Well, in this case the control does not really match the treatment group at all, so I would suggest that may be a serious flaw. If the hypothesis is that the pr

Well, in this case the control does not really match the treatment group at all, so I would suggest that may be a serious flaw. If the hypothesis is that the proprietary algorithm is responsible for the change in the treatment group, it is an odd choice to use an entirely different game as a control, rather than the same game with the proprietary adaptive algorithm removed. (Evidently it is possible to remove just the algorithm, as the demo version lacks the treatment algorithm.) I'm a bit skeptical of the utility of the study. (I also take minor issue with a study "sponsored by Alkili" being reasonably construed as independent, given that "the funder had a role in study conception and design, confirming data and statistical analyses, and conducting the study." But this of course does not necessarily invalidate the underlying science.)