Simon

Most systems in this space blur two different jobs. Predicting what a market will do and deciding what to do about it are separate problems, with separate failure modes. A system that fuses them cannot tell you which half was wrong when it loses.

Simon keeps them apart. The prediction layer states what it expects and how confident it is. The strategy layer decides what, if anything, to do with that. When the prediction layer does not know, it says so, and the strategy layer is allowed to do nothing.

Why it exists

Simon is where the studio’s fail-loudly principle gets tested hardest. A system that reports its real performance, including the periods it was wrong and the periods it was uncertain, is uncomfortable to look at and considerably more useful than one that reports only its wins.

That discipline carries into the other ventures. Caddie tells an instructor when it cannot assess a criterion for the same reason Simon tells you when it does not know.

Where it stands

Simon is in active development and is not available to anyone outside the studio. Its performance reporting is the part we will publish first, because a system that claims to report honestly should be judged on the report.

If you want to talk about the approach, email hello@royalcoastlabs.com.

Simon is a research system. It does not provide investment advice, and nothing here is a recommendation to buy or sell anything.