Does your forecast beat a fair naive benchmark?

An independent, out-of-sample answer — scored against a naive baseline that's genuinely hard to beat. Even when the answer is no.

A new, solo, independent practice — judge it on method, not on a track record it hasn't earned. This is an indicative statistical analysis for your internal use: not an audit, assurance, actuarial or expert opinion, not investment advice, and not intended for third-party reliance.

PASS
FAIL
INCONCLUSIVE

What it is

The Alpha Edge Test is an independent statistical check of whether a forecasting model beats a fair naive baseline on your own out-of-sample data. We score your model against a demanding, well-chosen naive baseline — random-walk, climatology, and seasonal naive — using standard, published, widely-used scoring rules (CRPS, pinball loss, calibration/PIT, Forecast Value Added). You get a plain-English PASS / FAIL / INCONCLUSIVE verdict and a clear report you can use in your own internal review and share with your team or advisors.

Why independent

We don't sell you a model, a forecast, or an edge. We only test whether the one you have beat a naive benchmark on your out-of-sample history. We have no stake in the result and nothing to sell you but the test itself — the verdict is the product. Beating a fair naive benchmark out-of-sample is genuinely hard, and many forecasts don't — so you deserve an independent, methodical read before you rely on a model.

Our report is one input to your own decision-making, not a recommendation to act.

How it works

  1. Send your model's forecast-vs-actuals history (the data you already have).
  2. We score it out-of-sample, walk-forward, against a demanding naive benchmark.
  3. You get an Alpha Edge Test Report: the verdict, the value-added, calibration, and a plain-English read.

Proper scoring

CRPS, pinball/quantile loss, PIT & tail calibration — not just accuracy theater.

Demanding benchmark

Beating naive is the real bar. We pick a well-justified, hard-to-beat naive baseline, not the easiest.

Out-of-sample

Walk-forward only. In-sample numbers don't count — and we won't quote them.

Honest by design

This is a new, solo, independent practice — judge it on the rigor, not on a track record it hasn't earned yet. It tests for an edge; it does not provide or promise one, and it guarantees no result. In our own founding research, the headline finding was no edge — exactly the kind of honest answer this service exists to deliver. A FAIL is a valid, valuable outcome — often the most valuable one. The method is standard, published statistics applied without flinching; the value is independence and honesty, not a secret formula.

Built by a working nurse, in public. See the experiment it came from → — including the honest "no edge" finding that started it.

Get an independent read

Fixed-scope, one-off analysis on your own data. Tell us the model and the question and we'll scope it.

Contact for an engagement

The Alpha Edge Test provides independent statistical analysis of data you provide, for the client's internal use. It does not provide forecasts, trading signals, investment advice, insurance, or any regulated service; it does not manage money; it is not an audit, assurance engagement, actuarial or expert opinion; and it does not guarantee, promise, or imply any result, profit, or "edge." Reports are indicative analyses and are not intended for third-party reliance. A new, solo, independent practice — no track record is implied. McLaughlin Capital Investments LLC is the operating entity; it is not a registered investment adviser, broker-dealer, or CTA, and provides no investment, securities, insurance, or actuarial services. © McLaughlin Capital Investments LLC.