a profit of $59.04, or 144% on the money at risk
5. If it resolves No: you lose the whole $40.96
6. Breakeven true probability: 0.40 + 0.04 x 0.40 x 0.60 = 40.96%
7. If you instead sell at 52 cents: receive $52.00, pay another taker fee of $1.00, net $51.00
8. Round trip profit on that 12 cent move: $10.04, after $1.96 of fees

Work through that list once and you have the whole model. You put up $40.96 for a shot at $100.00, and the only question that matters is whether the real chance of Yes beats 40.96%. Not 40%, which is the number on screen. The fee moves the hurdle, and anyone who skips it does not know what they are aiming at.

The last two lines are the ones every beginner guide I read leaves out. Selling early is a second trade, and if you cross the spread to do it you pay a second taker fee. On this example the two fees come to $1.96 against a 12 cent move worth $12.00, so roughly 16% of the gain went in costs. On a 3 cent move it would have eaten most of it. Active trading is not free just because the headline fee looks small.

One last thing about that 144% figure. A 40 cent contract pays 1.5 times your stake when it wins and loses everything when it does not, which is not an asymmetry you can exploit. It is the same bet priced at 40% instead of 60%, and the only edge on offer is being right about the probability.

The price is the probability, and that is the whole trick

If Yes trades at 65 cents, the market is pricing a 65% chance, and the matching No contract will sit near 35 cents because the two have to add up to roughly a dollar. There is no line to decode and no plus or minus number to reverse engineer. A sportsbook quotes you a payout and hides the probability inside it; a prediction market quotes the probability and lets you work out the payout.

The conversions, if you want to compare against a bookmaker, are short. Implied probability is the price itself. Decimal odds are 1 divided by the price, so 65 cents is 1.54. American odds split at even money: under 50 cents use 100 x (1 - price) / price as a positive number, so 25 cents is +300, and over 50 cents use -100 x price / (1 - price), so 75 cents is -300. Our odds guide works through every price point and the edge maths in full.

Two cautions. Convert the price you can actually fill at, not the midpoint on a chart, because on a thin market the best bid and the best offer can be several cents apart. And in a multi-outcome market each option trades as its own Yes contract, so nothing forces the set to total 100 cents. A four-way race showing 40, 30, 20 and 15 sums to 105, which is a 5 cent overround you are paying for a thin book.

What it costs: the fee formula and every current rate

1. Polymarket global formula: fee = contracts x feeRate x price x (1 - price), charged in USDC
2. Politics, finance, tech and mentions, rate 0.04: $1.00 per 100 contracts at 50 cents, $0.75 at 25 or 75 cents, $0.36 at 10 or 90 cents
3. Sports, economics, culture, weather and other, rate 0.05: $1.25 at 50 cents, $0.94 at 25 or 75 cents, $0.45 at 10 or 90 cents
4. Crypto, rate 0.07: $1.75 at 50 cents, $1.31 at 25 or 75 cents, $0.63 at 10 or 90 cents
5. Geopolitics and world events, rate 0: no taker fee at any price
6. Polymarket makers: never charged, and paid from a rebate pool worth 25% of taker fees in most categories, 20% in crypto and 15% in sports
7. Polymarket US takers: coefficient 0.0695, which peaks at $1.74 per 100 contracts at 50 cents
8. Polymarket US makers: a rebate of 0.0125 paid to you, worth $0.31 per 100 contracts at 50 cents
9. Kalshi takers: round up(0.07 x contracts x price x (1 - price)), so $1.75 per 100 contracts at 50 cents
10. Kalshi makers: a 0.0175 coefficient exists in the schedule but its multiplier defaults to zero, so a resting order on a standard series pays nothing

Here is the detail the big-brand explainers skip. Fidelity's primer says the platform collects a transaction fee and never names an amount, and one popular beginner comparison quotes a 0.10% taker fee for Polymarket US, which is not the published schedule. The real structure on both venues is the same shape: contracts x rate x price x (1 - price). That second term is a parabola peaking at 50 cents, so the fee is largest where the market is most uncertain and nearly vanishes at both ends.

Run it and the effect is big. On a crypto market, 100 contracts at 50 cents costs $1.75 in fees while the same 100 at 95 cents costs 33 cents. As a share of what you put up that is 3.5% against 0.35%. Coin flips are roughly ten times more expensive, per dollar risked, than near certainties, so a beginner hunting small edges should hunt them away from 50 cents.

