This looks excellent. I think this is a good intermediate step to meet the immediate needs. Let's merge and keep discussing if we want to create something that is fully automated or not.

Per conversation with Meikang and Rui, here are notes on two possible routes to implementing an automated system for selecting the number of runs, which would not be visible to the user and instead just run a scientifically-valid number of trials. TBD next steps in the new year. https://bmgf-my.sharepoint.com/:w:/g/personal/brittany_hagedorn_gatesfoundation_org/EV0DHld61gJAnuzZd3Yc2KIBmvtn9Ve1HjTc6gJ4n6zZow?e=bkRigJ

https://kb.palisade.com/index.php?pg=kb.page&id=85#:~:text=To%20have%20%40RISK%20monitor%20convergence,for%20any%20or%20all%20outputs.
This website documents how At Risk (an add-in monte carlo simulation tool for Excel) tests convergence. It has a series of articles on specifying a convergence criteria and things to consider based on the test statistic we choose (mean, sd, or percentile). The approach appears to be similar to option 2 we discussed, and we could use these articles as reference to corroborate our calculations.