Pre/Dicta has profiled 750 federal judges and built an algorithm that can predict how each judge will rule on a motion to dismiss based on the cause of action, the characteristics of the parties and the attorneys. It runs this prediction based on one piece of data – the docket number!
It does not read the complaint, it does not care about the law, the facts or the defenses. Are you as horrified as I am? The problem is that Dan Rabinowitz built a methodology and an algorithm that was trained on over 50,000 federal motions and is right 85% of the time!
Earlier this year Pre/Dicta acquired the data created by the state court analytics pioneer Gavelytics which had gone out of business last year.
The full article is available on LegalTechHub