State of the Election — Toronto 2026

The standing baseline of the biweekly record. Edition of August 14, 2026 — another complete refresh: the City Clerk's feeds re-pulled, every previously-absent candidate searched again, every previously dead or unreadable site re-fetched, and every candidate with new or changed evidence individually re-read under our strictest instrument yet. The receipts are at the bottom. Previous baseline: August 5. Ward by ward: the new Ward Profiles hub, one page per council ward. Province-wide: State of the Election — Ontario, all 2,289 races.

Reading this in real time: every field figure below — candidate counts, filings, web presence, incumbency — is a snapshot with figures as of August 14, 2026, taken directly from the City Clerk's candidate feeds and re-verified candidate by candidate the same day. Nominations close August 21, 2026, at 2:00 p.m. — the field is still moving (six people filed today), and every number here — provisional until that close — will be recomputed and re-stamped the day after.

The whole story in one paragraph

Seven days before nominations close, Toronto's election is still moving: 253 people have filed for the city's 55 offices — six of them today — and the field has widened further since our last full refresh. The thin edge is thinning: two-candidate council wards fell from five to three. The single seat this page has named for weeks as having attracted nobody at all — TCDSB trustee Ward 8 — now has a candidate. The deeper shape has barely moved: twenty-two of twenty-five sitting councillors have already re-filed atop 2022's landslide margins, and — now individually re-verified rather than assumed unchanged — about three in ten candidates still have no findable web presence at all, while fewer than half have published anything resembling a platform, and a much stricter, commitment-gated re-read of every scoreable candidate finds fewer than one in five actually name a specific, checkable promise. Around the 1.9 million people who may vote stand roughly 900,000 residents who legally cannot, a daily region of millions more, and a scattered online conversation with no common room. Every one of those facts, read correctly, is an opening — and the first organized act is free, takes thirty seconds, and is open to every person in this paragraph including the ones who can't vote: register for the conversation. If you've ever considered the larger act, the papers take one afternoon, and the window closes August 21 at 2 p.m.

1. The moving field — seven days can still change plenty

All figures as of 2026-08-14, from the City Clerk's official candidate feeds:

Why this matters to you: the field is provably movable, this week, by ordinary people — the seat this page named "attracted nobody at all" now has a name on the ballot because one person filed a form. In one of the three remaining two-candidate wards, a handful of organized neighbours is still the whole game. Here's what running takes. The clock says August 21, 2:00 p.m. — seven days.

2. The entrenchment — deeper than last edition, and nearly complete

Why this matters to you: unless something changes in the next seven days, your 2026 councillor is very probably your 2022 councillor, chosen by a fraction of your neighbours, by a wide margin. That is not an accusation of any person — it is a system running on defaults, and defaults are exactly what organized neighbours and real contests disrupt. Seven days remain to widen the field; four years of receipts-keeping begin the day after it closes.

3. The presence and platform layer — re-verified, and re-scored under a stricter test

What changed since the last edition is not just the numbers — it's a second layer of rigour on top of the wall-to-wall measurement we already had. This refresh re-pulled the Clerk's feeds, searched again for every candidate with no site on file at all, re-fetched every previously dead or unreadable site, and individually re-read every candidate whose evidence was new or materially changed under our strictest instrument yet — with an independent blind check built in: a second reader re-scored a random slice of the freshly-read candidates without seeing the first reader's answer, and came back in full agreement on class, every score within one point. These figures describe candidates, not our coverage:

A stricter test, new this edition. "Has a platform page" answers one question; "does that page name a specific, checkable commitment" answers a harder one. Every Toronto candidate was placed into exactly one class under our strictest, commitment-gated instrument — aggregate counts only, no candidate is ever scored, ranked or named negatively:

classn (of 256)%what it means
Names a checkable commitment4818.8%A specific, forward, checkable promise — a number, mechanism, or date
Published, but no checkable commitment7830.5%Real material online, but nothing specific enough to hold anyone to
A record, not a forward offer197.4%Past delivery described; no consolidated forward pledge
Nothing found10340.2%No published platform located — about a quarter of this group are brand-new filers not yet built out
Unreadable or unattributable83.1%Site dead, or content genuinely couldn't be verified as the candidate's own

Read plainly: this is not "fewer candidates are trying" — the field grew mostly with brand-new filers who haven't published anything checkable yet, seven days before close, which is exactly what you'd expect. One Toronto candidate's platform cleared this edition's absolute strictest bar — a specific, quotable, numbered commitment, independently double-checked — and joins the exceptional platforms list. The full measurement — every published platform assessed for breadth, depth and quality under a public rubric — is its own page: the Platform Baseline.

