Hi there!

First time poster here, I am hoping this is the right channel for this topic.

I've developed a new variant of Sequential Probability Ratio Test (SPRT) for online A/B test, based on the Sequential Probability Ratio t Test described in this paper (i.e. Hajnal Sequential t-Test).

I am quite happy with the new variant's performance ( in terms of statistical power, type 1 error rate and time to decision) using both simulated data and actual historical data we've seen in our online A/B test.

I am wondering if anyone would be keen to peer review this variant? If this variant has legs, we could potentially publish a paper about it. 

Thank you very much.

Derek