Every Polymarket market pays exactly $1.00 to the winning outcome. When you can buy every outcome for less than $1.00 combined — after fees — the difference is locked in. This scanner checks the live order books of the top 400 binary markets and 25 multi-outcome events for exactly that, using real asks, not last-trade prices.
📬 The PolymarketBro Daily Digest
Top movers, the biggest whale trades, and the one market to watch — in your inbox every morning. Free, one email a day, unsubscribe anytime.
No spam, no selling your address. Ever.
The morning scan found 6 opportunities right now, after fees. Best net edge seen anywhere in the universe: +6.39%.
| Market | Type | YES ask | NO ask | Combined | Net edge after fees | Size at best ask | |
|---|---|---|---|---|---|---|---|
| Next CEO of NVIDIA? | Neg-risk basket | Σ 28 legs | — | $0.928 | +6.39% | 1 sh (~$1) | Trade → |
| Democratic Presidential Nominee 2028 | Neg-risk basket | Σ 58 legs | — | $0.931 | +2.75% | 278 sh (~$259) | Trade → |
| Which Millennium Prize Problem will AI solve next? | Neg-risk basket | Σ 6 legs | — | $0.958 | +2.61% | 4 sh (~$4) | Trade → |
| Prime Minister of Israel after the next election? | Neg-risk basket | Σ 22 legs | — | $0.976 | +2.40% | 20 sh (~$20) | Trade → |
| Presidential Election Winner 2028 | Neg-risk basket | Σ 58 legs | — | $0.946 | +1.08% | 56 sh (~$53) | Trade → |
| Fed Decision in January? | Neg-risk basket | Σ 5 legs | — | $0.991 | +0.36% | 1,014 sh (~$1,005) | Trade → |
The 15 binary markets with the lowest combined YES + NO ask in the November 21, 2026 scan — the pairs closest to paying for themselves. A negative net edge means fees and the spread still eat the whole gap: interesting to watch, not a trade.
Real order books, not last trades. Market discovery runs through Polymarket's Gamma API, but every price on this page comes from the live CLOB order book — the best (lowest) ask on each side. Last-trade prices and mid-prices flatter arbitrage that isn't there; you can only buy at the ask.
Fees are deducted, always. Polymarket charges takers a fee per share that depends on the market's category and the price:
fee per share = rate × p × (1 − p)
Rates: crypto 7%, sports 5%, economics / culture / weather 5%, finance / politics / tech 4%, geopolitics 0%. The fee peaks at 50¢ (a 4% politics market charges 1.00¢ per share at 50¢) and shrinks toward the extremes. A pair's true cost is:
all-in cost = YES ask + NO ask + fee(YES) + fee(NO)
Two kinds of scan. Pairs: one binary market — buy YES and NO, one must pay $1.00. Neg-risk baskets: multi-outcome events (elections, champions, decisions) — buy YES on every outcome with a live market; exactly one pays $1.00. Baskets cover every outcome that has a live, priced market; placeholder outcomes with no market yet can't be bought and aren't counted. An event is only basket-scanned if its outcomes' prices sum near $1.00 to begin with — collections of independent bets are never dressed up as arbitrage.
Worked example. A politics market (4% fee rate) shows YES ask 46.0¢ and NO ask 51.0¢ — combined 97.0¢. Fees: YES leg 0.04 × 0.46 × 0.54 = 0.99¢; NO leg 0.04 × 0.51 × 0.49 = 1.00¢. All-in cost = 97.0 + 0.99 + 1.00 = 98.99¢ per pair of shares. One side must pay $1.00, so the net edge is +1.01%: buy 1,000 pairs for $989.93, receive $1,000.00, keep $10.07 — if every share fills at those asks.
Fees. As the example shows, taker fees consume roughly 1–2¢ per pair near even prices — more than most visible gaps. Any scanner quoting gross edges without fees is quoting fiction.
Thin depth. The best ask might be good for 40 shares. The size column shows what's actually available at the displayed price on the limiting leg; walking the book raises your average cost and can erase the edge.
Leg risk. Prices move between your first and second fill — especially around news. A half-filled "arbitrage" is just a position, with full directional risk.
Resolution risk. The $1.00 guarantee only holds if the outcomes truly partition the result — same market, same resolution source, same criteria. Never treat two different markets that merely look complementary as an arbitrage; if their resolution criteria differ, both can resolve against you.
Polymarket is watched by arbitrage bots that react in well under a second. Market makers quote YES and NO so the pair sums to $1.00 plus a spread, and any shock that knocks a pair under $1.00 is usually competed away before a human can click twice. Fees then bury whatever is left: a pair summing to 99.0¢ in a crypto market loses money after the 7% fee schedule. So an empty opportunities table is the honest, normal result — the watchlist above exists so you can see how empty, and how close the closest pairs are, on real numbers. Openings do happen — around breaking news, market opens, and resolution confusion — which is why the page rescans live in your browser and why the Telegram channel posts the morning scan.
Yes — this page. It scans the live Polymarket CLOB order books for binary markets whose YES and NO asks sum to less than $1.00 after taker fees, plus multi-outcome (neg-risk) events whose outcome asks sum under $1.00. The table is rebuilt every morning and you can re-run the scan live in your browser.
Every Polymarket market pays out exactly $1.00 per share to the winning outcome. If you can buy one share of every outcome for less than $1.00 in total — including taker fees — one of them must pay $1.00, locking in the difference. On a binary market that means buying YES and NO; on a multi-outcome event it means buying YES on every outcome.
Because Polymarket is efficiently arbitraged by bots. Market makers keep YES + NO priced at $1.00 plus the spread, and any genuine gap is usually competed away in seconds. On top of that, taker fees erase small gaps entirely — a pair summing to 99¢ can still lose money after fees. An empty scanner is the normal state of a healthy market.
Every row links to the market's live order book on Polymarket, where you can verify the real bid/ask before committing a cent. Open Polymarket →