AlphaFrame Lab

Our ETF allocation models, told properly: what they are, how we put them on trial and what survived the tests. The honesty pact: each engine's asset-class universe and character, always; the recipe (weights, positions, training details), never. Every figure is net of the Italian 26% tax and trading costs.

Ep 1The engines: the model fleet, universe and character

Ep 2The method: the distributional gate and the double bar

Ep 3Rigorous backtesting: why backtests lie

Ep 4Trained deep learning vs generative AI

Ep 5Performance and regimes: who wins (and defends) where

Ep 6The 26% factor: the tax drag that decides

Ep 7The alpha hunt: what we killed (and why)

Ep 8What survived: dated verdicts and honest value

Marketing communication for informational/educational purposes — not personalised advice, a recommendation or solicitation, and no return promise; capital is at risk. The investment advice service has not yet commenced and no portfolio management is provided: the models described are proprietary quantitative research, not a product or subscribable service. Figures quoted in episodes are in-sample/backtest where stated; past or simulated performance is not indicative of future results. Universe described by asset class/sector: no specific securities, no current allocations.