MPs and senators returning to federal parliament could hardly have had a better device to focus their minds than yesterday morning’s Newspoll, showing Kevin Rudd’s ratings still on the rise and his party’s vote firmly settled in landslide territory.

It seems like only yesterday that the prime minister was warning his troops against overconfidence. Now he’s warning them against panic.

But a common response from Coalition supporters is that they can imagine a substantial swing to Labor, but can’t see where the seats needed to win are going to come from.

If Newspoll is right, there’ll be no doubt. An eight or nine per cent swing will sweep all before it and produce a huge Labor majority. But with a lesser swing, it’s possible Rudd could win the votes but lose the election, just as Kim Beazley did in 1998.

Labor needs to pick up 15 seats (14 for a hung parliament). The first 15 seats on the pendulum are a diverse lot: some in the mortgage belt, some in the inner suburbs, some in the bush. Five are in NSW, three in South Australia, two each in Queensland, WA and Tasmania, and one in the Northern Territory.

So expect a lot more attention in the coming months on just where the marginals are and how both sides are targeting their campaigns there.

Which states will most receptive to Rudd’s message? Which areas will swing more than average, and which will defy the trend?

We know that the pendulum will get pretty close to the aggregate result, but in a close election those one or two departures from uniformity could be the difference between victory and defeat.