This is a really cool game. I was so off!

Then report the average of this metric over time with each game.

Does the orange mean your answer is within 25% of the absolute value? Or that your logarithm value is within 25% of the logarithm value of the true answer?

Thanks for making this, this is awesome

This is especially the case when the question is asking for a bounded number in the first place (eg a percentage). In fact I'm pretty certain you should _always_ succeed within 4 steps given +-10 on a percentage question and nearly always within 3 steps. ChatGPT says it's provably so but I'm not smart enough to verify. Rings true though.

Certainly made easier by knowing whether it's higher or lower, and especially with the yellow arrows if you're not too far off.

One UX change that might be nice is to have a "spoken" version of your guess live-update below the input. I keep having to count zeroes and it would be nicer to see "Eleven billion".

After about ~10 questions though, I started getting the same question every time. Like five times in a row.

For those interested, I did polish the initial app a lot: https://fermi-game.onrender.com/ (bad news though... I over-engineered it even further I think. It's my first real public project, so I learned my lesson to viciously descope the mvp). Some of the comments here (like scientific notation and sharing) are present in my project. I tried to re-share after polishing but the HN link sharing dynamics have been a bit opaque to me and kept the project buried when posted.

It's clear to me that there is a lane here for a fun brain teaser/exercise. Just getting the answer right on 2 tried on OP's version by guessing ~5% of 330M population buying new car was a nice hit of dopamine. Combining a little math and world-knowledge is pleasing, it would seem.

@danielfetz, any interest in collaborating?

Otherwise the UI and concept looks pretty interesting. But until that is fixed, it is unplayable for me.

(On an iPhone, using Safari)

So, I any guess above 1 000 is truncated. 1 000 is 1.

With a target of 20% accuracy, it won't make much difference, but I think that symmetrical error bounds are appropriate in this case - the factor by which the answer is wrong. so 2 times too big, is as good as 2 times too small.

I think I did find a solution to bring the focus a bit away from the binary search and would greatly appreciate feedback from all of you.

The game now shows a hint after the second incorrect guess. For example the hint "The US covers 1.87% of the Earth's surface." is displayed for the question about what percentage of the Earth's surface is land.

I hope that this brings into the game a whole new dimension where you have a second moment after the initial Fermi estimation to really think through your guess and the assumptions you have made. How does the new information received through the hint impact your model?

I think those things put together now make the game a very compelling training ground for getting better at Fermi estimation and updating your beliefs in light of new information without over or under-reacting.