Clarity on adjustments for sequential monitoring

[tmanndd]

Hello,

I've built out a program for monitoring tests sequentially which is based on this example: https://github.com/spotify/confidence/blob/master/examples/Confidence%20Examples%20-%20Sequential.ipynb.

I'm hoping to understand more regarding what adjustments are being applied. From what I understand, there is an adjustment that is being applied with respect to the information quotient to account for the inflated risk of peaking - am I right in thinking that, once the quotient is reached, there should not be any adjustments once the target users is met?

If this is the case, then I am still seeing the confidence interval bounds being adjusted past this point - are these adjustments, then, due to "multiple comparisons" with respect to the number of days the test has been live and therefore the bonferroni correction ?

I'd be interested to know more regarding what adjustments are being applied and for which situations (i.e., when is it adjusting for the information quotient, bonferroni, or both?).

Thanks for your time! This is a really cool repository and I've been having fun using it. Let me know if you need any more information.