In a first past the post system the party leading in the polls may not be the one that winds up with the most seats.
There are provincial elections due in both provinces in early October this year, on Oct. 6 in Ontario and Oct. 4 in Manitoba. Recent polls suggest both races will be tight but, interpreted at face value, the polls potentially provide misleading guidance about what to expect. In both provinces the Progressive Conservatives currently hold narrow leads in the polls, but in neither province would it be enough to win the most seats.
In Manitoba a December 20, 2010 Probe Research poll gave the opposition PCs are four point lead 42-38 overall but they trail by ten points in the City of Winnipeg. The PCs "waste" thousands of votes in the first past the post system in rural constituencies. A much more efficiently distributed NDP vote leaves the outcome in Manitoba very much in doubt. My seat projection model translates the Probe poll into a narrow NDP victory in terms of seats: NDP - 29, PC - 26, L - 2. However, there would be several close outcomes, enough to leave the final outcome in doubt. The Manitoba PCs have consistently trailed in the City of Winnipeg and that is their greatest obstacle to success. If trends continue they are likely going to be still counting votes in Manitoba when the polls open in Ontario 36 hours later.
A similar scenario applies in in Ontario where there is the distinct possibility that the election will produce a minority government. A recent Ipsos poll puts the McGuinty Liberals just three points behind the opposition PCs who have 38% to the Liberals' 35% with the NDP at 17% and the Greens at 9%. However, TC's model suggests this would produce a virtual dead heat with the Liberals having 45 seats to the PCs' 44 with the NDP at 18, and with razor thin margins in several seats. Clearly, both races are too close to call and it is not obvious from current polls which party will have the most seats when the votes are counted..