Set to 0 to simulate a null experiment (no real difference)
Observations collected at each interim check
How many times you check during the experiment
More runs = more stable estimates (max 50). Changes here require clicking Run.
📈
Configure simulation parameters and click Run to see results.
Naive Repeated Testing False Positive Rate
—
Sequential Testing (OBF) False Positive Rate
—
Naive Repeated Testing — Running p-values Across Simulations
Each line is one simulated experiment run.
Red lines crossed α at least once during the run —
a false positive when no true effect exists.
Sequential Testing with O'Brien-Fleming Boundary
The green dashed boundary demands a much more extreme result early in the experiment and relaxes
as more data accumulates — enabling valid early stopping without inflating the type I error rate.