....with a 42.5 kg load at 29 km/h. ...but I suspect you’re heavier than 42.5 kg and have two tyres. Multiply the power difference accordingly.
I don't think a weight multiplication is necessary is it? As you say he has 2 tyres, and his load is split between them, assuming 50/50 front/rear (bad assumption I know...) that's a 85kg rider, so the figures stand as they are. if you're massively more or less than 85kg then obviously need to adjust, but not by much really.
But back to the original point, tyres do make a big difference, as you've illustrated there even two 'quick' tyres have a difference of over 4W per tyre, so 8-10W in total. Cheaper/slower tyres can easily reach double that again.
It would be good to get some actual figures of mudguard impacts at different speeds, yaw angles, mudguard widths and clearances from tyre so we could know if the impact they have is in the fractions of a W, single digit W, or double digit W areas.
It's already been proven numerous times that rider position has the biggest effect, followed by clothing and tyres (order dependant on specifics), so even if performance of the rider was a constant (we know it isn't), unless the effect from mudguards is really quite large, it could easily be that all the extra effort people attribute to them is actually 'other things'.
I've been hunting but not found much other than the limited and not particularly rigorous BQ tests...does anyone have any links to some actual decent testing of aero effect of mudguards, with test descriptions and figures? All I can find is anecdata, phrases like "it's obvious", "clearly" and "it's well know that", and dubious/biased correlation with no implied causation. And as we well know many 'obvious' and 'well known' things are demonstrably false when actually investigated.
I want evidence dammit!