Polymarket Leaderboard 2026: What Top Wallets Tell You
The Polymarket leaderboard ranks 20 wallets by profit or volume and explains neither. We read the two 2026 studies on whether copying the top ones works.
Affiliate disclosure: contains affiliate links; we may earn a commission at no extra cost to you. 18+ only — see BeGambleAware.org. Last updated 7 October 2026.
Key takeaways
- The board shows 20 wallets and explains nothing. There is no methodology text anywhere on the page — no formula, no statement on whether open positions count, no mention that profit is quoted before fees.
- The Profit/Volume toggle is not two views of one ranking, it is two different populations. Sort by volume and you are mostly looking at market makers: one wallet near the top had quoted 141,322 markets.
- One wallet is not one person. The 2024 election's biggest winner ran four accounts, and two of them — Theo4 and Fredi9999 — sat on the all-time profit board simultaneously.
- Most people at the top are lucky, and the board cannot tell you which. A Yale and London Business School study of 1.72 million accounts found 3% genuinely skilled and another 29% winning with no statistically discernible skill.
- The one finding you can act on is about order type, not identity. A separate study of 588 million trades found winners post limit orders and losers take market orders — a habit you can copy without copying anyone.
On this page
The Polymarket leaderboard is the most persuasive page on the platform, and the least explained. Twenty wallets, ranked, with a profit figure next to each one — the top of the monthly board read +$3,654,335 when I checked it on 7 October 2026 — and not one sentence anywhere on the page telling you how that number was calculated.
That silence costs money, because the obvious inference is wrong. Of 1.72 million accounts studied over two years, only about 3% showed genuine skill, while 29% were winning on luck alone and collected the larger share of the gains. A board that ranks by outcome cannot separate those two groups, which means roughly nine in ten of the names you are reading got there by a route you cannot repeat.
So this guide does the work the page does not: what each column actually measures, which ranking is quietly a market-maker list, and what two peer-reviewed 2026 studies — one from Yale and London Business School, one from CEPR — measured when they tested whether past leaderboard performance predicts the next month. I read the live board and both papers on 7 October 2026 and quote them directly. New here? Start with our Polymarket explainer.
What the Polymarket leaderboard actually shows
The board lives at polymarket.com/leaderboard and is simpler than its reputation. Two ranking metrics, four time windows, a category filter, and twenty rows. Below the main table sits a second list, "Biggest wins this month", which ranks individual market positions rather than traders.
Here is every control on the page and what it does to the ranking.
That bottom row is the important one. Every third-party tracker that mirrors this data inherits the same gap, and most of them paper over it with the phrase "verified on-chain" — which attests to where the number came from, not to what it means. The absence matters most for the metric toggle, so start there.
Profit or volume: the toggle that rewrites the board
Switching the Polymarket leaderboard from Profit/Loss to Volume feels like re-sorting a list. It is closer to opening a different list. A volume ranking measures turnover, and the wallets with the most turnover are overwhelmingly market makers — accounts quoting both sides of a market to earn the spread, whose trades carry no opinion about who wins.
Bitquery's whale-tracking walkthrough put numbers on this for the week ending 21 September 2026. The wallet at number one had quoted 141,322 markets. Seven of the top eight had placed thousands of orders across hundreds or thousands of outcomes. And the eighth-placed wallet is the one worth memorising:
Bitquery's own summary is the sentence to keep: "Volume made it look worth following. The positions say otherwise." Note also that the figure is quoted before fees, and that this wallet paid $149,358 of them in a week — on Polymarket's fee schedule the charge lands when a trade fills, so a high-turnover strategy pays it hundreds of times over.
The profit ranking avoids this particular trap, and walks into a different one.
One wallet is not one person
Nothing in the data links a trader's accounts together. Split a position across four wallets and you appear as four traders; concentrate it in one and you appear as a single large one. The leaderboard cannot tell the difference, and the best-documented case in the platform's history is exactly this.
The French trader known as Théo, who backed Donald Trump through the 2024 US presidential election, ran four accounts: Fredi9999, Theo4, PrincessCaro and Michie. Polymarket confirmed all four belonged to one person. NBC News reported the accounts had collectively staked over $28 million, with the position valued around $28.6 million at the time, and the trade returned roughly $48 million. His two largest accounts, Theo4 and Fredi9999, both appeared on the all-time profit leaderboard.
