There is a regression model in there for sure, but it’s as far as I’m aware not fundamentally built upon just a regression model, but actually on an equation system of biochemistry, because we know exactly how much ATP can be produced anaerobically with glycolysis and aerobically with oxygen etc. Exactly how it works I don’t know, and the regression model is in there for sure to match the critical power tests probably to energy produced via different systems, and then the equations can be solved. But yes, this is INSCYD’s proprietary model. And 98% (97.5% to be exact) is also correct from what Sebastian has said.
Also, spot on about it needing to be an independent “true” FTP.