Time to make some money, then. At least I'll have something concrete when I eventually lose the coin, one way or another. And it won't destroy my agency either, hopefully. I turned back to my laptop to look at some stocks. My original plan was to pick 16 promising highly volatile stocks and after four coin tosses I would know which one of them went up most during the next week. Then I'd just buy the best one with as much leverage as I could get.

It took me maybe ten minutes to realize that while picking high-volatility stocks isn't that difficult, there's an easier way to do this, and it's way more efficient too. And even better, getting "caught" for insider trading or whatever this might get flagged as would be impossible too. I had already looked at the bitcoin price as a test for the coin before, how had that slipped my mind? All notable cryptocurrencies have high volatility and there aren't even that many of them. Instead of trying to pick the best one, I could just ask if it would go up at least 5% during the next week, then buy as much as I could and place a limit order selling it when it was up at least that much.

That could be improved even more, I thought. The actual peak could be determined quite precisely with binary search in just a few iterations. If the value would be falling instead, I could do shorting. And maybe it would be possible to pick a set of assets that would on average go up using some probabilistic tricks? Anybody actually knowing any maths could probably make way more money with this.

But even with a simple scheme it should be possible to make quite a bit of money quickly. But how much? Some back-of-the-envelope math would help. A quick search shows that ethereum has around 1.4% hourly volatility. ChatGPT, which I trust almost as much as the newspaper op-eds when it comes to financial decisions, computes a 78% probability that it will be up at least 5% during the next week given some questionable assumptions about drift and normal distribution. That makes no sense, unless in the remaining 22% of worlds you're losing almost all your money. Otherwise everybody would be doing this, magic coin or not. Whatever, at least I was completely insulated from losing money with this approach. That meant a single coin toss had expected return of 0.78 * 1.05 + 0.22 = 1.04. The average duration would probably be less than a week, perhaps 5 days would be a good estimate. With about eight tosses per day, I'd need 40 cryptos with eth-like properties for full saturation, which looked realistic. So that's about 5% weekly return, which would 12x the money per year, and 2000x it in three years. Nice exponential growth, but that wouldn't even make me properly rich in the next year, which was way too slow. Even taking all loans I could get without mixing people in real life, investing a six-digit sum on this wouldn't be easy, and KYC limits on legal crypto providers would make even that difficult.

Getting a better model should be my first priority then, if I were to make money. Single coin toss would be valued at about a billion dollars each in case I could get over the worst information asymmetry barrier in human history, and that made this quite frustrating. I should just use the coin to come up with ways to do this. But that seemed quite scary and stressful. Maybe getting started right now, just making a bit, would be good enough. Or at least better than doing nothing.

Ethereum price will be up at least 5% sometime during the next week. Land tree-side if yes, non-tree otherwise.

I didn't even bother typing it down, I just blurted it out. The coin landed on the scratched side. So much for getting started right now.