Pinnacle Sports Betting for US Bettors: How the No-Vig Sharp Price Predicts the Close
If you bet from the US, Pinnacle is not your book. In a statement quoted by FRONTLINE and The New York Times in 2015, Pinnacle said the company "pulled out of the United States in 2007." The price still matters to you. Pinnacle's line, with the margin stripped out, is the best public estimate of two things: who wins, and where the market closes.
This guide covers why sharp books set the benchmark, the one identity linking your edge at bet time to your closing line value, what two large tests found when bettors bet against the market's own numbers, where the benchmark breaks, and three ways to read the number from the US without an offshore account.
Sharp Books and Soft Books Price Differently
Tadgh Hegarty and Karl Whelan of University College Dublin describe two business models in their study of European soccer odds, which Oxford Economic Papers ran in its January 2026 issue. Soft books keep high margins, spend on marketing, and restrict bettors who win. Sharp books run low margins on high volume, take bets from professionals and syndicates, and use those bets to shape their odds. The authors add one more detail. Sharp books such as Pinnacle sell their probability estimates to retail European books as market makers.
The difference shows up in the data. The working-paper version of the study covers 84,230 matches from 22 leagues, the 2011/12 through 2021/22 seasons. The journal version runs the same comparison on more than 150,000 matches. The soft-book home, away, and draw market showed a strong favorite-longshot bias. The sharp-dominated Asian Handicap market on the same matches showed none.
| Market | Who dominates the market | Loss rate, longshot bets | Loss rate, favorite bets |
|---|---|---|---|
| Home, away, draw | Soft retail books | 9.4% | 6.1% |
| Asian Handicap | Sharp books | 3.5% | 4.1% |
Both rows compare bets across the same range of win probability, roughly 26% to 48%. On soft prices, longshots lost 9.4 / 6.1 = 1.54 times as much per dollar as favorites, 54% more. On sharp prices, longshots lost 3.5% and favorites 4.1%, with no clear pattern across price levels.
Read the table as a rule for de-vigging. The standard proportional method assumes the book marked up every side by the same percentage. A sharp line comes close to fitting the assumption. A soft line breaks the assumption, and the de-vigged longshot still looks likelier to win than the truth. De-vig the sharp book, then compare. The de-vig methods guide covers the corrections for lopsided soft lines.
Golf matchup data, with US books in the sample, puts a number on the margin gap. Data Golf collected opening and closing odds on 106,204 PGA Tour and European Tour matchups and 3-balls from 11 books, starting in January 2019, and published the results in December 2020. Blind betting every matchup at Pinnacle lost 3.31%. The same strategy at DraftKings lost 7.01%. Strip each book's margin out and every book landed near zero.
| Book | Blind return | Return after removing the margin |
|---|---|---|
| Pinnacle | -3.31% | -0.01% |
| 5Dimes | -4.60% | +0.05% |
| BetOnline | -4.88% | +0.05% |
| Betcris | -5.59% | -0.02% |
| Bovada | -6.80% | +0.08% |
| DraftKings | -7.01% | -0.34% |
| bet365 | -7.08% | +0.06% |
The thinner margin is half the case for Pinnacle. Accuracy is the other half. Data Golf rated Pinnacle's closing lines the most accurate in its database and found the best blended forecast put 95% to 100% of the weight on Pinnacle. BetOnline, 5Dimes, and Bovada mostly copied Pinnacle or bet365. Pinnacle, DraftKings, bet365, and Betcris set their own prices.
The Identity: Your Edge Now Is Your Expected CLV
Start with the conversion. Implied probability is 1 divided by decimal odds. A two-way line sums past 100%. Divide each side by the total and you have the fair probability.
De-vig a sharp line and price your bet (example numbers)
Sharp line: Team A -150, Team B +140
Implied A: 150 / 250 = 60.000%
Implied B: 100 / 240 = 41.667%
Total: 101.667%, a 1.67% hold
Fair B: 41.667 / 101.667 = 40.98%, fair odds +144
Your book: Team B +155, decimal 2.55
EV = 0.40984 × 2.55 - 1 = +4.51%
At +145 the same bet carries +0.41%. At +140, the sharp book's own posted price, you lose 1.64% per dollar. The fair price of +144 is your break-even line.
