Categories

Thursday, June 1, 2023

NBA Finals 2023 Bracket Math

Once the playoff bracket was set, and Miami barely got in to the field as the 8 seed, I gave them just a 0.7% chance to make the Finals, and only 0.1% to win the title. Out of 10,000 simulations, that's roughly 10 wins. That's it.

Their chances are a lot higher now! And their impressive abnormal(?) performance in the postseason necessitates a few assumptions to be changed in my play-by-play simulation of the Finals. I mean, I had them as the worst rated team in the field - and they did have a negative point differential in the regular season!

Per current injury reporting, I'm assuming Tyler Herro comes back for Game 3, while other injuries (like Victor Oladipo) eliminate those players from the series. I'm taking in to account Denver's significant home court advantage, which I've analyzed before, and so has Zach Lowe and others at ESPN.

The final component I had to decide on - do I use regular season data to underpin the model, or each team's postseason performance?

I didn't exactly have Denver as a juggernaut before the playoffs - they were the 4th strongest team in the West, behind Phoenix, Golden State, and Memphis. And when I use regular season only, I get a competitive series, with Denver as a slight favorite:

# GamesDEN in...MIA in...
48.2%4.9%
515.3%9.5%
617.6%13.3%
717.3%14.0%
58.3%41.7%

But the Heat have been on a run for the ages, and their "new normal" is maybe this elevated postseason play. So I reran everything off of playoff statistics only... only to see Denver strengthen:

# GamesDEN in...MIA in...
47.8%4.3%
517.1%8.5%
616.0%14.6%
718.7%12.9%
59.6%40.4%

Denver's chances as the favorite actually increase. My results aren't as aggressive as the betting markets (80%) and FiveThirtyEight, which has Denver as a 74% favorite

Ultimately I project Denver in 7 (I was wrong last year, when I had Boston in 6).

While Jimmy Butler has been otherworldly, Nikola Jokic has been even more dominant. Here is the projected average box score, using the playoff data as the input:

NamePosTeamPointsReboundsAssistsTurnoversStealsBlocks
Nikola JokicCDEN34.112.612.03.61.50.4
Jimmy ButlerPFMIA25.55.75.51.51.90.3
Tyler HerroSGMIA20.95.65.52.31.00.2
Bam AdebayoCMIA17.78.84.42.41.20.4
Jamal MurrayPGDEN17.43.14.51.31.20.1
Michael Porter Jr.SFDEN15.17.42.10.50.90.3
Aaron GordonPFDEN12.74.22.40.90.50.3
Bruce BrownSFDEN11.33.12.20.81.10.1
Kentavious Caldwell-PopeSGDEN9.52.41.50.61.20.2
Duncan RobinsonSFMIA8.11.52.10.80.50.0
Caleb MartinSFMIA7.83.11.20.60.60.1
Kevin LovePFMIA7.85.51.80.70.50.2
Max StrusSFMIA7.42.30.90.20.40.1
Gabe VincentPGMIA6.91.02.40.80.60.1
Kyle LowryPGMIA5.61.93.01.00.70.3
Ish SmithPGDEN5.42.04.41.80.40.1
Jeff GreenPFDEN4.81.90.80.50.30.2
Cody ZellerCMIA2.82.30.40.60.20.1
Christian BraunSGDEN2.81.70.40.50.50.1
DeAndre JordanCDEN0.50.30.10.20.00.0

Jokic averages a triple double in the Finals 52.4% of the time. That's insane. I have him dropping 50+ 2.7% of the time, compared to 0.2% for Butler, and 0.01% for Jamal Murray (no other player recorded 50+ once in any simulation).

Compared to regular season data as the input:

NamePosTeamPointsReboundsAssistsTurnoversStealsBlocks
Jimmy ButlerPFMIA26.05.56.21.42.30.2
Nikola JokicCDEN25.610.411.73.41.40.3
Bam AdebayoCMIA25.510.34.52.71.80.5
Jamal MurrayPGDEN18.13.26.41.81.00.1
Aaron GordonPFDEN17.35.53.41.30.90.4
Tyler HerroSGMIA16.24.34.31.80.80.1
Michael Porter Jr.SFDEN14.74.21.10.90.60.2
Max StrusSFMIA11.83.42.70.90.80.1
Bruce BrownSFDEN10.93.23.41.31.10.3
Kentavious Caldwell-PopeSGDEN10.22.22.51.01.40.2
Caleb MartinSFMIA9.24.21.91.01.20.2
Kyle LowryPGMIA7.92.64.11.20.90.1
Gabe VincentPGMIA7.91.72.51.01.00.0
Jeff GreenPFDEN7.82.11.20.70.30.1
Christian BraunSGDEN5.32.20.90.50.60.1
Duncan RobinsonSFMIA4.81.20.90.40.30.0
DeAndre JordanCDEN3.52.80.60.80.30.2
Kevin LovePFMIA3.32.00.90.40.30.0
Cody ZellerCMIA2.01.00.20.20.10.0
Ish SmithPGDEN1.90.71.60.70.20.0

Bam regresses quite a bit using playoffs only, Butler and Murray are both actually quite consistent, but Jokic rocketing up the leaderboard is the difference.