Even though it looks as though RII (which says that any strange diet produces brief weight loss, then a couple of months of plateau, then you fatten up again), looks like what's happening here, I'm just going to abandon all the R theories.
For various complicated reasons I think I was wrong to introduce them in the first place. So I'm just going to pretend I didn't.
So in fact I'll pretend my prior was:
My current state is some appetite loss (but not like the first month), belt notch=3 (although that feels slightly tighter than it did last month)
Likelihoods of that result according to various theories:
H 50/203 : W 5/134 : S 25/134
H 2000/203: W 180/134: S 225/134
rounding off to scores out of 100:
H 77: W 10: S 13
This seems pretty reasonable. The willpower theory is getting annihilated because it's not predicting appetite loss and there's no weight gain. Shangri-La is looking better by comparison, and Helplessness, having won this round, is looking good.
I think that's what I actually believe.
The standard Willpower theory says I should just get fatter if I eat more food. This hasn't been happening.
Shangri-La's month of success has given it some credibility even though I originally thought it mad. This month has damaged it though.
The Helplessness theory, that your weight stays the same whatever, and if you eat more you get slightly less hungry to make up doesn't look like quite as good an explanation for the observed results as Shangri-La, but on the other hand, it looked much more likely before I started, and the difference is not great enough to wipe out that initial advantage.
But I want to give Shangri-La its best shot. So next month I'm going to try my damnedest to make it work. Various types of oils, the sugar water version, noseclips, the works. If that works (again), then I'm going to believe that something interesting is going on. If that fails then I think I have my answer.