Breaking Down a Rivalry Matchup With Actual Numbers
Most people talking about classic matchups like Dodgers versus Red Sox are just throwing around names and headlines. The actual interesting part is in the player stats. When you look at the regular season data from the last few years, a pattern emerges that neither casual fans nor broadcast analysts really talk about enough. The pitching mismatches tell a different story than the overall team ERA suggests. Here is how I actually approach this. I pull the starting rotation data, then cross-reference each pitcher's split stats against the opponent's lineup construction. The Dodgers tend to chase high-fastball zones, and the Red Sox have three or four hitters who handle heaters better than league average. That matchup advantage compounds across six innings. It shows up in the xERA and whisker-ball% numbers before it shows up in runs allowed. One thing nobody warns you about is roster churn during the season. A player who hit .310 with runners in scoring position in June might drop to .240 by September because of a tweak to the opposing catcher's pitch framing data. I learned this the hard way when I spent two weeks building a breakdown on Mookie Betts using May and June splits, only to realize his production dropped off sharply against left-handed pitching by August. The fix was simple: I switched to rolling three-month windows and added a weighted recent-game factor. It takes about twenty minutes longer upfront, but it prevents you from building an entire projection on stale data.
For the Dodgers-Red Sox angle specifically, the advanced metrics point to three key levers. First is bullpen usage. The Red Sox historically lean on their setup man more aggressively in close games, which leaves their fifth starter exposed earlier in the game. Second is park factor. Fenway's dimensions inflate home run rates for right-handed pull hitters, and the Dodgers' lineup construction plays into that. Third is the umpire. Strike zone variance matters more than most people admit, and certain home plate umpires call a consistently tighter zone for curveballs. If you want to replicate this yourself, you can pull Statcast data directly from Baseball Savant. It is free. You do not need a subscription. The export feature lets you grab batted ball data for any series, and you can filter by pitcher hand, spray chart, and launch angle. From there, a basic spreadsheet with pivot tables will get you most of the way. I recommend adding a column for expected batting average on balls in play, or xBA, because it removes the luck factor and shows what actually happened on contact. It usually changes the narrative by a full win expectation in close matchups. The main limitation with this approach is that it cannot predict clutch performance or in-game managerial decisions. No stat sheet captures a manager pulling a starter after four innings because he looks uncomfortable. You have to watch the game to understand that layer. Also, small sample sizes in head-to-head matchups can be misleading. If two pitchers faced each other three times last year, that data point is noise, not signal. I ignore anything under fifty plate appearances between a pitcher and a specific batter split, unless the exit velocity and barrel rate are outliers, which is rare.
For a complete walkthrough with download links to the spreadsheets and a video breakdown of the process, check out the tutorial page linked on the official sabermetrics community hub. It covers the Statcast export steps, the pivot table setup, and a pre-built template you can adapt for any matchup, not just Dodgers versus Red Sox. It should cut your setup time from a couple hours down to under thirty minutes once you have the template in place.
Get the Full Details
