Last updated: July 13, 2026. This article is educational analysis, not betting advice. Betting involves risk; only adults 21+ should wager, and only with money they can afford to lose.
Here is the least advertised fact in American sports betting: if you get good at it, most sportsbooks will stop taking your bets. Not because you cheated. Because you won — or because their models predicted you would. This article explains how limiting works, what the public reporting shows, how fast it happens, and why prediction-market exchanges like Kalshi and Polymarket have no economic reason to do the same thing. It ends with the honest caveats about exchanges, because they have real costs too.
The practice: legal, routine, and mostly invisible
A sportsbook is a counterparty. Every bet you place, the book takes the other side. If you consistently bet at better prices than the market's eventual consensus, you are a projected long-term cost — so the book cuts your maximum stake, sometimes to a few dollars, or closes the account outright. In most US jurisdictions this is legal: books are private businesses and their house rules generally let them refuse action from anyone (as of July 2026; rules vary by state, so check your jurisdiction).
This is not a fringe claim. The Washington Post reported in 2022 on how major US operators use limits to restrict sharp customers. In September 2024, ESPN covered a Massachusetts Gaming Commission roundtable where representatives of BetMGM, Caesars, DraftKings, FanDuel, Fanatics, Penn and Bally's openly defended the practice of restricting how much sharp customers may wager — operators said limits target advantage-play patterns such as betting mistake lines or having a better model, not winning in itself. The debate regulators are having is not whether limiting happens — the operators confirmed it does — but whether it should be disclosed and constrained.
Why books limit: the math of the counterparty model
A standard two-sided market is priced at -110 / -110. Each side implies 110 ÷ 210 = 52.38%, so the two sides sum to 104.76% — a 4.76% overround. If the book balances $110 on each side, it collects $220, pays the winner $210, and keeps $10, about 4.55% of handle. That is the business: earn the vig on balanced recreational flow.
A sharp bettor breaks this. Sharps don't bet randomly; they bet only when a price is wrong — stale lines, slow-moving props, mispriced parlay legs. Books measure this with closing line value (CLV): if you bet a team at +150 (implied 40.0%) and the line closes at +120 (implied 45.45%), you beat the close by more than five percentage points. Do that repeatedly and the book's risk models flag you long before your account is even profitable. Multiple industry accounts note that bettors get limited while down money — it's the CLV pattern, not the bankroll, that triggers the flag.
How fast do limits arrive?
Faster than most people expect. Public reporting and bettor accounts describe:
- Limits arriving within days to a few weeks of a detectable sharp pattern — especially on player props and derivative markets, where books are most vulnerable.
- Cuts that are severe, not gentle: a flagged account may be allowed roughly a tenth or less of what a fresh account can bet. A market that takes $2,000 from a new customer might take $200 — or $20 — from a limited one.
- Market-by-market limiting: an account restricted to pocket change on NBA props may still be allowed normal stakes on mainline NFL sides, because the book's exposure differs by market.
The uncomfortable corollary: the skills that make betting +EV — line shopping, de-vigged fair-price comparison, beating the close — are exactly the signals that get accounts restricted. Any honest +EV education (including ours) has to say this out loud.
Why exchanges don't limit winners
Now the interesting part. Kalshi and Polymarket US are not sportsbooks. They are CFTC-regulated exchanges — Designated Contract Markets, the same legal category as commodity futures exchanges. The distinction matters mechanically, not just legally:
- A sportsbook takes the other side of your bet. Your win is its loss, so it has both the motive and the right to refuse your action.
- An exchange matches traders. When you buy 100 "Yes" contracts at 52 cents, some other trader (or market maker) sold them to you. The exchange holds no position; it earns a transaction fee whichever way the contract settles.
Your winning is not the exchange's loss — it's the exchange's revenue. A high-volume winner generates more fees, the way a profitable stock trader generates more commissions for a brokerage. There is no business reason to limit winners, and as regulated exchanges with impartial-access obligations, singling out profitable traders for exclusion would run against the model entirely. The constraint you face on an exchange is different: liquidity. You can buy only as many contracts as other traders are willing to sell at your price. Nobody bans you for being right; the order book just runs out.
A worked example on an exchange
Suppose de-vigged consensus across books says a team wins 55% of the time, and Kalshi's order book offers "Yes" at 52 cents. You take 100 contracts:
- Cost: 100 × $0.52 = $52.00.
- Fair value: 100 × $0.55 = $55.00, so your gross edge is $3.00.
- Trading fee: Kalshi charges a variable taker fee that peaks around the middle of the price range — roughly 1.75 cents per contract near 50 cents, less at extreme prices (as of July 2026; see the current Kalshi fee schedule). Call it about $1.75 here.
- Net expected edge: roughly $3.00 − $1.75 = $1.25, about 2.4% of your $52 stake.
The edge survives the fee — barely, in this example — which is why fee-aware pricing matters. EdgeFinder displays Kalshi quotes net of the trading fee for exactly this reason, and our free EV and no-vig calculators let you rerun this arithmetic on any line. For a deeper treatment, see Kalshi fees explained.
The honest caveats about exchanges
If exchanges never limited winners and were otherwise identical to books, this article would end here. They aren't identical:
- Thinner liquidity. Major markets (NFL sides, marquee games) are deep; niche props and smaller leagues can be shallow or nonexistent. A book might take your $5,000 prop bet instantly; the exchange order book might only have $300 at your price.
- Spreads. On an exchange you pay the bid-ask spread. In illiquid markets the spread can eat an edge the way vig does at a book.
- Fees. Kalshi's taker fees are real money, especially near 50 cents. Always compare prices net of fees.
- Narrower menus. Sportsbooks list vastly more props, alt lines and same-game parlays than exchange markets cover as of July 2026 — though exchange coverage has expanded rapidly.
- Unsettled legal terrain. Kalshi operates under federal CFTC regulation, but several states dispute whether sports event contracts are lawful there; sports markets are restricted or contested in a number of states, litigation is active in both directions (including a July 2026 New York federal ruling against Kalshi, now on appeal, and an April 2026 Third Circuit ruling in its favor), and Polymarket's US platform, launched in December 2025 under a CFTC-licensed exchange, faces its own state challenges. All of this is as of July 2026 — check your jurisdiction before trading.
What this means for a winning bettor
The practical takeaway is not "exchanges good, books bad." It's an accounting question. At a sportsbook, your ceiling is set by the operator's tolerance for your action, and for a winning bettor that ceiling drops fast. On an exchange, your ceiling is set by market liquidity, which grows as more traders show up — and no one revokes your access for being right. A comparison of the two models in full is in Kalshi vs. sportsbooks.
Whatever venue you use, the discipline is the same: know your fair price, count every fee, and track your results honestly. We hold ourselves to that standard too — every EdgeFinder model pick is logged pregame and graded publicly, wins and losses alike, at /record. No screenshots of winners only. That's the whole point.