I’m talking about maximizing logarithmic expectations rather than logarithmic Maximizing Expectations, which is correct in a compound interest environment. The implicit assumption is that it is a baccarat that can be tried over and over again an infinite number of times, and going bankrupt in the process means losing all investment opportunities from that point onward, so the opportunity loss is negative and infinite. On the other hand, the condition of being able to repeat the challenge an infinite number of times is impossible for people with a limited lifespan, so realistically, it would require an investment above the Kelly standard. Kelly criterion - Wikipedia
I’m running a simulation at the beginning of this article. [What’s important is (actually) what gambling teaches us: concentration and diversification and deduction by “good religion” - Objective Subjectivism https://objsbjvism.wordpress.com/2018/04/29/%E5%A4% A7%E5%88%87%E3%81%AA%E3%81%93%E3%81%A8%E3%81%AF%E5%AE%9F%E3%81%AF%E3%82%AE%E3%83%A3%E3%83%B3%E3%83%96%E3%83%AB%E3%81%8C%E6%95%99%E 3%81%88%E3%81%A6%E3%81%8F%E3%82%8C%E3%82%8B-%E9%9B%86/?fbclid=IwAR2Y3ax0ytPJjM2EUjETO-2MCocGCk2AjJTX8MKqnQ4p9PeLqdr6JV9azHg] This simulation has 10,000 trials, so even if we assume one trial per day, that’s about 30 years. If your total assets are 10 million yen, it is optimal to put 50,000 yen in a baccarat every day. Of course, unlike this simulation, the real-life baccarat has no known expected value or win rate.
- I wrote about a related topic in [Serialization Change your point of view
- Vol. 4 To Win Big, You Need to Bet Big: Change Your Perspective｜gihyo.jp … Gijyosha, Ltd.
- Talking about how the option that maximizes expectations is not the optimal choice.
- I’m connecting it to prospect theory from there.
This page is auto-translated from /nishio/対数期待値の最大化 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.