Mayoral forecast
What could happen
Estimates each candidate’s chance of winning from eligible polls and Toronto’s historical polling record.
Follow the forecast method ↓Methodology
How Toronto 2026 works
This site does three different jobs: it forecasts the mayor’s race, charts the public polling record, and identifies council races worth a closer look. Here is the evidence and reasoning behind each one.
Three views of the election
Polling trends
What pollsters reported
Preserves every public reading and smooths each candidate’s series to make movement over time easier to see.
Read the chart correctly ↓Council attention
Where to look closer
Orders the 25 wards by attention and explains the evidence around incumbents, open seats, candidates, and ward polls.
See how wards are assessed ↓01 · Mayoral forecast
From polls to chances of winning
The forecast builds a plausible full ballot, simulates election day many times, and publishes only the precision that survives its stress tests.
Which polling enters the forecast?
Once the candidate field is certified, the model uses polls that measure that field. It keeps the newest eligible reading from each pollster and gives every represented pollster equal weight. A firm does not gain extra influence simply by publishing more releases.
Polls do not always report every certified candidate. Reported candidates are compared within that poll, while Toronto’s previous elections inform how much support to reserve for candidates the poll left unmeasured. An omitted candidate is never treated as having zero support.
- Choose eligible polls
Use polls of the certified candidate field, keeping the newest eligible reading from each pollster.
- Balance the pollsters
Give each represented pollster equal weight, so publishing more often does not create more influence.
- Account for the full ballot
Use Toronto history to reserve support for certified candidates a poll did not measure. Missing never means zero.
- Model election-day uncertainty
Use past Toronto polling misses to create thousands of plausible full-ballot results.
- Stress-test the result
Repeat the forecast under other reasonable choices and after leaving out individual samples or pollsters.
- Publish a stable band
Show the narrowest plain-language range that survives the required checks; otherwise broaden it or withhold it.
How uncertainty enters
Historical Toronto elections supply two pieces the current polls cannot reveal by themselves: how much final-ballot support polls tend to leave unmeasured, and how far polling can land from the eventual result. The history calibrates uncertainty; it does not dictate who wins this election.
The model then reruns the forecast with more emphasis on the newest sample, lower and higher unmeasured-candidate shares, one sample or pollster removed at a time, and an incumbency-prior sensitivity. The incumbency prior is a check on the result, not part of the main model.
Illustration, not a current result
How a simulated result becomes a public band
SimulationA candidate wins roughly eight of every ten outcomes.
Stress testsRequired variants stay inside the same public range.
Published“About 4 times in 5”
If reasonable variants land in different ranges, the site uses a broader out-of-five band or withholds the quantity. It does not expose an unstable exact percentage.
02 · Polling trends
The dots are reports; the curve is a summary
The Polls page preserves the individual readings and adds one descriptive curve for each candidate.
Reading the chart
Each dot is one candidate’s reported share in one poll, positioned at the poll’s fieldwork date. The LOESS curve follows the local shape of those dots instead of drawing a straight segment from one poll to the next.
Each candidate is fitted independently. If a poll did not test a candidate, it contributes no dot and no inferred zero. Curves stop at the first and last observed dates, and a candidate with too few distinct observations appears as dots only.
This makes the polling record easier to read without turning it into the election forecast. The forecast uses its own full-ballot and uncertainty model described above.
See the polling record →A descriptive smoother
One candidate’s polls over time
Six reported poll values are shown as dots. A smooth line follows their local direction without passing through every dot.
03 · Council attention
Why a ward rises toward the top
Council evidence is thinner than mayoral polling, so the site organizes attention instead of assigning each candidate a win probability.
Four kinds of context
Open seats appear first because a race without an incumbent is inherently unsettled. For wards with incumbents, the Councillor Defeatability Index compares three measurements from the incumbent’s latest win. Separate ward facts, candidate-history hints, and any ward polls add context alongside that index; they are not blended into a hidden winner score.
The Councillor Defeatability Index was developed by Matt Elliott for City Hall Watcher. Each of its three measurements is ranked against the other scoreable Toronto incumbents, then the ranks are added with equal weight. A higher total means more comparative exposure on those measurements.
No incumbent is running, so the ward goes to the top of the attention order.
Compares vote share, eligible-elector support, and electorate growth against the winning margin.
Shows only candidate facts that match a separately supported historical association.
Preserves reported snapshots when they exist; they do not become a ward forecast.
Worked example · Ward 11, University–Rosedale
Why Dianne Saxe’s ward draws attention
These are concrete facts from Saxe’s 2022 win and the current electorate estimate—not a claim about who will win in 2026.
Saxe’s share of valid council votes in the win.
Votes for Saxe as a share of everyone eligible to vote.
Estimated additional electors compared with her 123-vote winning margin.
The first two values are among the lowest for Toronto incumbents, while estimated electorate growth is far larger than the previous winning margin. Together they explain the high index result and the separate growth flag without counting the same fact twice.
04 · Candidate-history evidence
A plausible idea is not automatically a public hint
The current upstream audit tested 35 candidate-facing flag definitions and publishes 12. Aggregate history means every confirmed previous elected-office race, including council; a hint may instead name one office type explicitly.
The evidence gate
The primary comparison uses candidates with confirmed identities in the 2010, 2014, and 2022 stable-boundary general elections. It accounts for election, the candidate’s role in the race, and field size, and groups uncertainty by council contest.
A categorical idea needs at least five matching candidacies in five contests and two elections, plus an adequate comparison group. A numeric idea needs at least 20 observations in ten contests and all three elections. Coverage only opens the door: the plausible effect must also stay entirely on one side of no relationship. The two Returning-councillor findings in open contests use an explicitly labelled small-sample tier: four contests, the same direction in all three elections, and additional influence and permutation checks.
