The numbers are at withouthotair, but often wrong
Prof MacKay is a very talented prose writer, a good academic in the field of information theory (lots there to keep our neural-network fans entertained), and a vociferous nuclear lobbyist.
However, he's no energy strategist, and many of his numbers are wrong.
If you want the real numbers, go to the peer-reviewed energy literature. Yes, it's harder work, because few writers are as fluent as Professor MacKay, and the papers aren't always available at zero cost unless you've got a Shibboleth or Athens login, but at least you'll be looking at some more reliable numbers.
Rob Gross & Phil Heptonstall on wind & intermittency. Graham Sinden on how rarely the country has no wind turbines generating at all, or on how much burning of fossil fuels wind turbines save. Michael Grubb on wind and the amount of fossil capacity displaced. And so on.
Or ask the Finns about their £3m/MW (and rising) construction costs for their new nuclear with the cracked welds. Have a look at Chernobyl and ask what would have happened if, instead of it being in a predominantly uninhabited wasteland, it had been in Britain. Pick a spot on a map of England in Wales, draw a 150km radius circle around it, and think about the consequences of rendering that area a deathzone.
No wonder the nuclear industry won't build new nukes, and won't decommission old ones, unless we the public agree to pay for any and all costs of accidents or disasters. It's quite impossible for nuclear to be competitive, once all costs are included, because meltdown is uninsurable.