MLB Rivalry Tracker

MLB Rivalry Tracker provides a daily comparison of three Major League Baseball teams—the Chicago Cubs, New York Mets, and Los Angeles Dodgers—by automatically pulling and processing the latest team and player statistics. Behind the scenes, it applies robust logic to ensure fair, meaningful comparisons and generates both individual leaderboards and overall team rankings.

Daily Stats from the MLB API

Every day, a backend job fetches both team and individual player stats from MLB's public stats API. By running after each day's games, the app guarantees that all leaderboards and rankings reflect the most up-to-date performances.

Fair Player Comparisons with Thresholds

To avoid skewed leaderboards caused by small sample sizes, the app enforces minimum-play thresholds before including any player in a given category:

Batters: Must have at least 3.1 plate appearances × the number of games their team has played to date.
Starting Pitchers: Must have thrown at least 1 inning × the number of games their team has played to date.
Relief Pitchers: Must have thrown at least 2 innings per week (week 1 begins March 27, 2025).

The 3.1 PA/game rate matches MLB's season-long qualification (502 PA over 162 games) and scales by each team's actual games played so far, using `plateAppearances` and `gamesPlayed` from DynamoDB.

Shohei Ohtani Exception

Because Shohei Ohtani is a two-way player, we handle him specially: only his hitting stats appear in the batter leaderboards—even though he also pitches, we exclude his pitching data to avoid double-counting him. All other players follow the normal role assignment logic.

Detailed Logic for Determining Stats

Scan & Deduplicate: Fetch all records from `PLAYER_STATS_TABLE` and keep only each player's most recent entry by date.
Threshold Calculation:
- Season anchor: March 27, 2025 marks week 1.
- Week number: floor((today – anchor)/7) + 1.
- Pitcher thresholds: 1 IP × gamesPlayed (per team) for starters; 2 IP/week for relievers.
- Batter threshold: 3.1 PA × gamesPlayed (from `TEAM_STATS_TABLE`).
Grouping Players: Group by team and role, applying the Ohtani exception; split pitchers into starters vs. relievers by gamesStarted/gamesPitched ≥ 0.5.
Selecting Top Performers: For each stat, filter out players below the threshold, parse the stat, then pick the best value (higher is better for positive stats; lower is better for ERA/WHIP). Ties share the same rank, and batting average is formatted to three decimals (e.g. .320).
Building Leaderboards: Assemble separate tables for batters, starters, and relievers, showing each team's top value and the league-wide leader.

Aggregating Team Scores

Each team's position (1st, 2nd, or 3rd) in every stat category is converted into ranking points (1 = best, 2 = middle, 3 = worst). Ties share the same point value, and the next rank is bumped accordingly (e.g., two teams tie for 1st → both score 1, remaining team scores 3). Summing these points across all categories yields an overall team score.

Automated Daily Updates (AWS Pipeline)

After 3 PM Tokyo time each day, AWS Step Functions orchestrates Lambda functions that fetch MLB API data, compute thresholds (including the Ohtani logic), group and rank players, and write results to DynamoDB. Notifications report success or failure, and the front‑end automatically refreshes to show the latest comparisons.

Loading...