Odds on Deck

Odds on Deck compares betting lines from more than ten sportsbooks across MLB, NFL, and NHL, converts them into vig-free market probabilities, and tracks how flagged props resolve against official results — so research is based on measured outcomes rather than marketing claims.

Sole founder and developer. I own the data model, ingestion, caching strategy, API budget, front-end, and deployment.

• —High-volume odds data under strict third-party API cost limits
• —Games and lines change in real time, so data freshness matters
• —Single-maintainer project — everything has to stay simple enough to run alone

A unified data model normalizes MLB, NFL, and NHL odds into one shape. A game-time-aware cache batches requests and only refetches when lines are likely to have moved, and scheduled jobs pull official results to grade tracked props automatically.

• +Batched fetches plus a game-time-aware prop cache cut third-party API costs by roughly 90%
• +One normalized schema across three leagues instead of per-league special cases
• +Automated result validation against official box scores so tracked props grade themselves
• +Vig removed from book lines to show a fair-market probability, not a marketing number
• +Kept the stack JavaScript-first (Next.js / React) for fast solo iteration

• +Approximately 90% reduction in third-party API usage costs through batched requests and game-time-aware caching
• +Unified MLB / NFL / NHL coverage from 10+ sportsbooks
• +Automated, auditable result validation