Are Polymarket's top traders bots? 61 of the monthly top 100 are, and they took 70% of its profit

The short answer

Yes for this month, mostly no for all time. Polymarket shows four profit leaderboards: all-time and the last 30 days, overall and in Crypto. We took every wallet that has made the top 100 of any of them, 503 in all, and checked each for the marks a machine leaves in its fills. In our run of October 27, 2026:

• 61 of the monthly top 100 trade like bots, and they made 70% of the board's profit. Eight of the top ten are bots.
• 79 of the Crypto monthly top 100 are bots, most of them in the 5- and 15-minute Up or Down windows.
• The all-time board is still mostly people: 20 of its top 100 trade like bots. Most of the rest stopped trading long ago.
• Bots also collect money the leaderboard leaves out. Rebates and rewards for the active bot-like wallets came to $12.4M over 90 days. The human-paced ones got $1.3M.

What counts as a bot here

Nobody labels a Polymarket account as a bot, so this is a judgement from behaviour. We use the same rules our copy-trading checks use. A wallet is bot-like if it breaks one of two:

• Market maker. Half or more of its fills are resting orders that someone else hit, and it also quotes both outcomes of a market, fills 100 or more times a day, or trades 50 or more markets a day.
• Machine pace. It fills more than 200 times a day, or half or more of its fills are in the 5- and 15-minute crypto windows.

Both are measured on the wallet's last 90 days of fills. A wallet that breaks neither we call human-paced. That does not prove a person is clicking: a slow bot passes. And a very busy person might fail. So read the counts as *trades like a machine*, not as a verdict on who is behind the account.

Board by board

The monthly board has 99 rows because rank 50 was missing from Polymarket's response that day. We give the profit share only for the overall boards, because the Crypto boards rank by Crypto profit, which the API does not give us per month.

The pattern holds by rank. On the monthly board, bots hold:

• 8 of ranks 1–10;
• 12 of ranks 11–25;
• 18 of ranks 26–50;
• 23 of ranks 51–100.

The higher the rank, the more machines. Two weeks earlier, on October 15, our copy-trading run counted 54 bots in the monthly top 100. That is a different day of the same rolling board, so treat it as a direction, not a trend line.

The monthly top ten

The two people in the top ten look nothing like the bots around them. joseph00677 fills about eight times a day and buys at a median 83¢. The wallet at #5 made $2.35M this month on 31 fills in 90 days, and all-time it is up only $74K, so this month undid a long run of losses. A bot earns its place by repeating something small all day. A person gets there with a few big calls.

Four kinds of bot

Of the 316 wallets in our set that are still trading, 202 (64%) are bot-like. They fall into three groups, set against the human-paced wallets for scale:

The four wallets that only buy both sides of a market at a human pace are left out of the table.

Market makers sit on the book and get hit. They are the biggest group on the overall monthly board, mostly in sports, and they hold #1. BreakTheBank fills about 2,400 times a day, 97% of them resting orders. Our market-making guide explains the three ways they get paid and the inventory risk that comes with it.

Minute-market bots trade the 5- and 15-minute Up or Down windows on Bitcoin and other coins. They fill in tiny pieces, a median $7, and many buy both outcomes of the same window. They are the Crypto monthly board: 63 of the 88 still-active wallets on it. Our minute-markets guide explains why the taker fee at 50¢ decides almost everything there.

Fast takers cross the spread instead of resting on it, hundreds or thousands of times a day. Most of them are in politics, and they buy expensive outcomes: a median buy at 79¢, and for the median fast taker 35% of its buys are at 90¢ or more, against 1% for a human-paced wallet. e46m3, #3 this month, buys at a median 97¢. That is a bot sweeping outcomes that are close to settled, the endgame trade our sniping guide describes. UpTheBlues is the sports version, taking about 10,900 fills a day.

By market, the bots are most crowded where the markets repeat fastest. Among active wallets:

• crypto: 86 of 118 bot-like;
• politics: 40 of 60;
• sports: 46 of 85;
• economics: 6 of 13.

Bot records are smoother, not always bigger

The median minute-market bot was up in 85% of its weeks and never gave back more than 5% of its peak. The median human-paced wallet gave back more than half of its peak at some point. That is what repeating a small edge thousands of times looks like next to a few large bets.

Two cautions before reading this as "bots are better traders":

• Survivorship. Every wallet here made a top-100 board. Bots that lost money never appear on one, so this table says nothing about the average bot. Our state-of-bots census found 268 of 592 bot and tool apps had gone quiet within a year.
• Size. Smooth is not the same as large. Over the last 30 days, by leaderboard count, the 80 active minute-market bots made $6.0M together. The 110 human-paced wallets made $16.6M.

The money the leaderboard leaves out

Polymarket's leaderboard counts profit after fees and before rebates. Rebates and rewards are paid separately, daily, and they go almost entirely to bots. Over the last 90 days:

That is about 90% to the bots. Some examples:

• UpTheBlues got $1.30M in taker-fee rebates over 90 days. The board credits it with $2.01M for the last 30.
• HomeRunHazard got $531K in taker-fee rebates, $150K in maker rebates and $26K in rewards.
• BreakTheBank got $134K in maker rebates and $46K in rewards.