Read that as a reader of the board would have: two independent elite traders agreeing on a position, in different sizes, across separate accounts. It was one man with a thesis and four wallets. Théo told the Wall Street Journal he had "absolutely no political agenda" and was "just making money" — he was an ex-bank trader, not a crowd.
This cuts both ways, which is why it is a structural problem rather than a scandal. A trader splitting wallets inflates the apparent number of successful independent opinions, and a trader consolidating them can top the board on a single call. Neither is visible from the outside. So the honest question is not who is on the board, but whether being on it predicts anything at all — and in 2026 that finally got measured.
What the 2026 research found about copying top traders
Two independent academic studies published within weeks of each other took the whole trade history and asked who wins and whether they keep winning. They are the best evidence available, and neither is cited by any of the leaderboard-mirroring sites I checked.
Two notes on reading that table honestly. The 44% persistence figure is far better than chance — Jensen's own framing is that skilled traders are "much more likely to be skilled in the next period than random chance would suggest" — so skill on this platform is real and measurable. And press coverage of the Yale/LBS paper has quoted the profit share as both 27% and over 30%; the figure above is the one Yale's own write-up uses, for skilled traders and market makers together.
What neither paper supports is the inference the leaderboard invites. Measuring skill took a sign-randomisation test across millions of accounts. The board offers you a single realised profit number over a window you chose.
Lucky or skilled: the 29% the board cannot separate
The Yale and LBS paper splits winners into two groups, and the sizes are the whole argument. About 3% of accounts were classified as skilled. A further 29% were "lucky winners" — profitable, with no statistically discernible skill — and that group captured the remaining 69% of gains.
So most of the money won on Polymarket was won by people the authors could not distinguish from coin-flippers, and since the leaderboard ranks on realised profit, those are disproportionately the names on it. The paper's conclusion about the wider user base is blunter than anything a tracker will tell you: "The remaining majority does not produce accuracy; rather, it funds it."
Jensen also expects the skilled share to fall below 1% as competition increases, which cuts against the idea that a profitable wallet today is a safe bet tomorrow. The practical consequence is a filter, not a ranking: a single window of profit tells you almost nothing, and the board only offers windows. Checking a wallet's own trade history is the only way round it — and there, one finding transfers cleanly.
What the winning wallets actually do differently
The CEPR paper's most useful result has nothing to do with who the winners are. It is about how they trade: successful traders tend to provide liquidity with limit orders, while unsuccessful traders take liquidity with market orders.
That is a mechanism, not an identity, which makes it the one thing on this page you can copy without copying a person.
That last row is where copying quietly breaks. If a wallet's edge comes from patient limit orders, then a follower who mirrors it with market orders inherits the position and pays the spread the original wallet was collecting. The strategy and its copy are not the same trade, and the difference lands on the side of the person following.
How to read a Polymarket wallet before you copy it
Every wallet on the board links to a public profile with its positions and history. That is where the answers are, and it takes about five minutes. Work through it in this order, because the early checks disqualify most candidates.
One more check belongs outside the table, because it applies to the trackers rather than the traders. On 7 October 2026 I looked at four third-party Polymarket leaderboards and they named four different wallets at number one — one listed a top-ranked trader showing zero volume, another put a wallet with a large negative P&L first because it sorted by turnover. Polymarket's own page is the baseline; treat anything that mirrors it as a copy that may have chosen a different window, a different metric or a stale snapshot. And before any of this matters, check you can trade at all: polymarket.com blocks trading from the United States and directs US residents to polymarket.us, and UK residents are restricted too.
Check a wallet, not a ranking
Every leaderboard entry links to a public profile with its full position history — the six checks above all live there. Verify your eligibility first; Kalshi is the regulated alternative where Polymarket is restricted.
Visit Polymarket.com →Affiliate link. 18+ only. Prediction markets carry risk of loss. Past performance is not indicative of future results.
My verdict on the Polymarket leaderboard
It is an excellent list of things that happened and a poor list of people to follow. The data behind it is real and on-chain, which is more than most of finance offers, and the research confirms that skill on this platform exists and persists better than it does among mutual fund managers.
The flaw is that the board ranks the one quantity that mixes skill and luck together, over a window you pick, with no fees column, no sample size and no way to see that two rows are the same person. Nine in ten names on it are there for reasons you cannot reproduce.
So use it as a sampling frame, never a ranking: let it hand you twenty public profiles, then disqualify nineteen of them with the six checks above. And take the one portable finding with you — post limit orders, stop paying the spread. That edge needs nobody's permission and survives being copied, which is more than can be said for the wallets themselves. If you would rather the venue did the verifying, read how disputed markets actually get settled before you size up.
