How much do Polymarket market makers make? About 1.4¢ per dollar traded, and one bad week can take months of it

The short answer

Market making on Polymarket is a thin business. A maker quotes both sides of a book, earns a little on each fill, and gets paid extra by Polymarket's rebate and reward programs. The question is how little "a little" is. So we took every wallet we track that quotes for a living and traded at least $5M in the last 30 days, and divided what it kept by what it traded.

• About 1.4¢ per dollar traded. The 51 wallets traded $995M and kept $13.84M: $10.29M from trading, after fees, and $3.55M from rebates and rewards.
• 37 of 51 ended the month ahead. 34 made money from trading alone, and payouts carried 3 more over the line.
• The market matters more than the size. Large politics makers kept 4.9¢ per dollar and sports makers 1.9¢. Crypto makers kept 0.56¢, and makers with no profitable category lost 2.4¢.
• Inventory is the risk. For the typical large maker, its worst week of the last six months wiped out 5.6 ordinary weeks. For 15 of 35 it wiped out ten or more.

Who counts as a market maker here

Our wallet checks call a wallet a market maker when at least half of its recent fills rest on the book, and it also quotes both outcomes of a market, fills 100 times a day or spreads across 50 new markets a day. 168 wallets that are still trading pass that test.

Most of them are small. 117 traded less than $5M in the month, $236M between them, and showed $21.9M of leaderboard profit. That is 9.3¢ per dollar, far more than a spread pays. A wallet with that ratio is making its money on where its positions end up, not on quoting. So the rest of this guide uses the 51 that traded at least $5M. They are the ones whose business looks like market making from the numbers too: a median $10M a month, about 580 fills a day, $24 a fill and 92% of fills resting on the book.

A cent per dollar

Here is where each dollar of the $995M went:

The trading line is more than the spread. It is the spread plus whatever the maker's inventory did when markets resolved, minus taker fees on the fills where the maker crossed. A maker that leans the right way into a game makes more than its quotes paid. One that leans the wrong way can lose more than a month of spread.

The payouts add about a third on top. The largest of the three for these wallets is the taker-fee rebate, which sounds odd for a maker. Large makers also take liquidity at times, to flatten inventory or to hedge, and at their volume they sit in the top fee-rebate tiers. That part of the income doesn't carry over to a small account, as the rebates study shows.

How the 51 did

Payouts were 26% of the group's total but only 12% for the median profitable wallet. The sports giants at the top collect most of them. RN1 traded $73.6M and was paid $289K, which is 28% of what it kept. BreakTheBank traded $56.6M, made $3.99M from trading and was paid $163K: for it the payouts were 4%.

Three wallets lost on the leaderboard and were carried into profit by payouts. The largest was ferrariChampions2026: it lost $120K trading $84.0M and was paid $403K.

Payouts don't save a bad month, though. mooseborzoii traded $51.9M, lost $2.51M and was paid $152K. Thunderous-Authoritarian-Snarl lost $2.36M on $14.5M and was paid $82K.

By market

"Main market" is the category where the wallet has made most of its all-time profit. "No profitable category" means it is down in every category on the leaderboard, which is why that group looks so bad.

Sports is where the big books are. 16 of the 25 makers that traded more than $10M are sports makers, and they are the steadiest group: 19 of 22 were ahead. A game resolves in hours, so inventory doesn't sit for long, but a goal or a break of serve moves the price faster than a quote can be pulled. Our sports guide covers the matching delay that protects makers there.

Politics paid the most per dollar, and that number needs care. Seven wallets, $45M of volume, and one month with a few big resolutions in it. A politics maker carries inventory for weeks, so its month is as much about which way that inventory leaned as about the spread.

Crypto paid the least. These makers quote the 5- and 15-minute windows, where the books are crowded with other bots and the price moves with Bitcoin second by second. The 9 large ones kept 0.56¢ per dollar from trading. 0x99a0…415f lost $637K on $17.2M in the month, yet was up in 94% of its weeks over its record. That is what a small edge with an occasional large loss looks like.

Bigger isn't better per dollar

Across all 168 market makers, sorted into quarters by 30-day volume:

Per dollar, the largest makers keep the least from trading and depend most on the programs. Volume earns rebates and rewards in proportion, while a good position doesn't grow with volume. The largest quarter is also where losing months happen: 15 of its 42 wallets lost on the leaderboard, against 5 of 42 in the smallest.

Read the small quarters with care. A wallet that traded $1M and shows $100K of profit is mostly a directional trader that happens to rest its orders.

The bad week

A maker's risk is the inventory it is left holding when the price runs. To see how often that bites, we took each large maker's weekly profit over the last 26 weeks from its PnL series. 45 of the 51 have at least eight full weeks in it.

The pattern is many small weeks and a few large losses. ferrariChampions2026 made a median $262K a week and lost $3.56M in the week to June 21, during the World Cup group stage: almost fourteen ordinary weeks in one. It still finished the 26 weeks $3.76M ahead. 0x2c33…0563 lost $1.40M in the week to September 20, nearly three of its ordinary weeks.

