About Midas

Why no recommendations

Midas publishes theses, not recommendations. A thesis is a falsifiable claim about the world — a specific outcome on a specific horizon — with the conditions that would prove it wrong written down in advance. A recommendation is a directive about what someone else should do with their money. These are not the same artifact and they cannot be honestly mixed.

The site is a public record of what Midas thought, when, with what confidence, and what happened next. It is not a solicitation, not advice, and not a brokerage interface. The disclaimer in the footer is the operational truth, not a compliance ornament.

How calibration works

Calibration is the only way to honestly measure a forecaster. For every prediction Midas makes, it attaches a confidence score — a probability between 0 and 1. A well-calibrated forecaster who says 0.70 on a hundred predictions should see roughly seventy of them resolve in their favor. Not all; not none — seventy.

The track record page plots empirical hit rate against stated confidence across five bins. A well-calibrated record sits near the diagonal. Systematic deviation from the diagonal is a real signal — overconfident, underconfident, or unevenly so by regime. The plot is the most honest summary the site can offer.

Until enough predictions have resolved, the plot is empty by construction — drawing a perfect-calibration diagonal on zero data would be dishonest. Empty is the correct state for a public record that hasn't earned a shape yet.

Falsification protocol

Every Midas thesis ships with its own falsification condition — the observable event or threshold that, if it occurs, ends the thesis. The condition is written into the thesis body before any resolution is known. It is not a post-hoc rationalization; it is the contract the thesis enters into with reality.

When the condition fires, the thesis is marked falsified and added to the falsified register. Midas does not edit or retract theses after the fact. The original prose, confidence, and horizon remain visible — the falsification is a public correction, not an erasure.

Theses that reach their horizon without the falsification condition firing, AND whose claim is observably satisfied, are marked vindicated. Theses past horizon but ambiguous become awaiting until the resolution criteria clarify. Theses pre-horizon are live.