Skip to main content
ABC News
Our New Live In-Game Win Probabilities Are Here For The NBA Playoffs

The NBA playoffs are almost here, with the play-in tournament tipping off later tonight. But while the regular season was careening toward a dramatic finish, we were hard at work creating a live in-game win probability model — and we’re pretty damn excited to share it. We officially added this model to our forecast page this morning, and you can read more about the ins and outs of how it works here.1

With this shiny new feature, we can see how likely each team is to win every game in real time — and how their odds to bring home the Larry O’Brien Trophy are impacted by each play as well. We’ve even backfilled in-game win probabilities for all 1,230 games played in the 2022-23 regular season. So to give you a sense of what to expect from the interactive during the playoffs, let’s take a look back at some of the most exciting and memorable games this regular season to see what our new win probability model thought about each contest.

Donovan Mitchell’s explosive performance in the Cleveland Cavaliers’ improbable overtime win over the Chicago Bulls on Jan. 2 was one for the books. Not only were his 71 points the eighth-highest game total in league history and most in Cavaliers team history, they also propelled an inspired comeback. The Bulls took the reins as in-game favorites early on, first earning the edge up 32-26 with 31.9 seconds left in the first quarter, with an Alex Caruso steal2 pushing their win probability up to 54 percent. The Bulls would hold onto their win probability advantage for the next solid three quarters. But the Cavs clawed back to tie the score, and Bulls guard Zach LaVine missed a 28-foot stepback 3-pointer with 50.4 seconds left in the fourth quarter, giving the Cavs a 55-45 percent edge in win probability.

The Bulls quickly wrestled back the advantage, though, earning a 130-127 lead and 91 percent win probability with 7.4 seconds left in regulation. But sending Mitchell to the line a few seconds later came back to bite the Bulls. After making his first free throw attempt with 4.4 seconds left in the game, Mitchell intentionally missed his second attempt, corralled his own offensive rebound and made his quick-release putback attempt. This game-tying flurry with the clock stopped improved the Cavs’ win probability by a whopping 38 percentage points. After Bulls guard DeMar DeRozan airballed a buzzer-beating 3-pointer, the game went into overtime — where the Cavs scored the first 10 points and cruised to an 11-point win, capping off one of the more exciting and drama-filled comebacks in the league this season.

Luka Dončić’s 60-point triple-double and last-second, game-tying bucket that fueled the Dallas Mavericks’ wild comeback against the New York Knicks on Dec. 27 might be one of the few bright spots for Mavs fans in what was an otherwise incredibly disappointing regular season. When Dončić stepped to the line to shoot two free throws with 4.2 seconds left at the American Airlines Center, the Mavericks were down by three points with only a 7 percent chance to win the game according to our new model. But Luka hit his first free throw,3 intentionally missed the second,4 corralled the loose ball5 and hit a game-tying bucket6 in a dramatic sequence that ended up forcing overtime.7 Finishing the season outside the play-in tournament was a disaster for a franchise with title aspirations, but Mavs fans can at least look back fondly on this game and their superstar’s performance.

The marathon double-overtime game between the Sacramento Kings and Los Angeles Clippers on Feb. 24 had pretty much everything you could want in an NBA game8 — two overtimes, multiple dramatic moments,9 a close finish and the second-highest point total in NBA history. The Kings may have not outplayed the Clippers throughout their game at Arena, but they certainly outlasted them.10 They fought for four full quarters and almost two full overtimes as underdogs before our win probability model thought they had a better than 50 percent chance to win, when they were up 176-175 with 29.8 seconds left in double overtime and Kawhi Leonard missed a 17-foot stepback jumper. This impressive win showcases the kind of stick-to-itiveness this young, exciting Kings team needed to finally end a 16-year playoff drought, the longest in NBA history.

With 10.9 seconds left in the third quarter of the Los Angeles Lakers’ Feb. 7 clash with the Oklahoma City Thunder at Arena, LeBron James hit a 14-foot fadeaway jumper — a shot he’s worked so hard to add to his scoring arsenal throughout his career — to break Kareem Abdul-Jabbar’s all-time NBA scoring record. These 2 points were a landmark in the career of one of game’s all-time greats, but ultimately came in a futile effort for the Lakers. After James’s historic shot, the Lakers were down 104-99 with only a 33 percent chance to win according to our model. They briefly retook the edge in the first few minutes of the fourth quarter, but ultimately let the game slip away. It was a discouraging loss for a superstar-laden team that came into the year with a title-or-bust mentality, but now finds itself needing to survive the play-in tournament to avoid missing the playoffs for a second straight season.

The Chicago Bulls’ double-overtime road win against the Philadelphia 76ers on March 20 was a thrilling, back-and-forth contest. The Sixers came into the game on an eight-game winning streak and were pretty hefty favorites according to our model, which gave them an 80 percent chance to win. But Joel Embiid, James Harden and company ran into a frisky Bulls team fighting for one of the last play-in spots in the East. The two teams went back and forth all game, with neither team leading by more than eight points throughout the entire game.

Our model’s favorite flipped back and forth a whopping 31 times during the second half and both overtimes. Ultimately, Zach LaVine’s offensive rebound after a Patrick Beverly missed 3-pointer with 57.8 seconds left in double OT turned the tide in Chicago’s favor for good. The Bulls held a 55 percent win probability after the rebound, and kept the edge until they finished off the Sixers by four points. It was a crucial win that helped them stay afloat in the playoff hunt in the Eastern Conference at the time.11

Those were just a handful of the games from a wildly entertaining NBA season. Lucky for us, our new win probability model does the same thing for each and every contest, analyzing how each play influenced each team’s odds to win. It will also be updating live over on our NBA forecast page during upcoming play-in and playoff games, allowing you to see how each turnover, stepback three and buzzer-beater affects your favorite team’s12 chances to be crowned world champions. We hope you’ll follow along with us!

Check out our latest NBA predictions.


  1. We’re particularly excited that it’s a pretty novel approach to modeling live in-game win probabilities. We use Poisson distributions and a tree-based endgame model instead of the more common logistic regression approach.

  2. Taking the ball from who else but Donovan Mitchell.

  3. Improving their odds to 8 percent.

  4. Improving their odds to 9 percent.

  5. Improving their odds to 13 percent.

  6. Improving their odds to 46 percent.

  7. Which our model thought the Mavericks had a 54 percent chance to win.

  8. Except maybe a win if you’re a Clippers fan.

  9. Including Malik Monk hitting a game-tying 3-pointer with 2.1 seconds left in the fourth quarter and De’Aaron Fox scoring a game-winning pullup with 36 seconds left in double OT

  10. Which, as Jeff Probst would attest, can be just as important in games like basketball or “Survivor.”

  11. They ultimately finished the regular season in the 10th spot in the East, needing to win two straight road play-in games to make it into the postseason for the second straight year.

  12. Or your biggest rival’s.

Ryan Best was a visual journalist for FiveThirtyEight.

Jay Boice was a computational journalist for FiveThirtyEight.


Latest Interactives