One market can do it on its own. At the US Open, BreakTheBank put $962K into the Shelton vs Alcaraz match and got $196K back: a $766K loss in a single binary market. That one is from the wallet's own trade and redemption history, not an estimate.

Our market-making guide explains how makers try to limit it: inventory caps, skewing quotes, pulling them before news and hedging what they can't flatten.

The last 90 days

The 30-day leaderboard row is the cleanest measure, because it is after fees. Over 90 days we have to use the profile PnL series, which is before taker fees, so these numbers run a little high:

Over a longer window, more of them are ahead and the margin is about the same: under a cent and a half per dollar. The 67% of up weeks is lower than the 85% our top-traders study found for minute-market bots. Market making on Polymarket is less smooth than its reputation.

What this means for a market-making bot

• Price it per dollar traded. At 1.03¢ of trading profit per dollar, a bot needs to trade about $1M a month to make $10K. With $10K of capital, that means turning the whole bankroll over a hundred times a month without one bad resolution.
• Don't count the big makers' rebates. About half of the group's payouts were taker-fee rebates, earned at volume tiers a new account starts without. Maker rebates and rewards came to 0.19¢ per dollar, and that part scales with what you quote.
• Pick the market for the inventory, not the spread. Sports inventory clears in hours. Politics inventory can sit for months, and a position that runs against you sits with it.
• Ask a seller for profit per dollar traded. A "market-making bot" with a big month on little volume made a directional bet. Divide the profit by the volume. Anything far above a few cents per dollar isn't spread capture.
• Start with code you can read. Poly-Maker is open source and was built as a reference: quoting over WebSockets, position merging and risk limits. Run it small, measure your own cents per dollar, and compare with the table above.

Poly-Maker

Open source · GitHub

Poly-Maker

Market Making

Open-source market-making bot for Polymarket. Quotes both sides of the book over WebSockets with configurable spreads, position merging and risk controls.

Botopen-source

Polybot (No-Code)

Market Making

No-code Chrome extension for building automated Polymarket trading bots right in your browser. Spin up a bot from any event page with one click, add stop-loss and take-profit orders, or run the Fair Value market-making…

Botfree

PolyScalping

Scalping

Real-time scanner and reward analytics for Polymarket. Surfaces scalping opportunities as they appear and shows the full-wallet math Polymarket hides — tracking LP rewards, maker rebates, referral rewards and yield…

Platformfree

How to check a market maker yourself

Three public calls give you the numbers in this guide for any wallet. Dollar volume over a window, with Unix timestamps:

curl -s "https://data-api.polymarket.com/v2/user-volume?user=0xWALLET&start=1760000000&end=1762600000"

Use volume_usdc. The volume field next to it is in shares, and for a maker it runs about twice as high. Then the payouts, and the 30-day leaderboard row:

curl -s "https://data-api.polymarket.com/v2/activity?user=0xWALLET&type=MAKER_REBATE,REWARD,TAKER_REBATE&start=1760000000&limit=500"
curl -s "https://data-api.polymarket.com/v2/leaderboard?user=0xWALLET&time_period=month"

Divide profit plus payouts by volume_usdc. Every wallet page under /wallets shows its 90-day payouts and its weekly record.

What this can't tell you

• It is one month. The 30 days to November 8, 2026 hold the NFL and college football seasons, the US midterms and the end of the baseball season. A World Cup month would look different.
• Only winners are tracked. Every wallet here has made a top-100 profit board at some point. Makers that never got that far aren't in the data, so the typical maker is likely doing worse than these 51.
• Capital is invisible. We can see what a wallet trades and holds, not the cash behind it, so this is a margin on volume and not a return on capital.
• The trading line mixes spread and direction. The leaderboard doesn't split profit from quoting and profit from where inventory ended up.
• The kinds are ours. "Market maker" is a rule on fills, explained above, and a few wallets sit near its edges.

Method

• Wallets. The 168 wallets in our wallet set that had traded in the last 14 days and pass our market-maker rule, as of November 8, 2026, 12:03 UTC. The guide's main group is the 51 with at least $5M of 30-day volume.
• Volume. volume_usdc from /v2/user-volume with start and end set to the window.
• Profit. The 30-day leaderboard row from the same rebuild: after taker fees, before rebates. The 90-day figures use the change in trade_pnl from /v2/user-pnl, which is before taker fees.
• Payouts. Every MAKER_REBATE, REWARD and TAKER_REBATE activity row dated inside the window, all pages.
• Main market. The leaderboard category with the most all-time profit.
• Weeks. Week-on-week changes in the PnL series over the last 26 weeks, using only gaps of six to eight days. "Ordinary week" is the wallet's median week.
• Script. scripts/market-maker-income.mjs in our repository does the pull. It is plain Node with no keys.