Fight Genie

Advanced UFC Fight Prediction & Analysis System

https://fightgenie.ai/

https://rudycorradetti.com/2024/12/04/fight-genie-ai-nodejs-discord-bot-ufc-predictions/

Fight Genie is an advanced UFC fight prediction system utilizing dual AI models (GPT and Claude Sonnet), real-time odds integration, and comprehensive statistical analysis to provide accurate fight predictions and betting insights.

Table of Contents

• Overview
• System Architecture
• Core Features
• Technical Implementation
• Data Processing
• Prediction System
• Security & Privacy
• Deployment

Overview

Key Features

• Real-time fight predictions using dual AI models
• Live betting odds integration
• Comprehensive fighter statistics
• Privacy-focused data handling
• Automated event management
• Detailed statistical analysis

Access Model

Fight Genie is now 100% FREE to use! All features are available to everyone.

If you find the bot valuable, please consider supporting its development and server costs using the $donate command.

System Architecture

Database Structure (SQLite)

Currently implemented as local SQLite with planned Azure migration path:

-- Core Tables
events
  - event_id, Event, Date, City, Country
  - event_link, is_completed, event_time
  - prelims_time, main_card_time

fighters
  - Name, Height, Weight, Reach, Stance
  - SLPM, SApM, StrAcc, StrDef
  - TDAvg, TDAcc, TDDef, SubAvg
  - last_updated

stored_predictions
  - prediction_id, event_id, card_type
  - model_used, prediction_data, created_at

-- server_subscriptions (REMOVED - No longer used)
-- payment_logs (REMOVED - No longer used)
-- solana_payments (REMOVED - No longer used)

Data Pipeline

Event Management System

Event Detection Flow:
1. Check current date against event database
2. Scrape UFCStats.com for verification
3. Update event status and fight card
4. Trigger predictions if needed

Fighter Statistics Pipeline

Data Collection Process:
1. UFCStats.com scraping
2. Data normalization
3. Statistical validation
4. Database storage
5. Cache management

Technical Implementation

Web Scraping System

UFCStats.com data collection using Axios and Cheerio:

Scraping Scenarios:
1. Event Detection
   - Upcoming events page
   - Event details and timing
   - Fight card composition

2. Fighter Statistics
   - Individual fighter pages
   - Career statistics
   - Fight history
   - Performance metrics

3. Fight Results
   - Completed event verification
   - Result confirmation
   - Record updates

Prediction System

Confidence Calculation (100 points)

Score Components:
1. Base Score (30 points)
   - UFC win rate (15)
   - Opposition quality (15)

2. Style Score (25 points)
   - Technical advantages (15)
   - Physical advantages (10)

3. Form Score (25 points)
   - Recent performance (15)
   - Activity/preparation (10)

4. Historical Score (20 points)
   - Career consistency (10)
   - Championship experience (10)

Edge Calculation

Edge Formula:
1. Convert odds to probability
   - Positive odds: 100/(odds + 100)
   - Negative odds: |odds|/(|odds| + 100)

2. Calculate edge
   - Edge = model_confidence - market_probability

3. Determine value rating
   ⭐⭐⭐⭐⭐: 20%+ edge, 70%+ confidence
   ⭐⭐⭐⭐: 15%+ edge, 65%+ confidence
   ⭐⭐⭐: 10%+ edge, 60%+ confidence

Payment Integration (REMOVED)

Payment processing (PayPal, Solana, Stripe) has been removed as the bot is now free.

Security & Privacy

Data Protection

Stored Data (Minimal):
- Discord Server ID
- Discord Admin ID (for specific admin commands)

Not Stored:
- Personal Information
- Payment Details
- User Messages
- IP Addresses

Prediction Accuracy

Tracking Metrics:
1. Winner prediction accuracy
2. Method prediction accuracy
3. Round prediction accuracy
4. Confidence correlation
5. Edge validation

Documentation Notes

Command Reference

User Commands:
$upcoming - Show current event
$predict - Generate predictions
$checkstats - Fighter statistics
$sub - Check bot status
$donate - Support Fight Genie's development

Admin Commands:
(Admin commands related to payments/promos removed)
- $advance - Advance to the next event (Admin only)
- $forceupdate - Force update current event data (Admin only)
- $syncpredictions - Sync predictions (Admin only)

Manual Tweet Triggering

You can manually trigger specific tweet types using the TweetAutomation.js script directly from the command line. This is useful if a scheduled tweet fails or needs to be posted outside the schedule.

Usage:

node src/utils/TweetAutomation.js <tweet_type>

Replace <tweet_type> with one of the following:

• Fight Analysis Thread:

  node src/utils/TweetAutomation.js fight_analysis

• Value Pick Tweet:

  node src/utils/TweetAutomation.js value_pick

• Model Competition Thread:

  node src/utils/TweetAutomation.js model_competition

• Weekly State Thread:

  node src/utils/TweetAutomation.js weekly_state

• UFC History/Promo Tweet:

  node src/utils/TweetAutomation.js ufc_history_promo

Note: This script respects the TWEET_TEST_MODE environment variable set in your .env file. If set to true, tweets will be logged to the console and tweet-logs.txt instead of being posted to Twitter.

Support

For technical support or inquiries:

• Email: rudycorradetti4@gmail.com

---

Technical Specifications

System Requirements

• Node.js 16+
• SQLite3
• Discord.js v14+
• Minimum 2GB RAM
• 10GB Storage