Next, link the two numbers. Closing line value compares your price to the fair closing probability. In an efficient market, any predictable move already sits in the price, so today's fair probability is the best forecast of the closing fair probability. Your expected CLV equals your EV at bet time.
Why EV now and expected CLV match
CLV = fair closing probability × your decimal odds - 1
Expected CLV = expected fair close × 2.55 - 1
Efficient market: expected fair close = fair now = 40.98%
Expected CLV = 0.40984 × 2.55 - 1 = +4.51%
Any single close lands above or below the target. Two possible closes for the same +155 ticket:
| Sharp close | Fair B at close | CLV on your +155 | Sharp move |
|---|---|---|---|
| A -140, B +130 | 42.70% (+134) | +8.90% | +1.72 points toward you |
| A -165, B +150 | 39.11% (+156) | -0.26% | -1.87 points against you |
Across hundreds of bets, your average CLV converges on the average edge you measured at bet time, on one condition: the benchmark was efficient when you bet. The CLV guide covers tracking. The next two sections test the condition.
Two Tests of Betting Against the Market's Own Number
Lisandro Kaunitz, Shenjun Zhong, and Javier Kreiner ran a large version of this test. They collected closing odds on 479,440 soccer matches from 818 leagues, January 2005 to June 2015, across up to 32 books, Pinnacle among them. For each outcome they averaged the decimal odds across books, inverted the average, and checked the result against how often the outcome happened.
The consensus tracked reality almost one for one. Regression slopes came in at 1.003 for home wins, 1.012 for away wins, and 1.081 for draws, with R² values of 0.999, 0.998, and 0.995. After a small constant for the margin, the average price worked as a probability. Their rule: bet whenever the best decimal odds on offer exceeded 1 / (consensus probability - 0.05).
| Test | Bets | Return |
|---|---|---|
| Ten-year simulation at closing odds | 56,435 | +3.5% |
| Random picks at the best price, 2,000 runs | 56,435 each | -3.32% |
| Simulation 1 to 5 hours before kickoff, Sept 2015 to Feb 2016 | 6,994 | +9.9% |
| Paper trading plus real money, $50 a bet | 672 | +6.2% |
Recompute the returns from the paper's own dollar figures. A $98,865 profit on 56,435 bets of $50 is 3.50%. A $34,932 profit on 6,994 bets is 9.99%. A $2,086 profit on 672 bets is 6.21%. Then the books responded. Within months of real-money betting, books cut the stakes they would take, held bets for manual review, or refused them. The authors stopped.
Data Golf ran the same idea with Pinnacle alone. The site treated Pinnacle's margin-free closing odds as the fair price and bet every other book's closing odds whenever the expected value cleared a threshold.
| Minimum EV against Pinnacle's close | Bets | Expected ROI | Realized ROI | Share of edge kept |
|---|---|---|---|---|
| 0% | 4,803 | 2.28% | 1.84% | 81% |
| 1% | 3,040 | 3.34% | 3.07% | 92% |
| 2% | 1,902 | 4.47% | 3.75% | 84% |
| 3% | 1,140 | 5.82% | 2.03% | 35% |
| 4% | 743 | 7.07% | 1.08% | 15% |
| 5% | 511 | 8.25% | -0.59% | below zero |
Share kept is realized ROI divided by expected ROI, for example 1.84 / 2.28 = 81%. Up to a 2% threshold, the Pinnacle close predicted the result almost one for one. Data Golf called the fit a rough one-to-one match between expected and actual ROI. From 3% up, the kept share dropped to 35% and lower. In this sample the biggest gaps against a sharp close were the worst bets. Treat a gap this size as a reason to recheck the line, the market rules, and the news, not a reason to bet bigger.
