Machine Learning Ball (@MLBallForecast) header
Machine Learning Ball (@MLBallForecast) profile picture

Machine Learning Ball

@MLBallForecast

Probabilities, published in advance. Graded in public. MLB · NFL · NBA · MLBarometer. Statistics-only. Not picks.

 Barnsdall, OK
machinelearningball.com
Joined January 2026
  • Tweets40.2K
  • Following60
  • Followers938
29.2KMedia
Photo by Machine Learning Ball (@MLBallForecast)Photo by Machine Learning Ball (@MLBallForecast)Photo by Machine Learning Ball (@MLBallForecast)Photo by Machine Learning Ball (@MLBallForecast)Photo by Machine Learning Ball (@MLBallForecast)Photo by Machine Learning Ball (@MLBallForecast)Photo by Machine Learning Ball (@MLBallForecast)Photo by Machine Learning Ball (@MLBallForecast)Photo by Machine Learning Ball (@MLBallForecast)Photo by Machine Learning Ball (@MLBallForecast)
NBA scorecard: Machine Learning Ball vs ESPN's BPI and DARKO, 2026-27 season, games of Oct 20 through Nov 8. How it's graded Every game uses our morning forecast, locked before tip, and ESPN's BPI game predictor captured the same morning. Same games, same morning, graded on both sides. No game is dropped or filled in after the fact. The main measure is the Brier score: the squared gap between the win probability and what happened (1 for a win, 0 for a loss), averaged over every game. Lower is better. A 50/50 call on every game scores 0.250. Games: us vs ESPN's BPI (146 games, all 146 graded on both sides) Brier score: 0.1885 vs 0.1880 Winners called: 105-41 (71.9%) vs 103-43 (70.5%) The two disagreed on the winner in 8 games. We went 5-3 in those. Games where our forecast was 70% or higher: 42-6 (ESPN also went 42-6 in those 48 games) Oct 20 to Oct 25 (43 games): ESPN 0.188, us 0.192 Oct 26 to Nov 1 (49 games): us 0.200, ESPN 0.207 Nov 2 to Nov 8 (54 games): ESPN 0.171, us 0.175 Through Nov 8 the season gap is 0.0005 in ESPN's favor, well inside the noise. The 95% interval runs from us 0.008 better to ESPN 0.009 better. The two are level. Players: us vs DARKO (2,830 player-games, points and assists) These are the player-games where DARKO's board was built from this season's data. The first three slates of the season, when its board still carried last season's minutes, are left out. Points, average miss: DARKO 4.40, us 4.57. DARKO is ahead, it was closer on 53.9% of player-games, and it has been ahead every week. This gap is outside the noise. Assists, average miss: us 1.27, DARKO 1.27. Even. Our player numbers are medians, the middle of each player's range. For points that runs below the average: our numbers came in 2.1 points under actual per player-game, and DARKO's came in 0.9 under. Rebounds aren't compared because DARKO doesn't publish them. Every game and every player-game is graded the same way, and the scorecard is updated each week. Sources: ESPN's BPI game predictor and DARKO's public player projections, each captured the morning of every slate. #NBATwitter #NBA machinelearningball.com
Machine Learning Ball — Statistical Forecasting Firm

Machine Learning Ball — Statistical Forecasting Firm

A free, public, statistics-only forecasting firm. Probabilities for MLB, NFL and NBA, graded in public.