Related guides
- How Does Polymarket Work? UK Beginner's Guide 2026 — YES/NO shares, wallets and redemption, from scratch.
- Polymarket Fees 2026: What You Actually Pay — why the fee lands at the fill, and what that does to high-turnover wallets.
- Polymarket Resolution 2026: When UMA Decides — the $750 bond and the two-hour challenge window.
- Polymarket vs Kalshi 2026: Fees, Access and Who Settles — the regulated alternative, compared properly.
- Polymarket UK 2026: Is It Legal? Honest Guide — the restricted-country list and what it means in practice.
- Polymarket Referral Code 2026: What You Actually Get — the four referral programmes, and which one pays you.
Polymarket leaderboard FAQ
What is the Polymarket leaderboard?
A public ranking at polymarket.com/leaderboard showing the top twenty wallets by either Profit/Loss or Volume, over four time windows — Today, Weekly, Monthly and All — with an optional category filter. A second section, "Biggest wins this month", ranks individual winning positions rather than traders. The page publishes no methodology, so how profit is calculated, whether open positions count and how fees are treated are all unstated.
Can you copy trade on Polymarket?
Polymarket has no built-in copy-trading feature; third-party tools offer it by watching wallets and mirroring their trades, usually for a fee. The structural problem is order type: a CEPR study of 588 million trades found successful traders tend to provide liquidity with limit orders while unsuccessful traders take it with market orders, so a copy that fills after the original, at a worse price, pays the spread the original wallet was collecting.
Are Polymarket leaderboard traders actually skilled?
A minority are. A Yale and London Business School study of 1.72 million accounts across 210,322 markets classified about 3% as genuinely skilled, with a further 29% profitable but showing no statistically discernible skill — and that lucky group captured 69% of all gains. Because the board ranks realised profit, both groups appear on it and nothing in the display separates them.
Does past performance on the Polymarket leaderboard predict future results?
Weakly, but better than chance. The Yale/LBS paper found 44% of traders classified as skilled on one random half of their trades were also skilled on the other half, which the authors say is far more than randomness would produce. CEPR's paper describes month-to-month performance as "modestly persistent" while warning it may reflect sample selection rather than skill, and one author expects the skilled share to fall below 1% as competition rises.
Why do Polymarket volume leaders sometimes lose money?
Because volume measures turnover, not accuracy, and the top of a volume ranking is mostly market makers quoting both sides for the spread. Bitquery's walkthrough of the week ending 21 September 2026 found the eighth-largest trader had traded $12,104,594 across 303 orders on six outcomes, won three of six, and was down roughly $434,000 before $149,358 of fees — while ranking inside the top ten.
Is one Polymarket leaderboard entry one person?
Not necessarily. Nothing links a trader's wallets together, so one person can occupy several rows. The French trader known as Theo ran four accounts during the 2024 US election — Fredi9999, Theo4, PrincessCaro and Michie — which Polymarket confirmed belonged to a single trader, and two of them appeared on the all-time profit board at once. A reader would have seen two independent elite traders agreeing.
Is Polymarket leaderboard profit shown before or after fees?
Before fees, and the page does not say so. That matters most for high-turnover wallets, since Polymarket charges at the moment a trade fills rather than at settlement: the eighth-placed wallet in Bitquery's September 2026 sample paid $149,358 in fees in a single week. Any profit figure you read on the board or on a tracker that mirrors it should be treated as a pre-cost number.
Why do third-party Polymarket leaderboards disagree with each other?
Because each picks its own metric, window and refresh rate, and none of them can see Polymarket's formula either. Checking four such trackers on 7 October 2026 returned four different wallets at number one, including one top-ranked entry showing zero volume and another led by a wallet with a large negative P&L because the page sorted by turnover. Use polymarket.com/leaderboard as the baseline and treat mirrors as snapshots.
Sources checked
- Polymarket leaderboard — metrics, windows, row count and the absence of methodology text, read 7 October 2026.
- Yale Insights — "Wisdom of the Few", on Gomez-Cram, Guo, Kung and Jensen, 10 June 2026.
- CEPR DP21615 — Akey, Gregoire, Harvie and Martineau, "Who Wins and Who Loses in Prediction Markets?", 12 June 2026.
- Bitquery whale-tracker walkthrough — the week ending 21 September 2026, market-maker footprints and the eighth-placed wallet.
- NBC News — the four accounts controlled by one French trader in 2024.