Openers Are a Weaker Benchmark
Run the same test on opening lines and most of the edge disappears. Data Golf bet Pinnacle's margin-free openers against every other book's openers.
| Minimum EV against Pinnacle's open | Bets | Expected ROI | Realized ROI | Share of edge kept |
|---|---|---|---|---|
| 0% | 6,098 | 3.36% | -1.09% | below zero |
| 1% | 4,486 | 4.41% | 0.30% | 7% |
| 2% | 3,259 | 5.52% | 0.13% | 2% |
| 3% | 2,294 | 6.80% | 0.50% | 7% |
| 4% | 1,651 | 8.09% | 0.64% | 8% |
| 5% | 1,227 | 9.35% | 1.29% | 14% |
| 6% | 948 | 10.49% | 1.73% | 16% |
| 7% | 720 | 11.77% | -1.96% | below zero |
| 8% | 568 | 12.93% | -2.59% | below zero |
| 9% | 429 | 14.39% | -1.94% | below zero |
| 10% | 338 | 15.71% | 3.13% | 20% |
The opener promised a 3.36% edge and the bets lost 1.09%. No threshold from 1% to 10% kept more than 20% of the promised edge, and the bets lost money at 7%, 8%, and 9%. A sharp opener is an early estimate. Pinnacle's prices moved the most of any book from open to close in the Data Golf sample, the pattern you expect from a book using early bets to find its number. Unabated's August 2026 primer on market makers describes Pinnacle's early years the same way: Pinnacle took in information from bettors and shaped its lines, and other books tried to move in lock-step. The primer also says Pinnacle now shares the market-maker role with other books, Betcris among them.
The leader also changes by market. On 1,401 golf matchups where the two books' margin-free openers differed by at least 5%, Pinnacle moved 54.7% of the way toward the Betcris opener by the close. Betcris moved 15.6% of the way toward the Pinnacle opener. Data Golf flags a timing caveat: on 72-hole matchups, about 70% of the sample, Pinnacle posted openers before Betcris. Pinnacle is the default benchmark, not the leader in every market. In a niche market, watch which book moves first.
Chasing a move after the fact is a separate trade. Tobias Moskowitz tested asset-pricing patterns on sports betting markets in the Journal of Finance in 2021. He found momentum in prices, consistent with delayed overreaction, with returns too small to cover transaction costs. A line move tells you where the sharp number went. The value sat in the gap before the move.
- Near the close, a sharp no-vig price is an estimate. In the Data Golf closing sample, gaps of 1% to 2% kept most of their edge.
- Against an opener, a gap is a hypothesis. Demand a bigger gap, or wait for the market to confirm.
- Against any sharp price, a huge gap is a prompt to check the data first.
Skip the Offshore Account
The number is free. The account carries the risk. The same 2015 reporting said American and European investigators had found thousands of US betting customers at Pinnacle since 2007, despite the company's statement. A 2012 Queens money-laundering probe charged 25 people tied to offshore sportsbooks. Prosecutors did not charge Pinnacle itself.
Congress passed the Unlawful Internet Gambling Enforcement Act in 2006, the year before Pinnacle's exit. Section 5363 of Title 31, added on October 13, 2006, bars anyone in the betting business from knowingly accepting credit, electronic fund transfers, or checks tied to another person's unlawful internet gambling. The law targets the payments. You need none of them to read a price.
Three Ways to Read the Sharp Number Without an Offshore Account
1. Odds screens and data feeds
Odds feeds carry Pinnacle's prices. The Odds API, a sports data provider, lists Pinnacle in its EU region with a note on delay: the odds come from Pinnacle's public website and carry a possible lag. A stale sharp number creates fake edges. Compare timestamps before you bet a gap. A 2% edge against a price from 20 minutes ago is an edge against an old opinion.
2. US exchange prices
The same feed lists Kalshi, Polymarket, Polymarket US, Novig, and ProphetX in its US exchange region. An exchange contract pays $1 if the event happens, so the price in cents is a probability. Add the best YES price and the best NO price, divide each by the total, and you have a no-vig fair number, the same way you de-vig a sportsbook line.
Exchange prices to a fair number (example numbers)
YES at 52¢, NO at 50¢: total 102¢, a 2.0% overround
Fair YES: 52 / 102 = 50.98%, fair odds -104
Taker fee: 0.07 × $0.52 × $0.48 × 100 contracts = $1.7472, rounded up to $1.75
Cost per contract with fee: $0.5375, decimal 1.8605, about -116
NO at 50¢: $0.50 + $0.0175 fee = $0.5175, decimal 1.9324, about -107
BettingPros' October 2026 explainer gives Kalshi's taker fee as 7% × price × (1 - price) per contract, rounded up to the next cent. The explainer's figures round the order total: $1.75 per 100 contracts at 50¢, and a full cent on a single 10¢ contract. On NFL game, spread, and total markets, a filled resting order pays a quarter of the taker fee. On most other markets, a resting order pays nothing. The fee changes what the exchange costs as a place to bet. The fair number stays the benchmark either way. Check a DraftKings or FanDuel line against Kalshi or Polymarket prices, even when you place the bet at the book. The prediction markets comparison covers the cost of trading there.
