Test in Prod
I've had this thought as my operating motto for a very long time. Not because I don't value security or bug tests. Thankfully, GitHub CI automates a lot of that.
It's because asset management is nearly impossible in a vacuum, as per the allocation wager post's position sizing. I first noticed this when I moved from that paper Thinkorswim account to real Ameritrade funds. The losses there made their way into the end of Nine to Noon.
The problem in markets is volume and supply. Your transactions create or remove demand, which cannot be accurately simulated by any central algorithm.1 In my case, it showed up as easy, constant profits at unrealistic fill prices.2
This came up recently in my research of Yanis Varoufakis's work for the DUNA donation post, plus my execution of the Cats retroactive consideration vote.3 Yanis remarked how an economist's job is impossible to simulate or backtest well. You can run models, forge theories, and meaningfully contribute to the field; but you can never go back in time and see what America looks like without the New Deal.
Asset management is really similar. You can ogle all you want at the number which goes up, but all that really matters was the choice you made 9 years ago. The next best hope is doing something smart now.
That's where money differs from the mental model held by most engineers. You can't pay someone in "sandbox capital" which will eventually merge into USD.4 It's results and formation now, or no substantive efforts.
I want to make this my life motto tattooed across my back. But I said that about a stylized ticker set after a particularly nice trade in EUR/NOK back in college. I'll sleep on it for now, but it is my operating modicum.
-
Trust me, the communists tried. See the first widespread minifarm in Brazil or the bureaucratic structure of early Leningrad. These are well-documented cases of trying early human-assisted AI up against markets, and the force of the crowd always wins. ↩︎
-
I really honed this later on to execute the toppers on my BTC trades. The old adage among asset managers of any reasonable size or leverage is "you have to sell when the market's rising, and buy when it's falling." Because at any other point, you'll practically need a market order to execute, hence poor slippage. ↩︎
-
As stated in smaller forums, thank you guys so much for the unilateral (non-Cats) support of this proposal. I know it was a significant portion of our treasury, written with a lot of typos, and completely unprecedented in our 2-year history. Your faith in my ability to deliver means a lot. ↩︎
-
Stock corporations help merge this divide with a sort of interim currency which may have future value. But it's not really taken as seriously as paychecks (or as a standalone replacement for them). This really shows when a nonprofit immediately converts received securities into cash to operate, since any other choice would require investment oversight. ↩︎