Repeating in the same direction in every election earns the stronger evidence label. A pooled result can still publish when it clears the defined gate, but the site records that weaker level rather than pretending the pattern was universal.
Definitions that count every race except council—or add council only for some candidates—are not eligible for the public catalog. The history used by an aggregate hint must match the complete confirmed history a reader can see.
- Name a testable idea
Define the candidate fact, comparison group, race type, and direction before looking at the result.
- Check coverage and identity
Require enough candidates, contests, and elections, with confirmed links to their earlier races.
- Estimate the association
Compare vote share after accounting for election, candidate role, and field size, with uncertainty grouped by contest.
- Compare elections
Record whether the direction repeats across 2010, 2014, and 2022 or appears only in the pooled result.
- Publish, supersede, or withhold
Publish only a standalone statement the evidence can carry; retain every other tested definition in the audit.
Published
Previously elected trustee
17 candidacies · 16 contests · 3 elections
Associated with more council vote share in the same direction in 2010, 2014, and 2022.
Published
Won at least one previous race
13 candidacies · 12 contests · 3 elections
For non-incumbent, non-returning candidates with prior races, having at least one win was associated with more council vote share in all three elections.
Published
Previous race experience
107 candidacies · 68 contests · 3 elections
Non-incumbent, non-returning candidates with any confirmed previous race received more council vote share than otherwise comparable first-time candidates in all three elections.
Published
An unsuccessful council run
88 candidacies · 60 contests · 3 elections
The audit retains this as a pair—stronger than having no previous race, but weaker than other prior experience—but the ward cards do not turn that awkward comparison into candidate copy.
Published
A former councillor in an open race
4 candidacies · 4 contests · 3 elections
The small sample was approved at an explicit limited-evidence tier after the direction repeated in all three elections and passed additional influence and permutation checks.
Withheld
Previously elected MP
1 candidacy · 1 contest · 1 election
The estimated difference was large, but one example cannot support a public candidate-level statement.
Withheld
Each additional previous victory
199 candidacies · 110 contests · 3 elections
Across all candidate types, the uncertainty included no relationship. Among prior winners, additional wins showed no clear dose response.
Withheld
Two or more unsuccessful council runs
16 candidacies · 14 contests · 3 elections
The uncertainty included no relationship, and the estimated direction did not repeat across all three elections.
“Withheld” does not mean disproven. It means the available coverage, identity evidence, or uncertainty cannot carry a public candidate-specific statement. The corrected reader-readable screen tested 21 definitions and approved all 10 that cleared its evidence gate. Together with the two previously approved findings, that produces the current 12-flag catalog. Overlapping findings are combined into concise facts on ward pages. Continuous margin findings are not given an up/down arrow because the audit supplies no candidate-level cutoff, and the paired unsuccessful-council-run comparison stays in the audit rather than appearing as awkward candidate-card copy.
05 · Limits
What the evidence cannot settle
The site narrows its claims where the data is thin, historically unusual, or sensitive to reasonable modelling choices.
Polling coverage
Public polls can miss late movement, turnout differences, or a systematic error shared across firms.
Simulation assumptions
Historical polling misses guide uncertainty, but the current election can still behave differently from the past.
Historical eras
Boundary changes and short observable career windows limit which elections can be compared on equal footing.
Council outcomes
Incumbent defeats are rare, leaving too little evidence for honest ward-level win probabilities.
Candidate identities
A past race is attached only when the person link is confirmed; unresolved histories are not guessed from names.
Association, not cause
A historical pattern can help describe context without proving that the candidate fact caused later vote share.
06 · Glossary
Terms used on the site
A short reference for the labels that carry methodological meaning.
- Poll
- A reported survey reading from one pollster and respondent sample.
- LOESS trend
- A smooth curve fitted locally through one candidate's reported poll results.
- Win-chance band
- A public range, expressed as a frequency such as “about 4 times in 5,” rather than a fragile exact percentage.
- Sensitivity check
- A rerun under another reasonable assumption to see whether the public conclusion changes.
- Open seat
- A council race with no incumbent candidate in the field.
- Prior win
- An incumbent's latest council victory before the current election, including a by-election when it is the latest.
- Councillor Defeatability Index
- A City Hall Watcher comparison of Toronto incumbents using their prior vote share, eligible-elector support, and electorate growth relative to their winning margin.
- Exposure trigger
- A concrete ward condition that adds attention, such as estimated electorate growth exceeding the incumbent's previous winning margin.
- Historical hint
- A candidate-specific fact matching a supported historical association. It is context, not a candidate score.
07 · Sources
Records, polling, and attribution
The public pages combine official election records, published polls, and versioned analytical work.
Election and candidate records
Official City of Toronto election results and candidate-registration data provide council results, eligible electors, margins, and the current candidate slate.
Mayoral polling
Public releases from firms including Liaison Strategies, Forum Research, Mainstreet Research, Pallas Data, Abacus Data, and Ipsos. The Polls page retains the firm and fieldwork date for every reading.
Forecast history
Previous Toronto mayoral elections provide the omitted-candidate and poll-to-election uncertainty used by the current forecast.
Councillor Defeatability Index
Matt Elliott’s City Hall Watcher methodology supplies the three-component incumbent comparison reconstructed and backtested for this site.
Candidate-history audit
The upstream Council Defeatability Index project retains the complete tested hint catalog, evidence status, sample coverage, uncertainty, and identity audit. This page reviews contract 2.1.0.
Methodology reviewed Site data as of Aug 21, 2026