The free MLB data sources I use to investigate whether a historical betting angle still fits today's pitcher, lineup, bullpen, park, weather, and price.
This is where I check whether surface results are supported by the actual contact and pitch data. A historical angle is useless if today's pitcher or lineup no longer fits the mechanism. Limitation: Recent samples are noisy. Expected metrics are est...
Baseball Savant
FoxEdge uses FanGraphs to challenge ERA, batting average, and other surface numbers before trusting a matchup. It is especially useful for starter skill, bullpen quality, and handedness splits. Limitation: Projection systems disagree, and some tools...
FanGraphs
This is the fastest way to verify historical game and player context before building a more technical query. It is useful for checking whether a narrative survives basic factual review. Limitation: Flexible splits make cherry-picking dangerously eas...
Baseball Reference
FoxEdge uses the official game page to verify the inputs that can invalidate an analysis immediately, especially pitcher changes and lineup news. Limitation: Official data is reliable, but the interface is not built for comparative research.
MLB Gameday
A model using the wrong starter is not slightly wrong. It is garbage with decimals. This is the first verification step before any pitcher-dependent research. Limitation: Probables can still change, and openers complicate the listing.
MLB Probable Pitchers
Season-long bullpen quality means little if the best relievers are unavailable. RosterResource helps identify who is healthy, active, and expected to handle leverage innings. Limitation: Roles change quickly and still need game-day confirmation.
RosterResource
I use MLB's official starting-lineup page to confirm active hitters, batting order, handedness, and late changes before finalizing sides, totals, or props. Limitation: Lineups remain subject to late scratches and should be rechecked near first pitch...
MLB Starting Lineups
Retrosheet supports original research that cannot be produced from a prebuilt leaderboard. It is one of the foundations for reproducible historical baseball analysis. Limitation: Raw files require technical processing and careful schema work.
Retrosheet
This helps turn manual website research into repeatable pipelines. FoxEdge uses code-driven research because thousands of games cannot be audited by scrolling harder. Limitation: Scraped sources can change, so caching and validation are mandatory.
pybaseball
The free data sources explain the inputs. FoxEdge combines those inputs with historical evidence, current price, and market context to decide whether a game is playable or a pass. Limitation: Historical findings do not guarantee future results, and ...