How readings are evaluated

Why this page exists

Avenyx publishes one governed market-state reading. A reading is only worth trusting if it can be checked, so this page states — ahead of time — exactly what each reading commits to and how anyone can grade it against the public record. Nothing here is advice, a signal, or a performance claim. We grade whether our descriptions matched the realized environment, never whether acting on them made or lost money.

What we grade, and what we refuse to grade

Avenyx does not know your portfolio, your timing, or your risk tolerance, and it issues no instructions. So we never score readings against profit and loss — that would imply an advisory relationship that does not exist. We score only the accuracy and discipline of the published labels themselves. A reading can be “right” in our terms and still have nothing to say about whether any particular person should have acted.

The four claims every reading makes

Each State Card commits, in advance, to the following checkable claims:

1. Risk-label calibration. Grouped across the whole archive, periods we labelled Risk Today: Elevated should, on average, be followed by rougher realized conditions — higher volatility and deeper drawdowns over the following weeks — than periods we labelled Moderate or Low. Success is the ordering holding across many readings, not any single call.
2. Trigger fidelity. When a reading’s stated “what would change this” condition is met, the posture should move in the stated direction by the next reading — or that next card must explain why it did not. This grades rule-following, not foresight, and it is checkable from the very first reading.
3. Data-quality honesty. When inputs fall below the bar to publish reliably, the reading is marked low or paused rather than dressed up as a normal posture. We grade ourselves on whether paused readings actually line up with genuinely noisy input periods — and on never quietly publishing through them.
4. Change discipline. The environment label should not whipsaw from one reading to the next without a stated, recorded reason. Every change is tied to a logged trigger; unexplained flips count against us.

How scoring is reported

We will not publish a scorecard until there are at least 50 readings or six months of record, whichever comes later, and quarterly after that. Each scorecard groups readings by their labels and reports the realized outcomes plainly, including the misses — not a curated highlight reel. Because this is a single asset over a short horizon, early numbers carry real sample-size uncertainty; we will say so on every scorecard rather than overstate what the record can support.

What would count as Avenyx being wrong

If Elevated periods show no higher realized volatility or drawdown than Low periods across the sample; if posture repeatedly fails to follow its own stated triggers; or if readings are published during conditions we ourselves flagged as low-quality — those are failures, and the scorecard will show them.

How to verify the record yourself

Every reading is timestamped, price-anchored, and hashed. The canonical record of each reading is published with its SHA-256 hash, and each reading page can recompute that hash in your browser — so you can confirm no reading was altered after the fact and re-run every check on this page independently. You do not have to take our scorecard on faith.

What this is not

This is not a backtest of returns, not a forecast, and not investment advice. Strong scores on these descriptive measures say our labels were well-calibrated and consistently applied. They say nothing about whether the market will perform or whether any decision you make will profit. Decision authority remains with you.