Makers, rebates and the October 2026 changes

The maker side is the biggest free lunch available to a careful beginner. Posting a limit order that someone else fills makes you a maker, and Polymarket does not charge makers at all, instead paying them from the taker fees collected in that market. On Polymarket US the maker coefficient is negative, meaning the venue pays you 0.0125 x contracts x price x (1 - price) for providing the liquidity. The cost of that patience is that your order may never fill.

Two live dates to know if you trade the regulated US venue. The extra weekly crypto maker rebate ended at 8am ET on 15 October 2026, and from 1pm ET on 19 October 2026 the maker and taker rebates move onto volume-based tiers, with maker coefficients ranging from 0.0035 to 0.0104 and taker rebates from 3% to 30%. The taker fee formula and its 0.0695 coefficient are unchanged. Until then the older tiers apply: 10% of taker fees rebated above $250,000 of monthly volume, 25% above $1 million and 50% above $10 million.

Both of those are rebate changes, not fee-formula changes, so nothing above moves. They matter to a beginner only as a reminder that these schedules are revised often, and that the rate you memorise today is worth re-reading before you size anything serious.

The win rate you need just to break even

1. Geopolitics on Polymarket, no fee: 50.00%
2. Politics, finance, tech and mentions on Polymarket, rate 0.04: 51.00%
3. Sports, economics, culture and weather on Polymarket, rate 0.05: 51.25%
4. Crypto on Polymarket, rate 0.07: 51.75%
5. Polymarket US as a taker, coefficient 0.0695: 51.74%
6. Kalshi as a taker, coefficient 0.07: 51.75%
7. A sportsbook pricing both sides at -110: 52.38%

Those figures are all for a 50 cent contract, where the fee peaks, and they answer the only question worth asking before you click buy: how often do I have to be right just to stand still? The rule behind every row is one line. Your breakeven true probability is the price plus the fee per contract, or price + rate x price x (1 - price).

It works at any price, and the hurdle shrinks as you move away from even money. At 65 cents on a politics market the fee is 0.04 x 0.65 x 0.35, or 0.91 of a cent, so you need 65.91% rather than 65%. At 90 cents the fee is 0.36 of a cent and breakeven is 90.36%. At 10 cents it is 10.36%.

Compare the bottom two rows and you can see why prediction markets have taken share from sportsbooks. A book pricing both sides of a proposition at -110 is implying 52.38% on each, which sums to 104.76% and leaves it a hold of about 4.55% of everything staked. The most expensive prediction-market category asks you to beat 51.75%, and the cheapest asks for exactly 50%. That difference, a little over half a point to two and a half points per trade, compounds across every position you take.

One caveat, and it is the one that actually costs beginners money: these rows price the fee and nothing else. Crossing a 2 cent spread on a 40 cent contract costs 5% of your stake before any fee, which dwarfs the 2.4% the fee takes. Cheap fees and a thin order book are not a cheap trade.

How a market resolves, and who decides

Every competing beginner guide waves at this and none explains it, and it is where a confusing loss usually comes from. A market does not resolve because the news says so. It resolves on its own written rules, which name the data source, the deadline and the threshold. Read them before you trade, because an obvious-looking outcome can resolve the other way on a technicality.

On Polymarket's global platform, resolution runs through the UMA Optimistic Oracle. Someone proposes the outcome and posts a bond in USDC, currently $750, which they forfeit if the proposal fails. The proposal then sits in a challenge period of two hours during which anyone who disagrees can dispute it, and a disputed market escalates into UMA's own dispute process rather than being settled by Polymarket staff. A proposer who jumps too early loses the bond, which is what keeps the system honest.

Kalshi works through its exchange rulebook instead. Each contract names its settlement source in terms filed with the CFTC, and positions resolve automatically once the outcome is determined, usually shortly after expiry, though the exchange's documentation notes that timing varies with the data source and whether manual review is needed. Contested determinations go through a re-determination process rather than an oracle vote.

The lesson is the same on either venue: open the resolution criteria and check which source decides, by what date, and what happens in an ambiguous case. Two exchanges can resolve the same apparent question differently, which is also why a price gap between venues is not always free money.

Where prediction markets stand legally in October 2026

Both venues are CFTC-regulated, so at the federal level event contracts are permitted rather than licensed state by state. For the ordinary non-sports market that is most of the story, and those markets have drawn little state-level challenge.

