If the theoretical maximum is relatively low, that is also not a good thing

But you have your eyes set on the IgNobel I take it :-)

Cool, but go Bayesian. The effect size estimate is the *entire posterior distribution*.

I would certainly tune in for that!

Let me guess, is it when you mistakenly preregistered a frequentist analysis, but then you learn that it can't actually answer your research question, which is fundamentally Bayesian? ;-)

And what exactly do we conclude or learnwhen a null hypothesis has a certain compatibility with the data? We want to quantify evidence and assess the credibility of hypotheses. Not a single mention of Bayesian inference is made...