Thursday, June 6, 2024

NBA Finals 2024 Bracket Math

Going in to the playoffs, I only had 4 teams with > 5% chance of winning the title - Boston led the way with almost a 50/50 chance to raise another banner, whereas Dallas was barely on the radar, with the 12th best odds at 0.38%. Last year, Denver was barely above this 5% threshold, and ultimately steamrolled a Miami team that I gave a 0.1% chance of winning the championship at the start of last year's playoffsDenver then had about a 60% edge last Finals on the back of an incredible playoff run by Nikola Jokic.

This season it's Luka Doncic's turn in the postseason, so I've based this season's Finals simulations off of playoff statistics only. This choice could seem to penalize Boston, as they have had much less impressive performances throughout these playoffs, compared to Dallas. Even so, they've only lost 2 games the entire postseason.

In Boston's favor, I am including Kristaps Porzingis at full-strength, and assuming his regular season workload, as he is not listed on the injury report and is expected to play in all games.

A lot has been made about how Dallas would be the 3rd biggest Finals upset in the modern era how Dallas would be the 3rd biggest Finals upset in the modern era - their odds are pegged at +185, or ~35%. 

And yet I actually have those market odds as overrating their chances:

# GamesBOS in...DAL in...Total
412.0%2.7%14.8%
521.7%5.8%27.5%
620.0%9.6%29.6%
718.1%10.1%28.2%
71.9%28.1%100.0%

Ultimately I project Boston in 5. I have them as 7 point favorites (63% win probability) at home, and 3.5 point favorites (58% win probability) on the road. This adds up to a very large series probability.

A big question in this series is whether the top tier talent of Dallas can overwhelm Boston's depth. Dallas might have the 1st and 2nd best player, but Boston is far deeper at 3, 4, 5, 6, etc.

Luka may have the best projected stat line, but Tatum and Brown are consistently really good, and that, plus Boston's depth, results in Boston as a very strong favorite.

NamePosTeamPointsReboundsAssistsTurnoversStealsBlocks
Luka DoncicPGDAL30.28.49.23.61.70.2
Jayson TatumPFBOS29.99.86.52.21.50.3
Jaylen BrownSFBOS24.65.83.22.41.40.3
Kyrie IrvingSGDAL24.63.86.62.51.40.2
Derrick WhiteSGBOS18.64.35.70.71.30.6
P.J. WashingtonPFDAL14.96.61.91.41.10.3
Kristaps PorzingisCBOS11.24.21.60.81.20.6
Daniel GaffordCDAL10.55.50.91.00.30.7
Jrue HolidayPGBOS10.44.74.41.31.40.3
Dereck Lively IICDAL10.16.41.91.00.30.5
Derrick Jones Jr.SFDAL9.63.21.60.90.50.4
Payton PritchardPGBOS7.42.22.60.70.30.0
Al HorfordCBOS7.25.31.70.50.70.3
Sam HauserSFBOS5.72.30.80.30.50.1
Josh GreenSFDAL5.42.41.40.40.80.1
Tim Hardaway Jr.SFDAL3.11.10.30.10.30.0
Luke KornetCBOS2.51.80.30.30.00.1
Maxi KleberCDAL2.31.20.80.30.10.1
Dante ExumPGDAL2.10.70.60.10.10.0
Xavier Tillman Sr.CBOS0.50.70.20.20.20.0

Tuesday, May 14, 2024

Clark vs DeJean, A Fairly Competitive Game of One-on-One

A follow-up to my post, "Lavar vs Jordan, An Incredibly Lopsided Game of One-on-One". Cooper DeJean, professional football player and former Iowa Hawkeye, made headlines when he declared he could win a game of one-on-one against Caitlin Clark, professional basketball player and former Iowa Hawkeye.

Setting aside Austin Rivers's commentary on the crossover appeal of basketball/football players, Clark holds the all-time points record in both men's and women's college basketball. DeJean was a very good high school basketball player though, scoring more points in the state of Iowa than Carolina great Harrison Barnes, and recording more steals in the state of Iowa than Carolina great Marcus Paige.

