Ontario

Here is what Orbit, our experimental model, sees next to Harfang des neiges, our classical model.

Orbit’s favourite

69% chance of finishing first

Seats won

58Harfang 68 8 in 10 chance between 35 and 80

Chance of a majority (63)

42%Harfang 82% PC

Same favourite as Harfang des neiges

92% of ridings

Orbit (quantum) vs. Harfang des neiges (classical): do they agree?

Same winner, different call on the majority

Orbit (quantum)

Harfang (classical)

50/50 blend

Majority: 63

What if we listened to both engines at once?

50/50 blend: PC 67 · Chance of a majority62%

The blend gives each riding the average of the chances announced by Orbit and by our reference forecast, then counts the favourites. In our Quebec replays (2018 and 2022, the day before the vote), the blend always beat the reference forecast alone: in 2018 the error drops from 6.5 to 4.5 seats, and the announced probabilities become far more reliable. The 50/50 split is a choice of principle, not a setting tuned on those two elections. Its first real blind test came in Quebec on October 5: the blend picked the right winner in 114 of 127 ridings, against 110 for the reference forecast and 113 for Orbit, and its probabilities were the best calibrated of the three. One election is not enough to conclude: British Columbia is its second test.

Party by party

PC −10 Harfang

Vote: 38.2%
Vote Harfang: 39.2%
Chance of a majority: 42%

OLP +10 Harfang

Vote: 33.0%
Vote Harfang: 30.4%
Chance of a majority: 6%

NDP = Harfang

Vote: 22.2%
Vote Harfang: 21.5%
Chance of a majority: < 1%

GP = Harfang

Vote: 3.3%
Vote Harfang: 4.4%
Chance of a majority: < 1%

OTH = Harfang

Vote: 3.4%
Vote Harfang: 4.6%
Chance of a majority: < 1%

The vote, with its margin of uncertainty

The solid bar is Orbit’s projected vote; the thin line, an 8-in-10 range; the black tick, our reference forecast.

PC 38.2%
OLP 33.0%
NDP 22.2%
OTH 3.4%
GP 3.3%

Where the two engines disagree

Cambridge Orbit (quantum): OLP · 50% Harfang (classical): PC
Hamilton Mountain Orbit (quantum): OLP · 40% Harfang (classical): PC
Milton Orbit (quantum): OLP · 58% Harfang (classical): PC
Mississauga-Centre Orbit (quantum): OLP · 59% Harfang (classical): PC
Mississauga—Streetsville Orbit (quantum): OLP · 62% Harfang (classical): PC
Newmarket—Aurora Orbit (quantum): OLP · 58% Harfang (classical): PC
Oakville-Nord—Burlington Orbit (quantum): OLP · 51% Harfang (classical): PC
Peterborough—Kawartha Orbit (quantum): OLP · 58% Harfang (classical): PC
Pickering—Uxbridge Orbit (quantum): OLP · 51% Harfang (classical): PC
Whitby Orbit (quantum): OLP · 51% Harfang (classical): PC

Orbit’s instruments

Voters who switched sides since the last election

6.7 at least, in points of the electorate

Average gap between pollsters

± 2.6 points, around the most likely state

Horizon

Election today No election date set: forecast if the election were held today, like our reference forecast.

How Orbit works, in three ideas

Opinion travels on a sphere

A 2-point gain does not mean the same thing for a party at 5% and a party at 40%. By placing opinion on a curved sphere, Orbit knows this by instinct, and it carries the national swing to each riding the way you carry an arrow across a globe.

Polls are measurements

As in quantum physics, each poll is a measurement of a hidden state: the electorate. Orbit tracks that state day after day, erases each pollster’s own lean and lets no one dictate the average alone.

Qubits instead of dice

Each riding becomes a small quantum register. Instead of rolling dice 50,000 times, Orbit counts the seats exactly, in seconds. A real quantum circuit gives the same answer, to the billionth.

Read the full explanation