Sports event contracts are the live fight, and in 2026 the appeals courts have split. The Third Circuit sided with Kalshi against New Jersey in April. The Ninth Circuit ruled against it in Nevada in August. On 25 September the Sixth Circuit held that Ohio and Tennessee may apply their state gambling laws to Kalshi's sports contracts, and Kalshi petitioned for rehearing on 9 October. The CFTC maintains that its jurisdiction is exclusive and has taken action against several states, so the question looks headed for the Supreme Court.

What that means for you is narrower than the headlines suggest. Access is decided by where you physically are, and both apps run a geo-check that is the final word for your situation. Polymarket US has paused trading in Nevada, and several states including Connecticut, Maryland, New Jersey, New York and Minnesota have active disputes. Published state counts conflict with each other and go stale within weeks, so treat any list, including ours, as a starting point and let the app confirm.

Our legality guide covers the federal position and the state nuances in full. The short version: non-sports markets are broadly available, sports contracts are contested in a handful of states, and the app tells you which bucket you are in before you deposit.

Your first month: a plan that does not blow up

1. Fund small. $100 is plenty to learn on, and Kalshi's minimum deposit is $10
2. Cap any single position at 10% of the bankroll, so $10 on a $100 account
3. Trade away from 50 cents while you learn: the fee at 90 cents is $0.36 per 100 contracts against $1.00 at 50 cents
4. Prefer markets with a 1 cent spread and visible depth over a tempting price on an empty book
5. Post limit orders rather than crossing the spread, because makers are not charged on Polymarket
6. Write down the probability you believed, and why, before you place the trade
7. Avoid combination or parlay-style trades at first: the Polymarket US combo fee adds a 0.06 x (1 - price) to the fourth power term that loads cost onto cheap legs
8. Review after 20 trades: compare your stated probabilities against what actually happened

The single habit that separates people who improve from people who churn is that last pair of lines. If you record the number you believed before each trade, then after twenty trades you can check whether your 70% calls came in about 70% of the time. That is a calibration record, and it is the only honest feedback this activity gives you. Profit and loss over twenty trades is mostly noise; calibration over twenty trades is information.

The position cap matters more than it looks on a small account. At 10% per position you can be wrong five times running and still have most of your bankroll, which is a perfectly ordinary outcome when you are trading 40 cent contracts. At 40% per position, the same five losses end the experiment. Nothing about a good process protects you from a run of bad outcomes, so size for the run.

One expectation to set honestly: a beginner's first edge is usually negative, and the fee table shows why. If your reads are worth one or two points, the fee at 50 cents eats them entirely and the spread eats them twice over. Aim to become calibrated and cheap to trade, and treat any first-month profit as luck.

Frequently asked questions

Sources and when we last checked

1. Polymarket global trading fees, help.polymarket.com article 13364478: the contracts x rate x price x (1 - price) formula, the per-category taker rates and the maker rebate shares
2. Polymarket US fee schedule, docs.polymarket.us/fees: the 0.0695 taker coefficient, the 0.0125 maker rebate, the combination-trade formula and the 15 and 19 October 2026 rebate changes
3. Polymarket market resolution, help.polymarket.com article 13364518: the UMA Optimistic Oracle, the $750 proposal bond and the two hour challenge period
4. Kalshi fee schedule, kalshi.com/docs/kalshi-fee-schedule.pdf: the round up(0.07 x contracts x price x (1 - price)) taker formula and the 0.0175 maker coefficient with its default multiplier
5. Kalshi market settlement, docs.kalshi.com: automatic resolution on determination and the variability of settlement timing
6. Appeals court coverage of state regulation of sports event contracts, September and October 2026, for the Third, Sixth and Ninth Circuit outcomes

Every figure in this guide was checked against those documents on 15 October 2026. Fee schedules on both venues moved more than once during the year and the Polymarket US rebate tiers change again on 19 October, so treat these numbers as a snapshot rather than a constant. The legal position on sports contracts is moving faster still, and the state counts published by various trackers disagree with each other, which is why this guide points you at the app's own geo-check rather than a list.

If you are sizing a real position rather than reading for background, open the live fee schedule for your category and the market's own resolution criteria first. For the specific market you are trading, those two documents override any guide, including this one.

Related guides