Our Methodology

StatScope sources all game data, player statistics, and roster information from MLB's publicly available Stats API. This is the same data feed that powers the official MLB.com application, ensuring accuracy and timeliness. We do not scrape websites or use unofficial data sources.

StatScope's pre-game win probability model estimates each team's chance of winning before first pitch. Unlike simple prediction systems that rely solely on team records, our model incorporates multiple factors that directly influence game outcomes.

Model Inputs

1. Starting Pitcher Quality: We evaluate the probable starters using a weighted combination of ERA, FIP, WHIP, and K/BB ratio. FIP receives the highest weight because it is the most predictive of future performance. Recent starts (last 5 games) are weighted more heavily than season-long averages to capture current form.
2. Team Offensive Strength: Measured by team wRC+ and OPS over the last 15 games. We use a rolling window rather than season-long numbers to better reflect current offensive performance, which can fluctuate significantly due to injuries and slumps.
3. Home Field Advantage: Historical data shows that home teams win approximately 54% of MLB games. Our model accounts for this baseline advantage and adjusts for park-specific factors that may amplify or reduce the effect.
4. Recent Team Form:We calculate a form index based on the team's last 10 games, factoring in run differential, win percentage, and quality of opponents faced. Hot and cold streaks are real phenomena that historical models have shown to have mild predictive value.
5. Bullpen Status: Aggregated relief pitcher ERA, recent usage patterns, and closer availability. A team whose top relievers pitched the previous two days faces a measurable disadvantage in close games.

Model Limitations

We believe in transparency about what our model cannot do. It does not account for weather conditions, specific lineup construction beyond aggregate team stats, or day-of-game scratches. No pre-game model can predict outcomes with certainty. Our probabilities represent the most likely outcome given available data, not guarantees. Baseball is inherently unpredictable, and that is part of what makes it great.

For each game, StatScope provides a side-by-side sabermetric comparison of both teams' active rosters. This goes beyond simple stat comparisons to provide actionable analytical insights.

Hitting Analysis

We aggregate individual hitter stats across the active lineup to produce team-level offensive profiles. Key metrics compared include team batting average, on-base percentage, slugging, wOBA, and ISO. We also calculate platoon splits (performance vs. left-handed and right-handed pitching) to identify potential lineup advantages against specific starters.

Pitching Staff Breakdown

Beyond the starting pitcher comparison, we analyze each team's bullpen depth. This includes individual reliever ERA, FIP, K%, and recent workload. We flag relievers who may be unavailable due to consecutive-day usage and identify the highest-leverage arms available for late-inning situations.

Player Form Index

StatScope calculates a proprietary form index for each player based on their performance over the last 7 and 15 games compared to their season average. This helps identify players who are currently hot or cold, which can influence in-game matchup decisions and pinch-hitting opportunities.

Our head-to-head matchup tool pulls historical batter vs. pitcher records and presents them alongside contextual analysis. When a batter has faced a specific pitcher multiple times, we display the full statistical record including at-bats, hits, home runs, strikeouts, and batting average.

However, we also caution against over-relying on small sample sizes. A batter going 1-for-10 against a pitcher does not necessarily mean they struggle against them. With proper context, we note when sample sizes are too small to draw reliable conclusions and suggest looking at performance against similar pitcher types (fastball-dominant, breaking-ball heavy, etc.) for a more meaningful analysis.

The matchup tool is most valuable when combined with other factors: the pitcher's current form, the batter's platoon splits, and the game situation. We aim to present data that enhances understanding rather than providing false certainty from limited information.

StatScope uses AI to generate pre-game analysis notes and in-game commentary that synthesize raw data into readable narratives. These AI-generated insights are clearly labeled and should be understood as data-driven observations, not expert predictions.

The AI commentary considers starting pitcher matchups, recent team performance, head-to-head records, and park factors to produce analysis that highlights the most interesting statistical angles for each game. This feature is designed to save fans time by surfacing the most relevant analytical talking points rather than requiring them to dig through raw data themselves.

We continuously refine our AI commentary system based on user feedback and analytical accuracy. If you notice any factual errors or have suggestions for improvement, please contact us at statscope.help@gmail.com.

StatScope is committed to providing accurate, unbiased, data-driven baseball analysis. Our editorial principles include:

• Data accuracy: All statistics are sourced from official MLB data and verified through automated consistency checks. When discrepancies are found, we flag them and default to the official source.
• Transparency: We clearly label the sources, methodologies, and limitations of our analysis. When our models have known blind spots, we say so.
• No gambling promotion: StatScope does not provide betting odds, gambling advice, or promote sports betting in any form. Our analysis is for educational and entertainment purposes only.
• Independence: We are not affiliated with MLB, any team, or any sports betting company. Our analysis is independent and unbiased.
• Corrections: If we make an error in our analysis or data presentation, we correct it promptly and transparently. Users can report issues to statscope.help@gmail.com.