How to Price Prop Overs Without the Over Tax
Beating a sportsbook's close on an over is worth far less than your tracker says.
In the video's test of more than 252,000 player props, every sportsbook overpriced overs, so an over priced only by sportsbooks carries a hidden markup and its CLV is better graded against a liquid exchange close.
Do not let sportsbooks alone price an over, check the money behind every exchange price, and grade your CLV against a liquid exchange close.
Sportsbooks build a markup into prop overs
Every book builds its cut, the vig, into both sides of a bet, and removing it gives a no-vig price: the book's honest estimate of how likely something is. The video removed the vig from 215,000 DraftKings player prop overs, which were priced to hit about 49.6% of the time and actually hit 48.1%. FanDuel, Caesars, BetRivers, Hard Rock and Bet365 all leaned the same way.
Kalshi and Novig, priced on the same numbers, were within about a third of a point. On a bet priced near 50%, a point and a half is about 3% of expected value, which is bigger than most edges people bet. The likely reason is that the public loves overs, so a sportsbook can charge a little extra and still get the action, while on an exchange there is no house setting the price.
Biggest over tax by stat (hit rate below the no-vig price)
The worst stat in the test.
Nowhere near the top, and receiving yards were about the same.
The tax is biggest on the overs people love to root for
By sport, MLB overs carried the largest tax at 2.1 points, then the NFL at 1.9 and the NBA at 1.7, while the NHL was basically zero. By stat, NFL rushing attempts and rushing yards were the worst, followed by NBA threes, MLB hits and NFL receptions.
That changes how you read a board. If an NFL rushing-yards over looks soft at one book and only other sportsbooks are pricing it, the over is about three points less likely than the board says, roughly 6% of expected value that was never there, so a bet showing +4% is closer to -2%. The video's quick rule for any over priced only by sportsbooks is to knock about 3% off the edge shown.
Why your CLV on overs can overstate your edge
Closing line value (CLV) assumes the closing price is right, so the video tested how much of it shows up in results on bets that beat DraftKings' close by at least two points. On 49,769 unders, 95% of the edge showed up, but on 53,495 overs only 33% did. FanDuel's close told the same story, 89% on unders and 34% on overs, so a 3% CLV on an over graded against a sportsbook close was really worth about 1%.
Graded against Novig's close instead, the over problem disappeared: beating it on overs was worth 142% of the edge across 2,856 bets, with unders at 109%. That is a smaller sample, but it fits the logic, because an exchange close is set by people trading their own money against each other and nobody is marking up the over. When the video regrades a set of NFL bets against Kalshi's price at kickoff, the CLV on the overs shrinks sharply while the unders stay the same.
If your bet tracker grades you against a sportsbook's close, your overs are probably being graded generously. Grade overs against a liquid exchange close such as Novig or Kalshi, not whatever closing line your tracker uses by default.
An exchange close only counts with real money behind it
The exchange rule breaks when a market is thin. On ProphetX, beating the close was worth about 52% of the edge when less than $100 had traded, 47% between $100 and $1,000, and 133% once $1,000 or more had traded. Novig showed the same pattern, about 79% under $1,000 on a smaller sample and 134% at $1,000 or more.
A price with $50 behind it has not really been tested, because nobody with real money has agreed with it yet. The median ProphetX prop in the data had only about $55 traded before the game, while 62% of the Novig props graded had $1,000 or more. On a thin market, treat the exchange like one more sportsbook.
In +EV Sniper, the number under each exchange price is how much money is sitting at that price right now. The video's example is an NBA over on 2.5 threes showing +12.6%, with $1 at the Kalshi price, $72 on Novig and $47 on ProphetX. Thin money on a high-tax stat is two strikes before you even consider the bet.
Set up +EV Sniper so exchanges help price your overs
Doing this by hand means pulling every book's price on every prop, removing the vig, checking which books price the over, checking the money behind each exchange price and saving your own closing lines. In +EV Sniper, the Matt's Prop Weights preset puts about 60% of the fair-value weight on Novig, Kalshi and ProphetX, because the exchanges barely carry the over tax.
Required books then keeps your fair price from coming from sportsbooks alone, which matters most on overs. Even with filters set, look at the money under the exchange prices before you decide, and treat a barely traded line like another sportsbook.
The +EV Sniper setup from the video
Open the weights panel
In +EV Sniper, click Filters, then click Weights next to fair value.
Apply the prop weights preset
Under Recommended, pick Matt's Prop Weights and click Apply.
Require an exchange
Under Required books, pick one exchange (Novig, ProphetX or Kalshi), especially when you are betting overs.
Raise the minimum sportsbooks
Set the minimum sportsbooks to three or more, then apply filters.
Check the money
Before you act, read the size under each exchange price and discount any line with almost nothing traded.
How many bets it takes for a real edge to show
With a real 3% edge on every even-money bet at equal stakes, the video's math still gives a 38% chance of being down after 100 bets and 17% after 1,000, and it takes about 3,000 bets to be 95% sure the edge is real. On sportsbook-graded overs where only a third of a 3% CLV is real, the edge is closer to 1%, which leaves a 38% chance of being down after 1,000 bets and needs about 27,000 bets to reach that 95% level.
Even a good under at even money loses close to half the time, which is how people end up quitting strategies that work and keeping ones that do not. Results take thousands of bets to show up, but CLV shows up right away, so it is the earliest read you get, as long as you grade it against the right books.
Article FAQs
What is the over tax?
It is the extra markup sportsbooks build into player prop overs. In the video's data, DraftKings overs priced to hit 49.6% after removing the vig actually hit 48.1%, and every sportsbook tested leaned the same way.
Should I stop betting prop overs?
Not necessarily. The problem is pricing and grading overs off sportsbooks alone. Graded against a liquid exchange close, the video found the over problem disappeared.
Which closing line should I grade my CLV against?
Grade overs against a liquid exchange close such as Novig or Kalshi rather than a sportsbook's close. If the exchange market had very little money traded, treat its price like another sportsbook.
How many bets does it take to prove an edge?
At even money, the video's math says a real 3% edge takes about 3,000 bets to be 95% sure, and a 1% edge takes about 27,000. Until then, CLV graded against the right close is the best early read.
Open +EV Sniper
Apply prop weights and required books so an exchange helps price every over before you decide on a bet.