I've just noticed the bar charts which seem to give a total incidence per million of 2-4 million - see Figure S5 on page 32.
Do they really mean that everyone is going to catch it 3 times? Really? With hardly any effect from who gets boosted?
Of course they're not actually simulating the UK, just a country which is the UK and which is being used to make predictions about what might happen in the UK, but if their model for what is happening right now is obviously wrong then I don't really see why their prediction can suddenly become accurate about the future.
Even if the UK currently has about 1000 deaths per week, that's about 2 per million per day. You would barely even see that on the scale they're using in Figure 2, which goes up to 150, but by eye they seem to be saying there are 10 per million per day
now and will peak at 50 per million per day. From Figure S5, this seems to correspond to a case rate of about 5000/million/day (3500/100,000/week) and I really wonder if you already had 330,000 cases per day earlier this season
Figure 2B does suggest that by boosting the oldest age group, the total deaths would only be about 300,000 instead of ~430,000.