Forecasting is hard
Page 106, thanks to Ed Conway 
I'll admit, I have an idiosyncratic sense of humour. But still, I laughed out loud at this tangle of lines, which shows the OBR's best attempts to forecast oil and gas revenues. It's reminiscent of the woefully optimistic IMF forecasts for Greek GDP , excel that instead of being consistently wrong in the same direction, it's more like a child just scribbled a lot of lines on the chart.
Unfortunately, the oil and gas revenues remain important. Thanks to the long-standing decline in productivity in the sector, a function of the drying-up of North Sea oil fields, it usually imparts a massive downward pressure on the quarterly GDP figures, which means that getting the predictions accurate is crucial for getting the overall figure accurate.
Migration saves us money
Page 147, thanks to Jonathan Portes 
If you care about public sector debt, really the absolute best thing you can do is remove restrictions on migration. Migrants are educated by their home country, and frequently retire there too; in the meantime, they work hard, pay their taxes, and have a lower-than-average crime rate.
The "high migration" scenario is of the average net migration being slightly more than double what the ONS uses as its baseline assumption, with 260,000 people coming in on net compared to 140,000. That's a lot more than normal, but it's not outside the realm of political possibility. Just think what a fully open-borders policy  could do for the national accounts…
At the other end, the ONS looks at what "zero net migration" would do. Remember that zero net migration is actually the government's explicit policy, so it's already a bit damning that the ONS instead works on the assumption that they will fail to hit it by 140,000 people. But when we look at the stats, it's clear that we should be glad of that. Zero net migration would push the debt:GDP ratio over 100 per cent by 2050.
Young people and old people cost money
Page 78, thanks to Chris Giles
Again, nothing which will blow your mind: the state spends money educating young people, caring for old people, and providing health services to both, while the people in the middle pay the bills. What's interesting are the two crossover points – roughly 23 and 67 years old – where people go from being, on average, a contributor to a benefactor or vice versa, as well as the curious level of the peak of tax contributions, at just under 50.
You are never going to retire
The thick line is the OBR's best guess of what changes to the pension age are going to do to the proportion of people between 65 and 74 working: around a 66 per cent increase, to just over a quarter of those people working by 2045. That already comes after a doubling of the rate in the last twenty years:
We are never ever ever getting time off work.
This is all just guesswork
Finally, an important reminder that the long-term projections are as vague as can be. In fact, discussing them in terms of fiscal policy is almost nonsensical. What they are instead is predictions of demographic change mapped on to current policy. So if the nation continues ageing as it looks like it will be, and if we fail to do reform the state pension in that time, then the national debt will start rising on current policies in 2037.
Obviously, it's nonsense to act as though all our policies will be the same in 2017, let alone 20 years after that, but it's the only way talk about the future at all.