I'm curious how they handle model error (the error when your model is totally wrong).

They punish it. That is, your stated credence should include both your 'inside view' error of "How confident is my mythology module in this answer?" and your 'outside view' error of "How confident am I in my mythology module?"

One of the primary benefits of playing a Credence Game like this one is it gives you a sense of those outside view confidences. I am, for example, able to tell which of two American postmasters general came first at the 60% level, simply by using the heuristic of "which of these names sounds more old-timey?", but am at the 50% level (i.e. pure chance) in determining which sports team won a game by comparing their names.

But it seems hard to guess beforehand that the question you thought you were answering wasn't the question that you were being asked!

This is the sort of thing you learn by answering a bunch of questions from the same person, or by having a lawyer-sense of "how many qualifications would I need to add or remove to this sentence to be sure?".

Do you have some links to calibration training? I'm curious how they handle model error (the error when your model is totally wrong).

For question 10 for example, I'm guessing that many more people would have gotten the correct answer if the question was something like "Name the best selling PC game, where best selling solely counts units not gross, number of box purchases and not subscriptions, and also does not count games packaged with other software?" instead of "What is the best-selling computer game of all time?". I'm guessing most people answered WOW or Solitaire/Minesweeper or Tetris, each of which would be the correct answer if you remove on of those restraiints.

But it seems hard to guess beforehand that the question you thought you were answering wasn't the question that you were being asked! So you'd end up distributing that model error relatively evenly over all the questions, and so you'd end up underconfident on the questions where the model was straightfoward and correct and overconfident when the question wasn't as simple as it appeared.

I've always believed having an issue with utility monsters is either a lack of imagination or a bad definition of utility (if your definition of utility is "happiness" then a utility monster seems grotesque, but that's because your definition of utility is narrow and lousy).

We don't even need to stretch to create a utility monster. Imagine there's a spacecraft that's been damaged in deep space. There's four survivors, three are badly wounded and one is relatively unharmed. There's enough air for four humans to survive one day or one human to survive four days. The closest rescue ship is three days away. After assessing the situation and verifying the air supply, the three wounded crewmembers sacrifice themselves so the one is rescued.

To quote Nozick from wikipedia: "Utilitarian theory is embarrassed by the possibility of utility monsters who get enormously greater sums of utility from any sacrifice of others than these others lose . . . the theory seems to require that we all be sacrificed in the monster's maw, in order to increase total utility." That is exactly what happens on the spaceship, but most people here would find it pretty reasonable. A real utility monster would look more like that than some super-happy alien.

It seems to me that, unless one is already a powerful person, the best thing one can do to gain optimization power is building relationships with people more powerful than oneself. To the extant that this easily trumps the vast majority of other failings (epistemic rationality wise) as discussed on LW. So why aren't we discussing how to do better at this regularly? A couple explanations immediately leap to mind:

1. Not a core competency of the sort of people LW attracts.

2. Rewards not as immediate as the sort of epiphany porn that some of LW generates.

3. Ugh fields. Especially in regard to things that are considered manipulative when reasoned about explicitly, even though we all do them all the time anyway.

Build a better one yourself? I'm tired of eating.

I don't understand why objectivists seem to be held in low regard here. My exposure is limited to browsing a forum of objectivists[1] - they were indistinguishable from those here, though much more focussed on personal instrumental rationality in their topics.

I know they are formally a closed loop belief system limited by the writings of Ayn Rand (which I've not read), and have heard this belief system is flawed in some way. That sounds like a straw man.

I'm only interested in the steel man. What is the difference between rationality and objectivism?
The only one that comes to mind: Objectivism implies there is only one true way of some things, while rationality allows for individual variety in thinking processes (resulting from different information, experiences, terminal values, etc.)

However, if for one person their most desired thing is happiness - which can only be achieved through quasi-altruistic deeds - then I cannot see it as anything but objective and rational to carry out those deeds. Objectivism applied to the fulfilment of one's desires appears indistinguishable from rationality to me. Where am I wrong on this - or am I playing semantics?

[1] Knowing what to look for, I discovered the site again. They indeed are very skilled at applying instrumental rationality to various areas of their lives (exempli gratia what type of plastic surgery yields the most natural results?) - however in the Philosophy and Ethics sections, Ayn Rand philosophy abounds. They are describing the intent of an Important Figure (scary), without doing so for the purposes of then breaking it down; those that try the latter attack straw-man versions and are refuted.

Could you try using smaller candy?

The way the feeder is built, that wouldn't really help. It dispenses a constant volume, not a set number of candies. I could try to reduce the dispensed volume further, but I think other techniques would be best to try first.

if the reward is in the system, I tend not to wait very long before using it.

This seems OK to me.

It's not a problem except insofar as it interferes with some of the rules.

Or perhaps giving 1 candy per N points? Or giving a candy with probability 1/N?

These are the two big options I'm considering for next time. I'm leaning towards the "1 candy per N points" model, because that allows me to "gamify" the system with a big XP bar.