Ask HN: Has anyone tried Ray Dalio's method of algorithmic decision-making?
Ray Dalio describes the way he makes decisions as: 1. making mistakes and 2. storing the lessons learned from those mistakes in a computer to better help with his decion-making criteria.
I'm wondering if anyone's done this and what that program would look like?
I've only done what I see described here in big-picture studies and simulations. It did help though. An example would be in financial planning & simulation, where there's a lot of objective leverage information available. Dates, growth rates, technical indicators, and so on. This can all be manipulated algorithmically. I even did some of it in hobby languages like SmallBASIC (the non-MS version, really interesting project).
In fact what he describes sounds like & overlaps a lot with a technical trader's mindset. In that world you are always tinkering with new criteria and storing the updates (mistake-derived as they often are) as scanners, models, etc. (You may find this "software" in ThinkOrSwim or TradingView scripts, etc.)
In other areas that aren't so tractable to pure algorithm, I keep it more subjective in the best way I can (subjectivity also has big strengths to leverage) and update my various systems & frameworks more qualitatively, like by updating & rearranging lists of effective steps, designing new thought models, and so on.
BTW, did he not go into further depth than that?
I remember reading one of Dalio's books some time ago and he was big on principles (200+) but didn't share many details and didn't seem to have created a unifying system. To me that'd be frustrating or maybe even a non-starter under the circumstances for a lot of reasons, i.e. the principles will necessarily start to contradict one another, access to principles by newcomers is necessarily less efficient, and so on.
Honestly, I think that his book Principles (which is the one I'm reading) is just really dense and reads like an encyclopedia. Possible that it's just really hard to get through?
The decision-making frameworks that he describes go down to a micro-level. I do think that a lot of this stuff maps to engs like us, esp. the tradeoffs we make in deciding to pursue projects, skills, and jobs.
I came to his book after working closely with someone who created a unifying system, documented it, and trained large groups in its use, so that may be why it stood out as a problem.
The difference was obvious. 200 principles may be "just" really hard to get through from one perspective, but from a systems standpoint it's really only at square one.
Not a judgment on the principles themselves of course. But unfortunately principle-compilers (more like broad describers of phenomena) and model-architects (more like deep designers of interventions) are often two different people entirely.
In fact what he describes sounds like & overlaps a lot with a technical trader's mindset. In that world you are always tinkering with new criteria and storing the updates (mistake-derived as they often are) as scanners, models, etc. (You may find this "software" in ThinkOrSwim or TradingView scripts, etc.)
In other areas that aren't so tractable to pure algorithm, I keep it more subjective in the best way I can (subjectivity also has big strengths to leverage) and update my various systems & frameworks more qualitatively, like by updating & rearranging lists of effective steps, designing new thought models, and so on.
BTW, did he not go into further depth than that?
I remember reading one of Dalio's books some time ago and he was big on principles (200+) but didn't share many details and didn't seem to have created a unifying system. To me that'd be frustrating or maybe even a non-starter under the circumstances for a lot of reasons, i.e. the principles will necessarily start to contradict one another, access to principles by newcomers is necessarily less efficient, and so on.
The decision-making frameworks that he describes go down to a micro-level. I do think that a lot of this stuff maps to engs like us, esp. the tradeoffs we make in deciding to pursue projects, skills, and jobs.
The difference was obvious. 200 principles may be "just" really hard to get through from one perspective, but from a systems standpoint it's really only at square one.
Not a judgment on the principles themselves of course. But unfortunately principle-compilers (more like broad describers of phenomena) and model-architects (more like deep designers of interventions) are often two different people entirely.