So I set out to simulate DeJean's claim that if he lost, it would be by one or two. Using Clark's college senior year statistics at Iowa, compared to DeJean's high school senior year statistics at OABCIG, I simulated a game to 21 by ones and twos.

And he's right! I have Caitlin Clark winning 53.5% of the time by an average score of 21 to 20 - a very evenly matched game.

Notwithstanding the relative quality of competition, DeJean's high school statistics are fairly comparable to Clark's last year at Iowa - but where Clark wins out is her 3-Point Attempt Rate. Not her 3 Point percentage - Clark shot 37.8% her (college) senior year, compared to DeJean's 39.5%. But 60% of Clark's field goal attempts were threes - a ridiculously high volume compared to DeJean's 21.5%. 

Especially in a game of ones and twos, Clark's range and 3-point volume would give her the edge.

Saturday, April 20, 2024

NBA Playoffs 2024 Bracket Math (Post Play-In)

The streak continues - no regular season 7 seed has failed to make the playoffs in the play-in era. However, this year's iteration came at great cost to the respective teams, both losing arguably their best player - Zion Williamson for New Orleans and Jimmy Butler for Miami appear to be looking at lengthy absences. 

But the playoffs are set, and there is one very clear favorite - Boston, with their dominant regular season, is almost 50/50 to win it all - more than the next 3 teams combined (Denver, OKC, and Minnesota).


Denver feels too low, even though they coasted through much of the regular season. And I think if you talk to just about any Boston fan, even with their historic season to date, they would agree Boston feels too high. Miami has the chance to do the funniest thing ever and beat Boston without Butler (10% chance).

Home court also matters! Note I have Philly rated higher than New York, but the Knicks are still a slight favorite in the first round.

The above reflects a lot less parity than last year, with only 4 teams with > 5% chance of winning the Finals (Boston, Denver, OKC, and Minnesota). Last year I had 8 teams meet this threshold - the lowest of which was Denver, who looked pretty unbeatable on the way to the championship.

Any other result other than these 4 teams would be surprising - combined, they raise the banner 88% of the time.

Tuesday, April 16, 2024

NBA Play-In 2024 Bracket Math

As always, winning 2 games in a row is really hard - just like it was last year, when both the 7 and 8 seeds ultimately made the playoffs. No 7 seed has ever ultimately missed the playoffs since the play-in began in 2021.

This year is no different - while the 7 and 8 in the East are clearly superior teams, the second best play-in team in the West is the 10 seed Golden State Warriors.

But they have to win twice in a row, on the road - and thus have the worst odds of making the playoffs out of the West.


This also speaks to how perilous a strategy it would be for the Lakers to intentionally lose the first play-in game to avoid Denver in the first round (something that has been suggested).

Even setting aside that I have the Nuggets rated lower than the Thunder, the Lakers would only have about a 55% chance to beat either the Kings or Warriors in a single game. That leaves 45% where they completely miss the playoffs - so the difference in beating OKC over Denver would have to be massive to maximize "likelihood to win the title".

Sunday, March 31, 2024

Final Four per the MDS Model (2024)

While Houston went bust, one of the other two top contenders are overwhelmingly our likely champion (85% combined). But due to the much easier path (on paper) for Purdue, they substantially outpace Connecticut to win it all - even though Connecticut looks pretty unbeatable and has led by 30+ in every game.

SeedTeamChampionshipChampion
Final Four
1Connecticut67.60%38.86%
4Alabama32.40%13.12%
1Purdue87.11%46.07%
11NC State12.89%1.95%

I did pick the champion last year correctly at this stage (Connecticut). In the likely outcome that it's Connecticut/Purdue in the final (60%), Connecticut would be about a 1 point favorite over Purdue.

SeedMatchupsTitle Game
1Connecticut52.89%
1Purdue47.11%
1Connecticut88.57%
11NC State11.43%
4Alabama35.04%
1Purdue64.96%
4Alabama77.25%
11NC State22.75%