How Polymarket Pricing Actually Works — AMM vs. Orderbook, Explained Simply

The single most confusing thing about Polymarket is pricing. Is 62¢ a probability? A price? Both? And why does a $50K buy sometimes barely move the market while a $10K buy somewhere else moves it 3¢? This post is the model you need to answer those questions confidently.

Start with the core idea: price equals probability

On Polymarket every share of an outcome pays out exactly $1 if correct and $0 if not. So the current market price of YES is literally the market-implied probability of YES. If YES is trading at 62¢, the market thinks YES has a 62% chance. That's also the price you pay to buy one YES share.

One immediate consequence: YES price plus NO price on a binary market always sums to about $1 (minus a penny of spread). If YES is 62¢ and NO is 35¢, there's a 3¢ arbitrage — someone will pick it off within seconds on any liquid market.

The old way: AMMs

Polymarket launched on Uniswap-style automated market makers. Liquidity providers deposited USDC plus YES and NO tokens into a shared pool. Prices were computed by a constant-product formula: x times y equals k. Any buy pulled tokens out of the pool, shifting the ratio and bumping the price.

AMMs have one huge benefit: they always have a price, even on tiny markets nobody has ever looked at. But they have brutal downsides for whales: impermanent loss, high slippage, and no true price-time priority. A whale couldn't rest a limit order — they had to take whatever curve price was showing, the entire way up.

The new way: CLOB (orderbook)

Today's Polymarket runs a central limit order book, just like a traditional exchange. Makers post limit orders (“I'll sell 1,000 YES at 62¢”), and takers hit those orders. The matching engine is fully on-chain — settled through Polygon — but the UX matches Kalshi, Robinhood, or any normal exchange.

Why the switch? Three reasons: (1) tighter spreads on liquid markets, (2) whales can now rest limit orders without getting destroyed by curve slippage, (3) no impermanent loss for liquidity providers. The tradeoff: thinly-traded markets sometimes have a 4–8¢ spread or no bids at all. The AMM would always quote something; the CLOB just shows an empty book.

Slippage, in plain English

Slippage is how much your own trade moves the market against you. Imagine the YES book looks like this:

If you buy $20,000 of YES ($20K / $0.62, about 32K shares), you eat all of 62¢ (2K shares) + all of 63¢ (5K shares) + most of 64¢ (25K shares). Your average fill is about 63.6¢ — 1.6¢ higher than the top-of-book quote. That 1.6¢ is slippage, and it's where most retail gets surprised.

Whales know this. They split their bets into multiple orders over minutes or hours, or they rest limit orders below the ask and wait for someone to cross the spread to them. This is why our Signal Engine weighs time to build position as a feature — a whale who built $250K over three hours tells you something different than one who market-smashed $250K in a single block.

Why whales move thin markets 5¢ in a second

On a market with $50K of total liquidity across all asks, a $40K whale buy is the entire top of the book plus half of the next level. Price jumps 5¢ instantly. Then either (a) market makers replenish liquidity at the new level (slow recovery) or (b) other traders notice the whale's move and pile in (fast move further away).

This is why market thickness matters more than market size for signal quality. A whale putting $100K into a $10M market tells you less than a whale putting $100K into a $200K market, because the second one is betting on a market where most other participants are wrong and the whale has sized that bet to pay for itself.

Fees, in one paragraph

Polymarket charges 2% on profits at resolution time (you keep 98% of what you win). Gas on Polygon is negligible (under $0.05 per trade). Kalshi by comparison charges per-trade fees that are higher in aggregate. When computing arbitrage, always subtract both venues' fee tiers from the gross spread — a 2.5% gross edge often collapses to 0.4% net, which usually isn't worth the execution risk.

What this means for reading whale positions

• Entry price above 50¢ on a YES bet = whale thought the probability was already real; they're paying for conviction, not arbitrage.
• Multiple whales entering a single market within 24h often triggers a herd reaction; consensus detection lets you see it before the price moves.
• A whale building slowly vs. market-buying tells you whether they're front-running a news event (slow) or responding to one (fast).
• Thin-market consensus bets are MUCH higher-variance; size your tails smaller on them.