Portfolio optimization
A backtest above the CAC 40 is not enough: costs, stability, sensitivity and uncertainty change the decision.
Read the studyFive complex systems. One discipline.
Welcome to the Applied Assessment Notes.
Here, we study models, protocols and algorithms that look convincing at first sight. We reproduce how they work, test their assumptions, explain what breaks and propose a stronger alternative when the initial approach does not hold.
Organizations are reconstructed. Calculations, simulations and charts are real, dated and reproducible.
5studies
33computed charts
FR / ENbilingual reading
PASSreproducibility check
Flagship study
Industrial compute capacity
Estimating the future cost of compute capacity, defining what could actually be traded, and showing why a credible market requires telemetry, forecasting and verifiable delivery.
Decision question
How can a synthetic cost curve become a deliverable compute-capacity market without conflating forecasting, valuation and physical availability?
+7.49%generation B convexity gap
0.0508%simulation error
8.82%maximum option gap
The five assessments
Results that survive the narrative
Finance, infrastructure, microstructure and biology—connected by one way of checking.
Protocol R7
A cross-chain bridge works under normal conditions. The real question is what it loses when its assumptions stop holding.
Read the study
High-frequency trading
An order whose price is touched is not necessarily filled. Queue, delay, fees and adverse selection erase the apparent edge.
Read the study
HELIX-96
An AI score ranks thousands of mutations. The question is not only what ranks first, but whether its probabilities travel to a new protein.
Read the studyThe spine
One method, not five promises
The domain changes. The question does not: what do we truly know, what breaks, and which decision survives the limitations?
Understand the method- 01
Reproduce
Make every assumption observable.
- 02
Stress
Restore costs, tails, latency and uncertainty.
- 03
Correct
Change the mechanism, not only the narrative.
- 04
Decide
Conclude as clearly with rejection as with approval.
Study framework
Readable like an investigation. Verifiable like an experiment. Honest enough to conclude NO-GO.
Study framework