Most Profitable Algorithm

The chart represents the algorithm's performance if you were to invest 1 unit in each prediction it makes throughout the season.The red line represents the threshold that separates positive from negative performance.

Currently, the algorithm is profitable for 1 out of the 4 seasons in which training and testing have been conducted, with differents algorithms.

Prebet

This repository contains all the components necessary to extract, manage, and analyze NBA game data, implementing machine learning algorithms for wins prediction, and evaluate profitability. Youtube channel

Below is the organization of the repository:

Repository Structure

1. Data Extraction

Folder: 1

Scripts and resources for extracting NBA game data from various sources:

• Web scraping from Basketball reference. 1.R and 1_odds.R
• Raw datasets in RData format.
• Documentation of data sources and extraction workflows.

---

2. Database Cleaning

Folder: 2

Processes for cleaning and preparing the raw data for analysis:

• Handling missing values, outliers, and normalization.
• Merging and restructuring datasets.
• Reports summarizing cleaning and validation steps.
• Creation of new variables ( rankings, victories in a row,...)

---

3. Machine Learning Prediction Algorithms

Folder: 3

Implementation of machine learning models for predicting game outcomes:

• Training and evaluation scripts for logistic regression, random forests, gradient boosting, etc.
• Hyperparameter tuning and cross-validation workflows.
• Model evaluation reports with plots visualitation
• ...

---

4. Profitability and Portfolio Analysis

Folder: 4

Analysis of profitability and portfolio management:

• Profit calculation scripts based on predictions.
• Portfolio simulation and management workflows.
• Probability modeling and visualization of profit distributions.
• ...

---

5. Automatic Betting Algorithm

Folder: 5

Development of automated betting algorithms:

• Early-stage scripts for automated betting decisions.
• Integration with betting platforms (if applicable).
• Debugging logs and testing files.
• ...

---

6. Marketing and Advertising

Folder: 6

Resources for promoting the project:

• Marketing strategies and analysis of target audiences.
• Branding materials (e.g., logos, banners).
• Drafts of social media posts and blog content.
• Tracking files for engagement and visibility metrics.
• YouTube channel
• Creating a shiny app for interacting with the data
• ...

---

How to Use This Repository

1. Clone the repository:

   git clone https://github.com/javicarela/prebet.git
   

We are dedicated to continuously improving this project. Features currently in development are designed to enhance functionality and user experience. Contributions and feedback are welcome.