Someone built a website that assigns Elo ratings to San Francisco judges based on how serious the charges are that they dismiss or divert. The project pulls from public criminal court records and asks visitors to vote on which "released charge sounds worse" to generate comparative rankings across the bench.

The dataset includes 8,156 eligible charges across 46 judges and 2,937 defendants. Higher ratings mean voters picked worse-sounding charges that got dismissed or diverted. Judge Daniel A. Flores tops the list with a 1124 rating, followed by Dawn R. Payne at 1121.4 and Loretta M. Giorgi at 1114. The methodology focuses on dispositions with no jail time, including diversions, dismissals, acquittals, and deferred judgments.

The project, hosted on GitHub under jamiequint/sf_criminal_court, transforms opaque court data into something searchable and interactive. By gamifying the information, it makes judicial outcomes accessible to the public in a way traditional court records never achieve.