I start this article by quoting the sixth, seventh and ninth paragraphs of that article.

''For two years now I have consistently been making four confident predictions. First, the next election will be held on October 26, 2013. Second, there will be no by-election during this, the 43rd Parliament. Third, Julia Gillard will lead Labor into the 2013 election. Fourth, Tony Abbott will lead the Coalition into that election.

''As to who will win, I have never made any actual prediction, merely varying probability statements. My present statement is a 90 per cent chance for Abbott and a 10 per cent chance for Gillard.

''Of my four predictions, the only one to remain contentious is the election date.''

In that article I gave good reasons as to why October 26 would be a suitable election date.

Anyway, it was always obvious that the election would be held in either September or October and I did not have the space to go through all the dates.

Suffice it to say that I ruled out September 14 on the grounds that this is the Day of Atonement for Jews, otherwise known as Yom Kippur, the holiest day of the Jewish calendar.

If Gillard had told me she was determined to go to the polls in September and asked me to suggest a date, I would have advised September 7.

There is a significance to my advised date, it being that Tuesday, September 7, 2010, was the day when it became clear (courtesy of Tony Windsor and Rob Oakeshott) that Gillard could command the confidence of the House of Representatives in the 43rd Parliament.

I say that, whereas our most recent election was held on Saturday, August 21, Gillard won that election on Tuesday, September 7.

The essential difference between the October of Malcolm Mackerras and the September of Julia Gillard is that I was thinking within the square whereas she was thinking outside the square.

It will make no difference in the long run but this much can be said for her decision. It makes it impossible for the Coalition to argue that she is postponing the election date to the last possible day.

Much of the commentary on her choice was ridiculous. But I nominate the article by Graham Richardson in The Australian as the silliest.

Richardson is one of those people who thinks it is a great advantage to a prime minister to be able to pick the election date. To him, therefore, it was unforgivable of Gillard to throw away that advantage by treating the 43rd Parliament as though it had a fixed term.

There are some circumstances when it is an advantage, but they are rare. Certainly this year does not provide such a case.

In my capacity as a volunteer guide at Old Parliament House I have written an article titled Early General Elections for Australia's House of Representatives. If any reader wants a copy and cannot find it on the OPH website I can send it to you by ''snail mail''.

Having examined every federal election since Federation, I ascertained that there were only four cases where a House of Representatives plus half-Senate election (following the previous election of the same type) could be said to have conferred an advantage on the prime minister.

The elections in question were those of December 1919 (won by Billy Hughes), September 1934 (Joe Lyons), October 1998 (John Howard) and August 2010 (Gillard).

The previous House of Representatives plus half-Senate elections were in May 1917, December 1931, March 1996 and November 2007.

The point is that the fixed terms for senators impose a limit on the ability of the prime minister to go early. In 2010, Gillard stretched that ability to the maximum. Consequently she lost all subsequent advantage by choosing August as the month for the 2010 election.

Since there was no advantage left to Gillard in 2013, it was sensible for her to make a virtue of necessity, announce early and avoid media speculation on this subject.

Perhaps it may be said Gillard has falsified the prediction I was making on the election date. If so, readers might well ask: do I stick to all my other predictions? Yes, I do.

Clearly Abbott will be prime minister come September. The question is: how successful will he be?

As a piece of history it is true of the period since Canberra became the national capital in 1927 that every conservative elected prime minister has come into office on a landslide.

Joe Lyons in December 1931, Bob Menzies in December 1949, Malcolm Fraser in December 1975 and John Howard in March 1996 all swept to office in landslides and remained prime minister for more than seven years, winning three or more consecutive elections.

I do not expect Abbott to have a landslide win. Not that it matters. Even if the landslide-prone electoral system for the House of Representatives gives him a large majority in seats in the lower house he will have a disaster in the Senate election.

That is where Abbott is sure to differ from Lyons, Menzies, Fraser and Howard, each of whom scored well in the Senate election at his first win.

Consequently I predict Abbott will be the least successful conservative elected prime minister since Canberra became the capital.

Abbott's main mission seems to be to tear down Labor's achievements. He cannot do that without a Senate majority. Consequently, I have a further prediction to make. The 44th Parliament will be the second shortest in our history, double-dissolved before even a year in being.

The shortest parliament so far was the 11th. It was elected on November 17, 1928, and first met on February 6, 1929. It was dissolved on September 16, 1929, and the next election was for the lower house only, on October 12, 1929.

Since the length of our Federal Parliament is measured from the first day of sitting, the length of the 11th Parliament was seven months and eleven days.

The 44th Parliament will be elected on September 14 and will first meet on October 21. Readers can work it out for themselves. It will not run for a full year. A spring general election next year for all members of both houses is a very good bet.

Malcolm Mackerras is Visiting Fellow at the Public Policy Institute, Australian Catholic University, Canberra Campus.

malcolm.mackerras@acu.edu.au