machinelearningball.com
89
NBA forecasts, Wed Nov 11: 10 games. Win probabilities, margin and total projections, and player projections for every game, issued 7:31 AM CT from 100,000 simulated games per matchup. Full slate and player boards: machinelearningball.com/nba/ Game forecasts in the replies. #NBATwitter #NBA
Image from Machine Learning Ball (@MLBallForecast)
10
182
GS at HOU, 8:30 PM CT Win probability: HOU 67%, GS 33% Margin: HOU by 6 80% range: GS by 11 to HOU by 23 Total: 224 80% range: 201 to 248 Players to watch (MIN / PTS / REB / AST) Jabari Smith Jr.: 34 / 15 / 5 / 2 Kevin Durant: 33 / 21 / 5 / 3 Amen Thompson: 32 / 14 / 5 / 3 #NBATwitter #NBA #DubNation #Liftoff
Image from Machine Learning Ball (@MLBallForecast)
90
GS at HOU: Confirmed, 8:02 PM CT Win probability: HOU 67%, GS 33% Margin: HOU by 6 80% range: GS by 10 to HOU by 23 Total: 221 80% range: 198 to 244 Players to watch (MIN / PTS / REB / AST) Jabari Smith Jr.: 33 / 14 / 5 / 2 Kevin Durant: 33 / 20 / 5 / 4 Alperen Sengun: 32 / 19 / 8 / 6 Updated with confirmed starting lineups; the morning forecast stays on the site. #NBATwitter #NBA #DubNation #Liftoff
Image from Machine Learning Ball (@MLBallForecast)
78
ORL at UTAH, 8:00 PM CT Win probability: UTAH 40%, ORL 60% Margin: ORL by 4 80% range: ORL by 21 to UTAH by 13 Total: 218 80% range: 194 to 241 Players to watch (MIN / PTS / REB / AST) Franz Wagner: 30 / 18 / 4 / 4 Desmond Bane: 30 / 15 / 3 / 3 Paolo Banchero: 30 / 18 / 6 / 4 #NBATwitter #NBA #MakeItMagic #TakeNote
Image from Machine Learning Ball (@MLBallForecast)
87
ORL at UTAH: Confirmed, 7:34 PM CT Win probability: UTAH 42%, ORL 58% Margin: ORL by 3 80% range: ORL by 20 to UTAH by 13 Total: 217 80% range: 194 to 240 Players to watch (MIN / PTS / REB / AST) Franz Wagner: 34 / 22 / 5 / 5 Desmond Bane: 33 / 18 / 3 / 3 Keyonte George: 31 / 21 / 3 / 5 Updated with confirmed starting lineups; the morning forecast stays on the site. #NBATwitter #NBA #MakeItMagic #TakeNote
Image from Machine Learning Ball (@MLBallForecast)
144
BKN at NO, 7:00 PM CT Win probability: NO 62%, BKN 38% Margin: NO by 5 80% range: BKN by 12 to NO by 21 Total: 231 80% range: 207 to 254 Players to watch (MIN / PTS / REB / AST) Michael Porter Jr.: 30 / 20 / 6 / 2 Julius Randle: 30 / 18 / 5 / 4 Trey Murphy III: 30 / 15 / 4 / 3 #NBATwitter #NBA #NetsWorld #Pelicans
Image from Machine Learning Ball (@MLBallForecast)
65
BKN at NO: Confirmed, 6:40 PM CT Win probability: NO 55%, BKN 45% Margin: NO by 2 80% range: BKN by 14 to NO by 19 Total: 230 80% range: 207 to 253 Players to watch (MIN / PTS / REB / AST) Michael Porter Jr.: 32 / 22 / 6 / 2 Julius Randle: 32 / 20 / 5 / 4 Zion Williamson: 31 / 20 / 5 / 3 Updated with confirmed starting lineups; the morning forecast stays on the site. #NBATwitter #NBA #NetsWorld #Pelicans
Image from Machine Learning Ball (@MLBallForecast)
75
LAL at MIL, 7:00 PM CT Win probability: MIL 16%, LAL 84% Margin: LAL by 14 80% range: LAL by 31 to MIL by 3 Total: 231 80% range: 208 to 254 Players to watch (MIN / PTS / REB / AST) Luka Doncic: 34 / 32 / 6 / 6 Austin Reaves: 33 / 21 / 3 / 4 Tyler Herro: 28 / 17 / 4 / 3 #NBATwitter #NBA #LakeShow #FearTheDeer
Image from Machine Learning Ball (@MLBallForecast)
149
LAL at MIL: Confirmed, 6:36 PM CT Win probability: MIL 13%, LAL 87% Margin: LAL by 15 80% range: LAL by 31 to MIL by 2 Total: 230 80% range: 207 to 253 Players to watch (MIN / PTS / REB / AST) Luka Doncic: 34 / 33 / 6 / 6 Austin Reaves: 34 / 21 / 4 / 4 Tyler Herro: 29 / 18 / 4 / 3 Updated with confirmed starting lineups; the morning forecast stays on the site. #NBATwitter #NBA #LakeShow #FearTheDeer
Image from Machine Learning Ball (@MLBallForecast)
121
DEN at MEM, 7:00 PM CT Win probability: MEM 41%, DEN 59% Margin: DEN by 3 80% range: DEN by 20 to MEM by 13 Total: 231 80% range: 208 to 255 Players to watch (MIN / PTS / REB / AST) Nikola Jokic: 35 / 24 / 11 / 9 Jamal Murray: 35 / 24 / 4 / 6 Christian Braun: 32 / 11 / 4 / 2 #NBATwitter #NBA #MileHighBasketball #GrindCity
Image from Machine Learning Ball (@MLBallForecast)
78
