I am puzzled over the "stress test" relating to "five cold windless days" and the conclusion that lots of backup needs to be built especially for wind. Given that few people expect wind to have more than a very modest capacity credit, is it really necessary? Wind is a fuel saver, rather than a "plant saver" and so, if it is missing on cold winter days, does it matter? – nobody will be relying on it. As the model looks at average electricity needs, won't 20 GW of OGT backup be needed in a no-wind scenario? (20 GW is the about the difference between peak demand and average demand). Puzzingly the "maximum demand" scenario, which has no wind, suggests 7 GW of backup is needed.
The concept of "dedicated backup" is novel. There is no dedicated backup for nuclear, but the electricity system survived during the winter of 2008/9, when half the nuclear output (about 5 GW) was missing for a long period – much longer than five days. The deficit was made up from the Plant Margin (old definition – thermal plant capacity minus expected peak demand). The plant margin under that definition needs to be about 24%. The statistical analysis behind this figure ensures that the system does not suffer undue stress. If wind is present during the peak demand periods, it may improve the system reliability; if it's not, it doesn't matter.
I am concerned that the Pathways methodology seems to give credence to the notion that, "if you build wind, you need to build extra backup." That is going to distort the economics. You do not need to build backup for cold, windless days, but only to cover the additional short-term uncertainties. However, there does not seem to be any discussion of that issue.