Sports Betting: The 50/50 Myth, and Who's Really Winning Through Three Weeks of College Football

Last version of this explainer got the basic mechanic a little too clean, so let's fix that first, because it matters.

The textbook explanation goes like this: a sportsbook wants exactly half the money on each side of a bet, so it collects its small fee no matter who wins, like a casino raking a poker table. Real people who run sportsbooks say that's basically not true. Nick Bogdanovich, a longtime oddsmaker who ran risk for William Hill and Caesars, put it bluntly: that's a fallacy, it never happens that way, the books always have decisions. Jay Kornegay, who runs the Westgate SuperBook in Vegas, called the 50 50 idea an urban legend. Chris Andrews, who directs the sportsbook at South Point, said it plainly: there's almost always more money on one side than the other, either naturally or by design.

That's the old school Vegas view. The modern, publicly traded version of the same admission came straight from the top of DraftKings. CEO Jason Robins said out loud what old bookmakers only used to say in back rooms: this is an entertainment activity, people who are doing this for profit are not the players we want. That's not a company describing itself as a neutral middleman collecting a small fee off balanced money. That's a company saying it wants to choose its customers, not just price a game and let anyone bet either side.

So if books aren't just trying to split the money down the middle, what are they actually doing? Three things, and all three leave fingerprints you can see if you know where to look. First, when a book believes the crowd betting a team is wrong, it will happily sit on a lopsided position on purpose, betting against the public with its own money instead of just collecting a fee. Second, when it does move something, it's usually reacting to a small number of sharp, professional bettors, not the large number of casual ones, even when those two groups disagree about which way to go. Third, and this is the one people ask about most: if a book decides it really doesn't like a customer, it can cap how much that person is allowed to bet, or in rarer cases cut them off completely.

Is that common? It depends which of those two things you mean, and the real numbers are surprising either way. Outright bans, cutting a customer off entirely, are actually rare, state gaming regulators in Massachusetts found just over one half of one percent of bettors get banned, and Wyoming's regulator found under one percent. But quietly capping how much a good bettor can wager, without banning them outright, is far more common and far less tracked, one investigation into the practice put the number as high as one in ten accounts at some books. DraftKings, BetMGM, and PointsBet get named most often as the strictest limiters, while Caesars and WynnBET have reputations for letting winners keep betting bigger. It's a big enough deal now that New York has a bill, the Fair Play Act, that would make it illegal to limit someone just for winning too much, and Massachusetts now requires books to explain why a customer got limited at all. So: banning is rare, quietly shrinking your max bet is not.

Here's how you actually spot a book taking a side. If the number moves toward the side getting fewer bets, that's called reverse line movement, and it means the book is siding with the smaller, sharper crowd over the bigger, louder one. It really happened in Week 1: Wake Forest was getting only 38 percent of tickets against Akron, yet the number moved further in Wake Forest's favor, from 22.5 points to 24.5. Houston got just 29 percent of tickets against Oregon State, and the line still grew in Houston's favor, from 18.5 to 20.5. Both times, the book leaned away from the crowd and toward the smaller side, which is the opposite of balancing anything.

Now here's who actually won and lost money with real dollars behind it. Week 1, LSU didn't just beat Clemson, they demolished them 51 to 10 as the favorite, an easy cash for anyone who bet LSU. Then Michigan needed a miracle finish to survive Western Michigan, and BetMGM's trading manager, Adair Horne, admitted it straight: we went from a big winner to a loser. The book had six times more money riding on Michigan than the other side, and one customer turned a 10,000 dollar bet on Michigan at long shot odds into a 280,000 dollar payday on the final play. Bettors won both of those, big.

Books got theirs back the same week, though. Colorado was one of the most popular teams in the country to bet on, and Georgia Tech beat them outright, 14 to 13. All that public money on Colorado just vanished into the book's pocket. Then Week 3 flipped the script hard: LSU, fresh off destroying Clemson, lost outright on the road as a favorite to Ole Miss, 32 to 24, and Texas A&M, favored by about two and a half touchdowns at home, lost outright to Kentucky, 31 to 21. Betting the big name favorite is the single most common thing a casual bettor does, and both of those bets failed on the same weekend.

Put it all together and here's the honest read: the book always wins story is too simple, and so is the public always loses story. What's actually true is that books take real, on purpose risks against the crowd, actively manage who gets to keep playing, and this season the actual results have gone both ways so far, close to even. Three weeks isn't enough games to call a winner in that fight, but it is enough to throw out the old explanation of how the fight works in the first place.