DEN at MEM: Confirmed, 6:32 PM CT Win probability: MEM 41%, DEN 59% Margin: DEN by 4 80% range: DEN by 20 to MEM by 13 Total: 232 80% range: 208 to 255 Players to watch (MIN / PTS / REB / AST) Nikola Jokic: 34 / 24 / 11 / 8 Jamal Murray: 34 / 23 / 4 / 6 Christian Braun: 31 / 11 / 4 / 2 Updated with confirmed starting lineups; the morning forecast stays on the site. #NBATwitter #NBA #MileHighBasketball #GrindCity
Image from Machine Learning Ball (@MLBallForecast)
53
SA at CHI, 7:00 PM CT Win probability: CHI 13%, SA 87% Margin: SA by 15 80% range: SA by 32 to CHI by 2 Total: 237 80% range: 214 to 261 Players to watch (MIN / PTS / REB / AST) Caleb Wilson: 30 / 16 / 5 / 2 Josh Giddey: 28 / 12 / 5 / 7 De'Aaron Fox: 27 / 14 / 3 / 5 #NBATwitter #NBA #PorVida #SeeRed
Image from Machine Learning Ball (@MLBallForecast)
52
SA at CHI: Confirmed, 6:34 PM CT Win probability: CHI 12%, SA 88% Margin: SA by 16 80% range: SA by 33 to CHI by 1 Total: 235 80% range: 212 to 258 Players to watch (MIN / PTS / REB / AST) Caleb Wilson: 31 / 17 / 6 / 2 Josh Giddey: 30 / 13 / 6 / 8 De'Aaron Fox: 30 / 15 / 3 / 5 Updated with confirmed starting lineups; the morning forecast stays on the site. #NBATwitter #NBA #PorVida #SeeRed
Image from Machine Learning Ball (@MLBallForecast)
67
CLE at NY, 6:00 PM CT Win probability: NY 61%, CLE 39% Margin: NY by 4 80% range: CLE by 13 to NY by 21 Total: 223 80% range: 199 to 246 Players to watch (MIN / PTS / REB / AST) Jalen Brunson: 32 / 23 / 3 / 5 OG Anunoby: 31 / 15 / 4 / 2 James Harden: 30 / 15 / 3 / 4 #NBATwitter #NBA #LetEmKnow #NewYorkForever
Image from Machine Learning Ball (@MLBallForecast)
1
59
CLE at NY: Confirmed, 5:38 PM CT Win probability: NY 63%, CLE 37% Margin: NY by 5 80% range: CLE by 12 to NY by 22 Total: 223 80% range: 199 to 246 Players to watch (MIN / PTS / REB / AST) Jalen Brunson: 31 / 23 / 3 / 5 James Harden: 31 / 15 / 4 / 5 OG Anunoby: 31 / 15 / 4 / 2 Updated with confirmed starting lineups; the morning forecast stays on the site. #NBATwitter #NBA #LetEmKnow #NewYorkForever
Image from Machine Learning Ball (@MLBallForecast)
85
BOS at TOR, 6:00 PM CT Win probability: TOR 31%, BOS 69% Margin: BOS by 7 80% range: BOS by 24 to TOR by 10 Total: 217 80% range: 193 to 240 Players to watch (MIN / PTS / REB / AST) Scottie Barnes: 31 / 17 / 6 / 5 Jayson Tatum: 31 / 22 / 6 / 5 RJ Barrett: 30 / 17 / 4 / 3 #NBATwitter #NBA #DifferentHere #WeTheNorth
Image from Machine Learning Ball (@MLBallForecast)
25
BOS at TOR: Confirmed, 5:34 PM CT Win probability: TOR 35%, BOS 65% Margin: BOS by 5 80% range: BOS by 22 to TOR by 11 Total: 216 80% range: 193 to 239 Players to watch (MIN / PTS / REB / AST) Immanuel Quickley: 35 / 17 / 4 / 5 RJ Barrett: 33 / 20 / 5 / 3 Jayson Tatum: 32 / 24 / 6 / 5 Updated with confirmed starting lineups; the morning forecast stays on the site. #NBATwitter #NBA #DifferentHere #WeTheNorth
Image from Machine Learning Ball (@MLBallForecast)
104
WSH at IND, 6:00 PM CT Win probability: IND 56%, WSH 44% Margin: IND by 3 80% range: WSH by 14 to IND by 19 Total: 228 80% range: 205 to 252 Players to watch (MIN / PTS / REB / AST) Trae Young: 32 / 17 / 3 / 7 Anthony Davis: 30 / 19 / 8 / 2 Pascal Siakam: 30 / 20 / 5 / 3 #NBATwitter #NBA #ForTheDistrict #YesCers
Image from Machine Learning Ball (@MLBallForecast)
2
114
WSH at IND: Confirmed, 5:34 PM CT Win probability: IND 60%, WSH 40% Margin: IND by 4 80% range: WSH by 13 to IND by 20 Total: 229 80% range: 206 to 252 Players to watch (MIN / PTS / REB / AST) Trae Young: 33 / 17 / 3 / 7 Pascal Siakam: 32 / 21 / 6 / 3 Andrew Nembhard: 31 / 16 / 2 / 7 Updated with confirmed starting lineups; the morning forecast stays on the site. #NBATwitter #NBA #ForTheDistrict #YesCers
Image from Machine Learning Ball (@MLBallForecast)
126
OKC at POR, 9:00 PM CT Win probability: POR 48%, OKC 52% Margin: OKC by 1 80% range: OKC by 18 to POR by 16 Total: 228 80% range: 204 to 251 Players to watch (MIN / PTS / REB / AST) Deni Avdija: 30 / 19 / 5 / 5 Shai Gilgeous-Alexander: 27 / 23 / 3 / 4 Toumani Camara: 27 / 10 / 3 / 1 #NBATwitter #NBA #ThunderUp #RipCity
Image from Machine Learning Ball (@MLBallForecast)
1
106
Load more