Excel LADZ - Weekly Update: Tuesday 18 July

G’day lads, I’m excited to release the NBA In-Season Tournament video on Thursday, 20 July. As the model has already been posted to Patreon, I’ll run through the reasoning behind it.

The first step was deriving an attacking and defensive strength of each team. For this, I used the 2022-23 NBA Season stats. To work out a team’s attacking strength, it was simply the number of points scored per game by the team (throughout the 2022-23 season), divided by the league average points scored per game. The same method was used to find the defensive strength.

However, I have recently added a slight modification to the model. To adjust for the uncertainty in the season ahead (e.g. the Spurs may have a significant increase in strength), I have reverted every team ¼ of the way to the mean. For example, a team with an ATT Rating of 1.04 in the 2022-23 NBA Season, will have a corresponding ATT Rating of 1.03 in the model. This has a noticeable impact on the final Win Probabilities; weaker teams have a higher chance, stringer teams have a lower chance.

The second step of the model was to use the ATT & DEF Ratings to find Expected Points for each matchup in the groups. This involved Team 1’s ATT * Team 2’s DEF * League Average Score. However, to simulate this Expected Points figure, I’m required to simulate free throws, 2’s and 3’s separately.

Take the Grizzlies in the example below. Based on their Expected Points, it is expected that they will have 113.0 shots in the match. 50.0% of these will be 2 Pointers, of which they are expected to make 54.8% of. Not only am I simulating the number of 2 pointers shot in the match, but also how many are made. Multiplying this figure by 2, and I can see how many points shots from 2-Pointers have contributed towards the Grizzlies’ final score. Doing this for FT’s and 3’s, and their final regular time points of 122 will be achieved (for this single simulation).

After simulating these matches, the final standings of the group stage can be determined. For here, it’s just a simple knockout bracket that decides who makes, and wins, the final. This whole process was simulated 10,000 times to give you the win probabilities uploaded to the Patreon. However, the model attached only does 1,000 simulations, to make sure your computer doesn’t crash when opening the file.

Short Term

• NBA In-Season Tournament Model YouTube Release
• AFL Model uploaded to Patreon
• NRL Model uploaded to Patreon

Medium Term

• EPL Model
  • With the English Premier League Season starting in early August, I am planning to release a string of models. These include predicting the whole season, halftime markets, corner markets, and much more.

If you have any questions lads, leave a comment here or on Discord! I look forward to hearing about your ideas and/or questions!