I think I’ve finally unraveled what Sallet is talking about, though the media reports sure make it very confusing. The 7.04 W/kg value is apparently Froome’s power at V02max, and based on 500 watts at 71 kg. Someone upthread said it was peak power, but that is not quite correct; it is peak aerobic power. The figure 425 watts also appears in media stories. Apparently Sallet believes—on the basis of some mathematical model—that Froome’s utilization is 85% (425/500), that is, he can put out 85% of his V02max for a sustained climb. Assuming 425 W and 71 kg, his sustained output is 5.98 W/kg, which Sallet was initially reported to have found.
In the podcast he says 408 W, which further confuses the issue. However, here he means sustained power, not maximum power, and he's also correcting for weight. His original calculations assumed a weight of 71 kg, but after Kerrison provided the 67.5 kg value, Sallet uses this for the weight. Thus he gets 500 W x .85 x (67.5/71) = 404 W (close enough to 408).
So there have two sources of confusion, first and most important, Sallet is talking about Wmax, which is not normally the figure one determines from climbs or from power meters; I don't think the other guy, Pickering, on that podcast understood this, nor have most other people discussing this. This is a better value to use if you know it, as it combines two of the three key parameters needed, V02max and utilization, but how Sallet thinks he knows it I don't know, apparently he thinks he can get it from his model. And second, his original W value (though not W/kg value) was changed when he got a different weight value from Kerrison.
In any case, if you use Sallet's figures, you find that:
V02max x GE = 20.2
Not knowing either of those two values, one can’t solve for the other, but one can solve for one by making assumptions about the other. If we assume V02max is 90, which of course is extraordinarily high, then GE = 22.5. That is fairly high, but certainly not inconceivable. The highest power/weight value in the study of GT elite riders by Santalla et al that reported an inverse relationship was for an individual with a V02max of 80, which would require a GE of 25.5. Though GE values were not determined in this study, from the DE values it appears this rider probably exceeded that. Probably no other rider in the study did, though. The highest power value in the Lucia study was for a rider with a V02max of 70 and an efficiency of 28. The product of those numbers is 19.6, a little lower than the 20.2 calculated above. So Froome’s power, according to Sallet, is about as high as that of the highest of 12 elite riders in one study, and higher than that of any of 11 elite riders in another study.
Interestingly, though, Sallet’s own work reported very high efficiencies (mean of 25.6%) with no mention of an inverse relationship. Using that mean, Froome would need a V02max of just 79 to put out the power calculated by Sallet. I’m a little surprised that in his interviews he emphasizes the likelihood of cheating when his own work reports efficiencies that could account for such a high power. However, maybe he thinks he has some information bearing on Froome's efficiency as well. There is clearly much about the model that has not been reported.
According to the podcast, Sallet is going to furnish more data as rebuttal to Sky/Kerrison. Stay tuned.