I don’t think you’ve quite gotten it right. Maybe I wasn’t clear in the post, in fact I feel fairly confident that I was not as clear as I should have been, so let me try again:

Imagine that we are playing this game following actual purchases, that is, we have a price number in the back of our heads.

Imagine also that each participant knows of the existence of some nonzero number of potential players who will always answer with the price they actually paid. Their reasons for acting are irrelevant: They may be Christians who are commanded not to lie, or Kantians who will tell the truth because doing so is an end in itself, or they may be insomniacs who think they will sleep more soundly if they are honest. They may actually be deluded about the nature of the game or misunderstand it entirely. It hardly matters; in each case, it behooves you to mimic their behavior.

If you guess higher, then the honest person wins, and you are penalized. If you guess honestly, then you both gain that amount. If you guess lower, then you will likely be rewarded with less than you might otherwise have gotten by guessing honestly.

Moreover, because these considerations induce other players to guess honestly as well, the people who are honest for honesty’s sake end up setting a pattern that others will follow. They are the seeds around which a pattern of honesty coalesces.

And, in any case, the existence of receipts is not a triviality and should not be dismissed out of hand. To my mind it’s rather like a biologist deciding to study humans and then declaring that nerve cells are extraneous to his project. If you want to make claims about the market, then study the market. Don’t study some parts of it and assume that you can generalize for the whole.

So talk of “receipts” misses the point. The backstory (compensation for lost goods) is mere window-dressing, an inessential feature of the thought experiment. We could presumably imagine any number of different scenarios where the payoffs / decision matrix formally parallel the case described here.

Market cues also make no real difference to the case, as explained here. (We already have a tempting “default” value, namely the maximum of 100. The reasoning applies no differently to this than it does to a market-given starting point.)

So what is the point, you may ask? As the article explains, the TD shows that the sort of “economist’s rationality” behind the $2 outcome is (1) descriptively inaccurate, insofar as it mispredicts the actual choices people make; and (2) normatively inadequate, insofar as we would end up worse off if we followed it.

This doesn’t necessarily mean there’s anything wrong with capitalism, of course. It’s simply pointing out these two fundamental shortcomings in the understanding of “rationality” assumed by most theoretical economists.