9 Comments
Apr 12, 2021Liked by Leon Lin

Interesting observation: If you change the "stake" amount to something more reasonable, like say betting 10% of your current wealth instead of all of it, the results are much better. Also increase the number of flips to 1000 to get a better idea of long-term behavior. Somewhere between 0% and 5% of people lose capital, and the average gain is much higher too.

Expand full comment
Apr 12, 2021Liked by Leon Lin

Which overall makes sense. The more risk you take, the more likely that loss will be incurred, but the larger the potential payoff.

Expand full comment
author

Yep! changing the % bet (or the % win, % loss etc inputs) will affect the overall return profile, both for the individual and the ensemble.

I'm less sure changing the number of flips makes that much of a material difference though - the trend should be pretty clear with around 100 flips - though feel free to play around with the colab notebook to adjust

unsure if you mean to reference the kelly criterion in the comment, and yep that's another key point regarding bet sizing and how to reduce risk. Got a bit tired at the end of the post so gave a bullet on it to reference an older post: https://avoidboringpeople.substack.com/p/having-faith-in-the-kelly-criterion

Expand full comment
Apr 4, 2021Liked by Leon Lin

Really good reminder on path dependence. Thanks!

Expand full comment

this is mind blowing, best piece I've read in 2021!

Expand full comment
author

thank you!

Expand full comment

Good stuff, thanks Leon!

Expand full comment
author

You’re welcome!

Expand full comment
Comment deleted
Expand full comment
author

ty! LMK if you learn anything interesting - I only just got around to understanding the topic

Expand full comment