NBA Parlay Bets

Bet Types

Bet type	Variables	Model Type	Depiction
"Will Player X Make More Than Y Points in This Game?"	Player Context 1. Average points per game (season and last 5-10 games) 2. Field goal % (FG%), free throw %, and three-point % 3. Play Time 4. Injury status (game-time decision, minor injury, etc.) Game Context 1. Opponent’s defensive rating 2. Home vs. away game 3. Back-to-back games or rest days	Regression Problem - A continuous value	Linear Regression Model
"Will Team M Beat Team N?"	Team Stats 1. Team’s offensive and defensive ratings 2. Field goal %, three-point %, free throw % 3. Average turnovers per game 4. Points per possession (PPP) 5. Rebounding rate (offensive and defensive) Game Context 1. Home court advantage 2. Back-to-back games or rest days 3. Head-to-head record Situational Stats 1. Injury to key players	Classification Problem - "yes"/"no" output	Logistic Regression Model
"Will Player X Have More Than K+ Three-Pointers?"	Player Context 1. Average three-pointers made per game 2. Three-point attempt rate 3. Minutes played in recent games Game Context 1. Opponent’s 3P% allowed 2. Injury of key-players from same team (lead to more shooting opportunities)	Classification Problem - "yes"/"no" output	Logistic Regression Model
"Will Player X Have More Than K+ Rebounds?"	Player Context 1. Average rebounds per game 2. Rebounding Rate (offense & defense) 3. Height and position (e.g., centers and forwards generally grab more rebounds) 4. Minutes played per game 5. Presence of other rebounders on the team Game Context 1. Opponent’s rebounding rate 2. Opponent’s shooting percentage (more misses = more rebounding opportunities) 3. Home vs. away game 4. Lineup changes or injuries (e.g., if another key rebounder is injured, this player might get more opportunities)	Classification Problem - "yes"/"no" output	Logistic Regression Model