3. A consensus of independent books
The Kaunitz test used no single sharp book. The average across many books did the job. Build the same consensus from US books with one adjustment from the Data Golf work: count only books setting their own prices. A book copying Pinnacle or bet365 adds the same opinion twice. In golf, DraftKings priced on its own and added a bit of predictive value to other books' closing lines, perhaps none to Pinnacle's.
Our homepage shows this check in a real Pro Sharp Bot card from October 18: Titans moneyline +290 at BetRivers against a sharp +262, fair 26.6% against a market 25.6%, from 3 sharp books and 10 public ones. Recompute the card. The +290 price implies 100 / 390 = 25.64%. EV = 0.266 × 3.90 - 1 = +3.74%, the card's EV +3.7%.
Calculator: Edge Now, CLV Later
Enter the sharp two-way price on both sides, your odds on Side A, and, once the game starts, the sharp closing price. Switch the format to cents for exchange prices. The defaults load the worked example, with Team B entered as Side A. The edge line is your expected CLV, provided the sharp price was efficient when you bet.
Sharp benchmark: bet-time edge and closing CLV
The Workflow, Bet by Bet
- Pull the sharp two-way price with a timestamp. Run both sides through the no-vig calculator for the fair number. The odds converter handles decimal and implied probability.
- Compare the fair price to the best price across your books. The best price sets your edge.
- Bet gaps against a near-close sharp price. Against an opener, require a bigger gap or wait.
- Size from the fair probability. In the worked example, full Kelly on +155 against a 40.98% fair price is (1.55 × 0.40984 - 0.59016) / 1.55 = 2.91% of bankroll. Quarter Kelly is 0.73%. The Kelly calculator runs the numbers, and the Kelly guide explains the fractions.
- Log the sharp close on every bet and grade the CLV. Judge the process on average CLV across hundreds of bets, since one result tells you little.
- Expect limits. Books limited the Kaunitz team within months of real-money betting. The limits guide covers how books flag accounts, and the line-setting guide covers where retail lines come from.
Look around the free Discord room first, no card needed.
The verdict: Pinnacle's no-vig price forecasts the result and the close. Your edge against the sharp price at bet time is your expected CLV. Against Pinnacle's closing lines, Data Golf's bets kept 81% to 92% of the predicted edge at thresholds up to 2%. Against openers, the same method lost 1.09%. Kaunitz and coauthors made 3.5% across 56,435 simulated bets with the market's own consensus, and books limited them once real money went in. You need no offshore account. Read the sharp number from a feed, an exchange, or a consensus of independent books, bet the gap near the close, and grade every bet by CLV.
Sources and Method
Hegarty and Whelan (Oxford Economic Papers, 2026, and their 2023 working paper) supply the sharp and soft business models, the market-maker role, and the loss-rate comparison between the home, away, and draw market and the Asian Handicap market. Kaunitz, Zhong, and Kreiner (2017) supply the consensus calibration, the betting rule, the three test results, and the account limits. Data Golf's December 2020 study supplies the golf blind returns, the copying and independent books, the closer and opener threshold tables, and the Betcris movement figures. Moskowitz (Journal of Finance, 2021) supplies the momentum finding. Unabated supplies the market-maker description. FRONTLINE and The New York Times supply Pinnacle's 2007 statement and the Queens probe. Cornell's Legal Information Institute supplies the text and date of 31 U.S.C. 5363 and the act's short title. The Odds API supplies the feed regions. BettingPros supplies the Kalshi fee formula. The Pro Sharp Bot card comes from our homepage.
Every conversion, fair price, EV, CLV, Kelly fraction, fee, and share-kept figure above comes from the formulas shown, and you recompute each one in a spreadsheet. The worked lines and exchange prices are example numbers, not live markets. Limits: the Data Golf tables cover golf matchups and the Kaunitz tests cover soccer, so treat the exact thresholds as a guide for other sports. Proportional de-vigging assumes an even margin, which fits sharp lines better than soft ones.