Taker-fee rebates are tiered by volume, so the busiest accounts get back the largest share of their fees. Our fee guide explains the tiers. For a bot that trades all day, the rebate can be the difference between losing and winning. That is also why the same code makes less on a small new account.

Bots on the board you can buy

Some catalog sellers run their bot on a public Polymarket account, and four of those accounts are on the Crypto monthly top 100 right now. All four trade almost only the 5- and 15-minute crypto windows.

Rebates here are maker rebates, rewards and taker-fee rebates added together. They are paid on top of the profit column. For the first three, 90 days of rebates come to about four-fifths of a month's leaderboard profit, so they are a large part of what the bot keeps. Each catalog page reconciles the seller's numbers with the account's own feed and points out where the record is narrower than the sales page.

The three our editors picked:

PolymTradeBot

Editor's Choice

Crypto Markets

Automated Python bot for Polymarket's 5- and 15-minute Bitcoin & Ethereum up/down markets. Trades live or in paper mode with configurable risk controls, self-hosted from full source.

Botone-time

GoodFallen

Editor's Choice

Crypto Markets

Self-hosted Python bot for Polymarket's 5- and 15-minute crypto Up or Down windows, sold as full source for $399 once and backed by the public Polymarket account its seller runs it on, @GoodFallen. It does not call direction: through each window it buys both Up and Down in small pieces, some at the ask and some as resting bids, never sells, and redeems the winning side after settlement. Taker fees eat most of the trading result, and Polymarket's daily rebates, which grow with volume, make up half of what the account keeps. Paper mode, a daily loss limit, an exposure cap, a stop file and Telegram reports are built in.

Botone-time

StepAhead

Editor's Choice

Crypto Markets

Self-hosted Python bot for Polymarket's 5- and 15-minute crypto Up or Down windows, sold as full source for $399 once, or for $899 with installation on your server and a private monitoring dashboard. It runs the same rules on every window: when Up and Down together cost less than $1 it buys both and keeps the difference at settlement, and when the order book leans one way it backs that side and holds a small position on the other. It never sells; positions are held to settlement and redeemed. The seller runs it on a public Polymarket account, @stepahead, which today trades almost only Bitcoin's 5-minute windows. Practice mode on live markets, a daily loss cap, an exposure cap and Telegram reports are built in.

Botone-time

Before you buy any of them, run it in paper mode at your own size. A new account starts at the bottom rebate tier. Our cost guide adds up the rest.

How to tell a bot from a person

Every fill on Polymarket is public. Open a wallet's activity and look for these:

1. Pace. Hundreds of fills a day or more, around the clock, with no gap for sleep.
2. Resting orders. Most fills are limit orders someone else hit. Polymarket's fills show maker and taker separately.
3. Both sides. It buys Up and Down, or Yes and No, in the same market.
4. The 5- and 15-minute windows. Mostly "Bitcoin Up or Down" and similar short markets.
5. Tiny fills. A median fill of a few dollars, in strings of near-identical sizes.
6. Daily rebate payouts in the activity feed.

One mark alone proves nothing; plenty of people use limit orders. Three or four together are a machine. Every wallet page under /wallets shows the pace, maker share, both-sides share, minute-window share and rebates we measured, and says which rule a wallet broke.

What this means for you

• If you copy wallets: most of the monthly board cannot be copied. A market maker earns from the very spread a copier pays, and a minute-market bot has closed its position before a copy lands. Our copy-trading study and the list of wallets worth copying start from the ones that pass.
• If you want to run a bot: you are competing with these accounts, at their rebate tier and their latency. Start with what a Polymarket bot is and whether bots are allowed, then how to build one.
• If you read a leaderboard screenshot: check the pace and the rebates before you believe the profit was made by picking outcomes. Our scam red flags cover the rest.

Method

• Wallets. Every wallet that has appeared in the top 100 of Polymarket's overall or Crypto profit leaderboard, all-time or 30-day, since we started measuring on October 11, 2026: 503 in all. Board positions are those of October 27, 2026.
• Data. Polymarket's public Data API: leaderboards (profit after fees, before rebates), trades, activity, positions and rebate payouts. The weekly win rate and drawdown come from the daily profile series. No account or API key is needed.
• Pace. Fills over the last 90 days, from up to 1,000 of the latest. For busy wallets the 1,000 fills cover hours, not 90 days, so their fills per day is the pace of that recent stretch. 196 of the 202 bot-like wallets hit that cap, so their daily counts are rounded here and should be read as an order of magnitude.
• Resting orders. The wallet's latest 500 fills compared with its taker-only fills over the same window.
• Kinds. A wallet counts as a minute-market bot if it is bot-like and half or more of its fills are in 5- and 15-minute windows; otherwise as a market maker if it breaks the market-maker rule; otherwise as a fast taker. Still trading means a fill in the last 14 days.
• Profit shares use the leaderboard's own profit for that board's period, over the wallets on it. The table of four Crypto boards gives no share, because the 30-day Crypto profit is not published per wallet.