As we all know by now, the ALP is sixteen seats short of a House of Representatives majority, and you can plot those sixteen most marginal government-held seats along the pendulum.

Of course, it doesn’t really work that way. No-one expects the seats to fall in the order they appear on the Mackerras horseshoe. But how might they fall? I’ve made my own list of the first seats to fall, or “the seats to watch”.

I’ve started with current margins and performed a few twiddles to reflect my version of reality. For example, everyone would agree that Malcolm Turnbull’s on paper 2.5 percent margin overstates his vulnerability because of Peter King’s 2004 candidacy. So I gave Malcolm a couple of percent breathing space.

Seats in which members are retiring lose a percent to account for loss of personal vote – except for Makin, because the retiring member Trish Draper’s rorting problems contributed to the 2004 pro-Labor swing, which was the biggest in South Australia. Makin’s margin is already deflated.

Seats that changed hands in 2004 get an extra percent to account for the personal vote of the new incumbent. If it was a sitting member they defeated they get another point because Labor has lost that old personal vote.

Another exception: Labor’s Julie Owens gets the one percent for her own personal vote, but not for the loss of defeated member Ross Cameron. Like Draper, he produced much of the 2004 swing.

And so on. I factor in a greater than average swing to Labor in Queensland because they all love Kevin Rudd up there, and a lower than average one over in Western Australia because they’re all rolling in money these days. (Stop me if I start generalising.)

Unlike most observers, I reckon the classic mortgage belt demographic will swing to Labor less than average, rather than more, so I’ve factored in that, plus a bunch of other vague assumptions, to come up with this list of seats to watch.

The 16 closest seats on pendulum

My key 17 seats to watch

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

Kingston (SA) 0.1

Bonner (Qld) 0.5

Wakefield (SA) 0.7

Parramatta (NSW) 0.9

Makin (SA) 0.9

Braddon (Tas) 1.1

Hasluck (WA) 1.8

Stirling (WA) 2.0

Wentworth (NSW) 2.5

Bass (Tas) 2.6

Moreton (Qld) 2.8

Solomon (NT) 2.8

Lindsay (NSW) 2.9

Eden-Monaro (NSW) 3.3

Bennelong (NSW) 4.2

Dobell (NSW) 4.8

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

Bonner (Qld) 0.5

Parramatta (NSW) 0.9

Wakefield (SA) 0.7

Cowan (WA) 0.8

Moreton (Qld) 2.8

Makin (SA) 0.9

Braddon (Tas) 1.1

Eden-Monaro (NSW) 3.3

Page (NSW) 5.5

Lindsay (NSW) 2.9

Kingston (SA) 0.1

Solomon (NT) 2.8

Wentworth (NSW) 2.5

Cowper (NSW) 6.7

Stirling (WA) 2.0

Corangamite (Vic) 5.3

Bass (Tas) 2.6

My list has seventeen seats instead of sixteen, because my manipulations on the WA seat of Cowan – including the retirement of a popular sitting member – moved it to the government side of the ledger.

The order of my list is also very different to the pendulum’s. Missing from it also are Bennelong (NSW), Dobell (NSW) and Hasluck (WA). I don’t think they will be in the first 16 to fall.

In their place are three regional seats Corangamite, in Victoria and the NSW North Coast National seats of Cowper and Page.

You might like to try this at home. After the election we can see how close our lists were.