What are the statistics on governments that lose a vote of confidence and then win the subsequent election? The chances don't seem great if Joe Clark and Paul Martin were any indication. Maybe the Liberals can at least hope that history is on their side?
If I remember correctly, the Conservatives were still polling behind the Paul Martin Liberals when the government fell in 2006.
The important factor is
why they lost the vote of no-confidence. It's usually because of unpopular decisions, which should already be reflected in polling. As such, polling is almost certainly a better predictor of results than the means of a government falling, exceptions like 2006 not withstanding.