The Case Outcome Predictor estimates the probability that a specific judge will rule in favor of the moving party on a specific motion type. It is a risk assessment tool, not a guarantee.
What the predictor uses. The model is trained on thousands of California superior court tentative rulings. For each prediction, it draws on: (1) the target judge's historical grant/deny rate for the motion type, (2) patterns in the judge's tentative ruling language that correlate with outcomes, (3) attorney observation data about what arguments and presentations have fared well or poorly, and (4) any contextual facts you provide in the optional case description field.
Running a prediction. Navigate to the judge profile and click Run Prediction in the AI Tools panel, or access it from a hearing in your case tracker. Select the motion type, enter an optional description of your case theory and key facts, and click Generate. Results appear in 10–30 seconds.
Reading the results. The prediction shows a probability percentage (e.g., '68% likelihood of grant') with a confidence band (e.g., ±12%). The confidence band reflects data richness — wider bands mean less historical data for this judge and motion type. Below the estimate, the Key Factors section explains which elements of your case description and the judge's history most influenced the prediction. The Risk Flags section surfaces specific patterns that have correlated with denials before this judge.
How to use predictions responsibly. Treat the prediction as one input among many, not as a final answer. A 68% grant probability still means a substantial chance of denial. Use the Key Factors output to strengthen your brief and the Risk Flags to identify arguments that need preemptive addressing. Do not share prediction percentages with clients as if they were case-specific legal analysis — they are population-level statistical estimates.
Plan limits. Pro users receive 5 predictions per month. Firm and Enterprise users have unlimited predictions.