Assignment 2 Feedback

[venpopov]

I see you got a headstart on the 2nd assignment! I think you are on the right track with the last couple of simulations. The idea that there are fewer ways to get more than the median that to stay under it is compelling. As the goal for this assignment is more to get a feel for how simulations help us understand mechanisms, and to get a bit of practice with the github+quarto system, what you have is already sufficient.

But your idea is good, so I would advise you to explore this a bit more to establish better the mechanism. You note that:

https://github.com/HubiAtGit/hubert_baechli_ICMB_HS24/blob/d11dbf92948c0516e36a6dcf22d824400a65c205/hw2/index.qmd#L338

and you are correct :) It would be more precise to say that to go above a 1000, you would need to meet not someone with more than thousand, but that your combined wealth is at least a 1000. E.g. if you have 300, you need to meet someone with X, such that 300+X > 1000 AND that the random proportion you get of the common pot p_split is such that p_split * (300+X) > 1000. That makes it a bit more complicated than your description, but also more difficult to understand. It could be useful to simplify the simulation further and assume that the smallest unit is 1$ ( so that you don't get fractions), that you have a small set of people (10?), starting each with a small sum (5?) and try to figure out the space of possible outcomes and their probabilities. Not sure if it will work, but worth checking if you want to pursue this further to establish your explanation better.

Apart from that, just clean up your document so that it is a bit easier to follow. This is fine for while you are still developing it (I often have a similar level of detail), but just a bit more text would help add some clarity not only for me, but also for yourself when you go back to a simulation project in a few weeks/months (in a real setting) and suddenly you wonder why you did some things)

[venpopov]

(not sure if you wanted feedback at this stage - I was just excited to see that someone had already taken a stab at the assignment and got pretty far already)

[HubiAtGit]

Dear Ven First of all, thanks for your feedback, it's much appreciated!

Yesterday I tried to calculate the probability to get above the median (should be online now)

However, this is not absolutly correct at the moment.

• For one thing, the data are not from a single simulation
• On the other hand, there are 101 possibilities in the exchange with my simple calculation only 100

I already know how to fix this, and I plan to do so before the submission in two weeks. But I think it should be accurate enough that there shouldn't be many surprises.

The exciting thing is that the population median is stabilizing again after a few iterations... ... I will also get to the bottom of this 😉

I'm also planning a version with a better syntax and description by then. My English and programming skills are unfortunately quite beginner-like, which is why I prefer not to write or talk anything. But I will try my best

Best regards

[venpopov]

Oh, your current simulation is really clever! Nice. You are close to figuring it out. Yes, I noticed that the median chance of “getting over the median” seems to be stabilizing and I think understanding why that happens could be key. I will leave you to it, so you don't feel pressured by the real-time feedback and will comment again after the deadline