What survived the audit

2026-09-21 · Natkal Research

Track record · 21 September 2026

What survived the audit

We spent a week deleting our own best numbers. This is what was left standing — including the parts that are not working.

First, what we removed

The flattering numbers went first.

Every realized return we published had been measured against a stale entry price — a cached snapshot rather than the price the market actually printed. One call was reported as +50.78% when the real move was +21.96%. We rescored the book.

Then we found 27 verdicts produced by a council run in which no seats returned at all. The system had published a confident neutral call built on nothing. Twelve of those had reached the scoreboard; they are excluded now, and the code refuses to publish a verdict with no seats behind it.

Then 72 more, where the published call contradicted the council’s own distribution — a decisive signal sitting on the smaller of two opposing peaks, or a direction declared on 9% agreement. Those are marked contested, which is what they always were.

None of it was deleted. Every corrected record keeps its original verdict, and every call that was really published stays on the scoreboard — including the ones that were wrong. Only the calls that no council ever made were removed, because there was nothing there to be right or wrong about.

What is left

Conviction compounds with time.

Measured at 21 days and beyond, because below that every tier we run is indistinguishable from a coin flip and we would rather say so than quote a number that flatters us.

Horizon	Calls scored	Directionally right
T+30	803	58.7%
T+60	884	62.9%
T+90	505	70.3%

That is the top tier, and the shape is the point: it gets better as the horizon lengthens. An eleven-seat adversarial council is slow, expensive and wrong about next week. It is built to be right about next quarter, and the ledger now says it is.

70.3%top tier, T+90, across 505 scored calls

27,477calls scored at T+21 or longer

54.0%blended across every tier, including the ones failing

What is not working

The blended number is 54%, and we are not hiding why.

Averaging everything we run gives 54.0%. The gap between that and 70.3% is not a rounding detail — it is tiers that are underperforming, and we would rather name them than bury them in an average.

Tier	Horizon	Calls	Right
Mid tier	T+30	467	48.8%
Broad screen	T+60	4,347	45.4%
Experimental	T+30	120	35.8%

The experimental tier is worse than a coin flip and is labelled experimental for exactly that reason. The broad screen decays badly past a month. Both are being reworked. Neither is what a subscriber is paying for, and neither is quietly folded into a headline figure.

Why publish this

A record that only improves is not a record.

Every number above got worse this month, because we went looking for reasons our own scoreboard might be lying to us and found three. The 70.3% is what remained after we stopped grading ourselves generously.

You can check it. Every call carries the date it was issued, the price it was issued at, the distribution the council produced, and what the market subsequently did. That is the product: not a confident paragraph, but a position you can audit and a ledger that counts the misses.

Figures re-derived from the live ledger on 21 September 2026, minimum horizon 21 days, sample sizes shown. Past performance is not indicative of future results. Nothing here is investment advice.