Baseball Savant and FanGraphs for Betting: Extracting Wagering Value from Free Data Tools
Available in US
Two Free Platforms That Professional Bettors Use Daily
When people ask me what tools I use, they expect me to name some expensive subscription service or a proprietary model that costs hundreds per month. The truth is less glamorous. The two platforms that form the backbone of my daily research are Baseball Savant and FanGraphs – both free, both publicly accessible, and both producing the same data that all 30 MLB analytics departments use internally. The edge does not come from having exclusive data. It comes from knowing which data to pull and how to apply it before the line moves.
AI-driven tools have pushed engagement on betting platforms up by as much as 25%, and a significant portion of that engagement is powered by the same Statcast data that Baseball Savant makes available for free. The democratisation of data means the playing field between sharp bettors and casual bettors has narrowed – but only for those willing to invest the time to learn the platforms.
Baseball Savant: Key Reports and Statcast Leaderboards
Baseball Savant is MLB’s official Statcast portal, and it houses the raw data that underpins most modern pitching and hitting analysis. The three reports I check before every betting session are the Statcast Leaderboard, the Expected Stats page, and the Pitch Arsenal analysis.
The Statcast Leaderboard ranks every player by measurable physical outputs: exit velocity, launch angle, sprint speed, spin rate, and dozens of other metrics. For betting, I sort this leaderboard by expected wOBA (xwOBA), which estimates a hitter’s offensive production based on quality of contact rather than actual outcomes. A hitter whose xwOBA exceeds his actual wOBA by a significant margin has been unlucky – balls hit hard are finding gloves. That gap signals positive regression, and his props lines may be set too low based on recent results rather than underlying contact quality.
The Expected Stats page is the regression-hunter’s best friend. It shows expected batting average (xBA), expected slugging (xSLG), and xwOBA for every hitter and pitcher. When a pitcher’s actual ERA is 4.50 but his expected ERA based on contact quality is 3.20, that pitcher is due for improvement – and the market is pricing his starts based on the inflated ERA rather than the underlying batted-ball data. I run this check for every scheduled starter within the first five minutes of my daily research routine.
The Pitch Arsenal analysis shows each pitcher’s repertoire with movement profiles, usage rates, and whiff rates per pitch. This is invaluable for prop betting: a pitcher who has increased his sweeper usage from 15% to 25% and the whiff rate on that pitch is 38% is becoming a better strikeout pitcher, even if his overall K% has not yet caught up. The props market prices based on recent strikeout totals, but the arsenal data tells you where the totals are heading.
FanGraphs: Projections, Splits, and Betting-Ready Exports
Kyle Boddy, founder of Driveline Baseball and a Red Sox special advisor, has advanced pitching analysis through simulation models using machine learning to explore the full range of mechanical outputs. That analytical ethos permeates FanGraphs, which translates raw data into contextualised metrics that are directly applicable to betting decisions.
FanGraphs’ projection systems – particularly Steamer and ZiPS – provide preseason and in-season win probability estimates for every team. I compare these projections to the sportsbook’s implied probability for each game. When FanGraphs projects a team at 54% win probability and the line implies 48%, the discrepancy is worth investigating. The projections are not perfect, but they provide a disciplined baseline that prevents me from over-relying on gut instinct or narrative.
The Splits tool on FanGraphs is where I spend the most time. I can pull a pitcher’s stats against left-handed hitters in night games at home – or any combination of variables. For first-five-innings betting, I isolate each starter’s performance in specific platoon splits and cross-reference it with the opposing lineup’s handedness. The data is exportable to CSV, which means I can feed it directly into my model without manual transcription errors.
The Roster Resource page is another daily stop. It shows each team’s projected lineup, bullpen order, and injured list. Knowing who is in the lineup before the line adjusts is a timing edge that requires no analytical sophistication – just the discipline to check before the rest of the market does.
Building a Pre-Game Research Routine with Both Platforms
My daily routine takes 20 to 25 minutes and follows a fixed sequence. I start with FanGraphs’ probable pitchers page to confirm the day’s starters and check for any last-minute changes. Then I pull each starter’s FIP, xFIP, and recent spin-rate trends from Baseball Savant. Next, I check the opposing lineup’s wOBA splits against the starter’s handedness on FanGraphs. Finally, I compare the expected stats from Savant to the actual stats to flag regression candidates on either side.
The output of this routine is a short list of games where the data diverges from the line. Most days, that list contains two to four games worth a closer look. On some days, the data and the lines align perfectly and there is nothing to bet – which is itself a valuable outcome, because the discipline of passing on marginal spots is what separates profitable bettors from busy ones.
Both platforms update their data daily, which means the edge from yesterday’s research expires overnight. The routine must be repeated every day, without exception, for the full six-month season. That consistency is the price of admission, and it is why most recreational bettors never bother. They want a shortcut; the data tools are the opposite of a shortcut – they are a daily practice that compounds into seasonal profitability.
For a broader understanding of how these data tools feed into a complete sabermetric betting workflow, the advanced stats for betting guide connects the platform-specific techniques to the underlying analytical framework.
Data Tools FAQ
Which Baseball Savant reports are most useful for daily MLB betting research?
The Expected Stats page, the Statcast Leaderboard sorted by xwOBA, and the Pitch Arsenal analysis are the three most useful daily reports. Expected Stats identifies regression candidates by comparing actual performance to batted-ball quality. The Statcast Leaderboard ranks players by objective physical outputs like exit velocity and sprint speed. The Pitch Arsenal analysis reveals changes in a pitcher’s repertoire and whiff rates that have not yet shown up in aggregate stats.
How do FanGraphs projections compare to sportsbook implied probabilities?
FanGraphs projections and sportsbook implied probabilities are based on similar inputs but serve different purposes. The projections estimate true win probability based on team talent, while the sportsbook line includes a margin (juice) and reflects both sharp and public money. Discrepancies between the two are common and worth investigating. When FanGraphs projects a team at 54% or higher and the line implies 50% or less, the gap may represent genuine value – but always cross-reference with pitching matchup and park factor data before committing.
Written by the editors at mlb Betting Statistics.
