Labor goes into tomorrow’s Victorian election as a barely backable favourite, with offerings on a Coalition win ranging from $4.25 from Sportsbet to $4.65 from Flemington Sportsbet.

However, the once-popular notion that betting markets offer special insights into election outcomes has taken a bit of a knock of late.

On the eve of the Western Australian election in September 2008, bookies were offering between $3.45 and $3.75 on the Coalition victory, which duly transpired.

Late odds at the federal election were also better for Labor than the situation warranted, with Centrebet offering $1.52 on Labor and $2.44 on the Coalition.

The wisdom of crowds proved no sharper immediately after the inconclusive election result, when there was $3 to be had on Julia Gillard retaining office against $1.33 for Tony Abbott.

In the Victorian case there is also the cautionary precedent of the 1999 election, when victory for Jeff Kennett’s Coalition was so widely taken for granted that the Sunday Herald-Sun didn’t wait for a result before reporting it as fact.

A range of parallels between the current situation and 1999 were noted yesterday by Paul Austin of The Age: a Premier perceived as arrogant, an Opposition needing 13 seats to win, a feeling that this was a few more than even party optimists thought possible, and the latent possibility that a regional backlash could nonetheless make it happen.

If today’s 50-50 Galaxy poll is accurate, there could be further added a late hardening of anti-government sentiment that the polls detected (Newspoll and Morgan correctly indicated a dead heat in 1999, although ACNielsen erred heavily in favour of the Coalition) but pundits proved too conservative to take on board.

Labor would have particular cause to worry that John Brumby’s strong leads over Ted Baillieu on a range of measures (preferred Premier, most trusted to keep promises and better campaigner) have translated into a paltry Galaxy primary vote of 36%.

This might be seen to indicate that Labor will indeed face what a party source, quoted by John Ferguson of the Herald Sun, describes as a “f-ck you vote”, in which wavering voters conclude that whatever the merits of the two parties, “the government’s had enough time”.

For all that, any prediction of a Coalition victory would have to be rated a brave call.

The electoral playing field in Victoria is tilted in Labor’s favour: assuming a perfectly uniform swing, the Coalition would need 52% of the two-party vote to form government either in its own right or with the backing of rural independents, and no poll conducted in the past term has put them anywhere near this.

The obvious objection here is that swings are never uniform, and the 1999 election is a case in point. Whereas the Liberal vote largely held up in Melbourne, where the swing was under 3%, a regional swing approaching 6% resulted in enough unexpected defeats to cost the Coalition its majority.

Significantly, three conservative heartland seats were won by independents who ultimately threw their support behind Labor, just as Tony Windsor and Rob Oakeshott would do at the federal level a decade later.

While Labor is facing comparable threats from Greens and independents in former inner-city strongholds, it’s a lot harder to see them crossing the ideological divide to back a change of government (excluding a scenario where the Coalition wins 44 of the 88 seats, in which case they will have little choice).

That leaves the Coalition with the daunting task of picking off 12 seats directly at Labor’s expense.

The most efficient scenario for a Coalition win would involve a 4% swing across Melbourne, delivering eight seats (nine if you count South Barwon in Geelong), and a regional backlash of approaching 7% to serve up another four or five.

That would still require a statewide two-party preferred vote approaching 51%, unless they could pick off a few upsets further up the pendulum — the most widely canvassed being Bellarine and Yan Yean, both on 7.9%.

That being so, the Galaxy poll offers encouragement for the Coalition, but probably not enough.

Final poll results from Morgan, Nielsen and Newspoll, probably in that order, will follow later today.