It's Academic

I know this is only a "stylized mathematical model" of how Windows and GNU/Linux interact in the marketplace, but it's more akin to a Swiss cheese model, so many holes does it exhibit. For example:

The model captures what we believe are the most important features of the Linux-Windows competitive battle (faster demand-side learning on the part of Linux and an initial installed base advantage for Windows), but makes important assumptions regarding other aspects.

"Faster demand-side learning" has almost nothing to do with it these days: issues like control, stability and security are more to the fore.

And then this is a completely erroneous assumption, too:

Our paper introduces a dynamic mixed duopoly model in which a profit-maximizing competitor (Microsoft) interacts with a competitor that prices at zero (Linux), with the installed base affecting their relative values over time.

Nobody equates GNU/Linux with zero price anymore: even if TCO is a slippery concept, it is certainly more realistic than simply looking at the price tag, as this study does.

And so on, and so on. The fundamental problem is that open source is driven by so many complex - and often non-economic - factors that any simplistic mathematical modelling is doomed to fail from the start.