Over the past 60 years, I have become the most frequently quoted forecaster of Australian election results. My strategy has been to make a forecast long out from polling day and then stick to it - except in extreme circumstances where my reputation for good political judgment required a change of prediction.

Published on this website on Thursday 20 September was an article by me under the heading “How will Australia’s wealthiest electorate vote?”. In that article, I made a probability statement on Wentworth which was to give the Liberal candidate Dave Sharma a 55% chance, the Independent Kerryn Phelps a 35% chance and the Labor candidate Tim Murray a 10% chance. The rest of the article justified that thinking by arguing that the order of votes after preference distribution would place Sharma first, Murray second and Phelps third. Since one cannot win from third place, I predicted a Sharma win.

The essence of my reasoning was that Labor has a good candidate and, for a variety of reasons, could not “run dead”. As I now see it that was my first error. Labor’s campaign has proved that it can “run dead”. Every piece of evidence available to observers is that Labor is “running dead”. Their reasoning for such a strategy is that Labor cannot win the seat but Phelps can. Therefore, Labor can inflict the most damage on Scott Morrison by coming third and hoping for a highly disciplined transfer of preferences from Murray to Phelps.

There are 16 candidates but the real question is this: “how many significant candidates are there?” My original answer was three, Sharma, Murray and Phelps. The course of the campaign, however, tells me that the answer now is two, Sharma and Phelps.

Recently the Sydney Morning Herald arranged for a local debate between five candidates, the five being those deemed by the SMH to be significant. In ballot paper order, the five SMH significant candidates are Dominick Kanak of the Greens, Murray, Sharma, Licia Heath, another Independent, and Phelps.

My guess is that there will be about 300 “donkey votes”, they being those votes which go straight down the ballot paper. For the following “donkey vote”, I am not wasting space by giving the party of the candidates deemed by no one to be significant – so here goes.

My “donkey vote” is 1. Robert Callanan. 2. Kanak. 3. Shane Higson. 4. Steven Georgantis. 5. Murray. 6. Ben Forsyth. 7. Tony Robinson. 8. Samuel Gunning. 9. Sharma. 10. Angela Vithoulkas. 11. Deb Doyle. 12. Andrea Leong. 13. Heath. 14. Barry Keldoulis. 15. Phelps. and 16. Kay Dunne.

The effect of such a “donkey vote” is that Murray gets that vote but when Murray is excluded (as I now think that’s highly probable), it is then classified as “leaking” to Sharma, since he is placed ninth and Phelps fifteenth.

The big question, therefore, is this: “How much Labor leakage will there be?” The answer I feel now compelled to give is that, say, one fifth of Labor preferences will “leak” to Sharma. For that reason I have changed my probability statement. I now give Phelps a 55% chance of being the next member for Wentworth, with a 45 % chance for Sharma.

Malcolm Mackerras is Honorary Fellow of Australian Catholic University. malcolm.mackerras@acu.edu.au