HandsDown AIToday's best bets
Our three straight bets, led by what's been paying: a player prop, a spread and a winner. Each one updates with injury news until 25 minutes before the game (after NBA starting lineups are in), then it's locked and tracked. The spread and winner only show up when our chance beats what the price needs to profit.
Loading today's picks…
What's paying
Profit by bet type since Oct 22. Early days, so small samples swing a lot.
For fun: long shots
Parlays pay big, but the book's cut stacks on every leg, so they lose money over time. They're tracked separately from the best bets. See today's parlays →
Yesterday
Best bets record
Upcoming games
Our predicted margin, win chance and spread pick for each game, graded in the Record tab. A gold stripe means our model on its own disagrees with Vegas by 4 to 6.9 points.
Player props
For each NBA game, the over or under DraftKings thinks is most likely to hit (no worse than -300), with our projection next to it. A game only gets a prop when that side is at least 65% likely.
Parlays
Long shots for fun. Parlays pay big, but the book's cut stacks on every leg, so over time they lose money. Every one is locked 25 minutes before its first game and tracked here, wins and losses.
Today's long shots
The best-value parlay across all sports, and a same-game parlay: our spread pick plus a player prop.
Parlays by sport
The best 2-game and 3-game spread parlays per sport.
Today's results
Live scores for every game we picked, and how each bet is doing right now. Scores refresh every minute.
Yesterday's recap
Track record
The headline is our straight best bets (prop, spread, winner). Parlays are long shots, tracked on their own. The by-sport records grade the "Our prediction" pick and "Our spread pick" shown on each game; college spreads are leans, not bets. Beating the spread takes about 52.4% to break even.
Profit by bet type
By sport
Winners: did the team "Our prediction" favored win the game?
Spread picks: every game, the side "Our spread pick" takes against the line. Our model on its own always leans the same way; blending with Vegas only changes how strongly.
Gold-edge picks: only the games where our model on its own disagrees with the line by 4 to 6.9 points (the gold edge on the Games page). This is the test of whether the big disagreements pay off.