None of this is contempt, and absence of a platform is never treated as a fault of the candidate. Candidates are running largely alone, unstaffed and unfunded, many with days left on the clock — the vacuum is the shape of the invitation, and it is also the standard question this project puts to every candidate identically: by voting day the city should know exactly what each candidate proposes to do, written down, costed, dated.

Why this matters to you: you are hiring 26 executives for a multi-billion-dollar organization plus 29 school trustees, and as of today more than half have published no platform, three in ten cannot be looked up at all, and fewer than one in five clear a strict test for a checkable promise — facts we can now state about every candidate, individually re-verified, not assumed carried over. You would not hire a contractor this way. The library's honest briefs and question banks exist so that any candidate — including one who files next week — can meet a professional standard in days, and so voters can demand it.

4. The turnout floor — and live-verified doors

Why this matters to you: at 29.7%, a municipal ballot is one of the highest-leverage acts a citizen has — not because any ballot weighs more, but because the deciders are three times fewer. If you missed the online registration window, the in-person path is still open with seven days to go.

5. Who can participate — the claims register holds

July's row-by-row reconciliation stands, unchanged by today's field refresh:

Why this matters to you: if you can vote, you are one of 1.9 million holding a proxy for nearly a million neighbours who can't. Your ballot is not only yours. If you can't vote, you are not voiceless: every other instrument in this project — the conversation, the vision bank, the volunteer bank — is open to you on identical terms, and the count of non-voter participants is a number this movement reports out loud.

6. The fragmented conversation — the thesis, sharpened

Toronto's civic conversation still happens in a thousand disconnected rooms — neighbourhood Facebook groups, language-community WhatsApp and WeChat circles, Reddit, X, campus Discords, congregation newsletters — each partial, ephemeral, unaccountable, and walled off from the rest. A false claim refuted in one room lives on in forty others; a thousand rooms each concluding "someone should do something" sum to nothing. This edition's re-verification adds the sharpest version of the supporting fact yet: the conversation is so scattered that roughly three in ten of the people asking to lead it still cannot be found in it — up, not down, from a month ago — and our re-verification pass again found live candidate pages carrying ward numbers that no longer exist and campaign sites carrying repurposed content from other campaigns and other years, each case flagged in the registry, none treated as disqualifying on its own. No common room exists in which Toronto can hear itself. The answer is not "everyone leave your rooms" — the rooms are assets. The answer is one thin, common layer over many rooms: a single registration — register for the conversation — that any resident of the city, the region, or the world can complete in thirty seconds, that returns receipts instead of newsletters, and that carries what each room concludes into one shared, archived record. Your room stays yours. The city finally gets one.

7. What we still don't know — on purpose, out loud

Honesty gaps are listed, not hidden; nothing below is asserted anywhere above. Open verification lanes (each has an owner and a next step): per-ward and poll-level 2022 turnout (blocks any "your ward stayed home" claim — the single most motivating stat this page could hold); age-cohort turnout; the exact count of municipally excluded adult residents (blocks upgrading "hundreds of thousands"); the current city population ("a city of three million" is common usage, not a held figure); Toronto's non-permanent-resident count; the naturalization share of immigrant Toronto; one ward's 2022 baseline discrepancy (that ward is cited nowhere on this page by name); and trustee incumbency, which the TDSB's 22-to-12 ward restructure broke the records for. New honest classes from today's pass: 20 candidate sites unverifiable by plain fetch (held, neither present nor absent), 9 candidate rows browser-pending (today's re-fetch hit a script wall; the prior finding is carried, not counted as a change either way), 3 flagged registry residuals (names on our books whose only official filing is a different race — excluded from every field-figure count above), and the standing certainty that a late-filing wave will move every field figure between now and August 21 — seven days out — which is why every number carries today's date, and why this page returns every two weeks until voting day.

8. The leverage math — unchanged, only more so

No formula converts electors into a victory number, and we don't publish one. What the honest math says: where seven in ten stay home, a crowd the size of a good summer festival is decisive — and Toronto produces those every summer weekend. Every section above shrinks the crowd it takes: a 29.7% floor (§4), three two-way wards and a school-board seat that just filled (§1), defaults nearly fully filed (§2), a platform vacuum any prepared candidate or informed voter instantly stands out in (§3, sharper this edition than any before it). Apathy is not our obstacle. Apathy is our opportunity — and the pool of potential participants is not 1.9 million electors but everyone the decision touches.

You are not one voice in three million

You are one of a festival-sized crowd it would actually take — and if you can't vote, you are one conversation registration, one problem card, one vision statement away from counting anyway. Here's the one act every reader of this page can take today, whether you can vote in this election or not: register for the conversation. The same thirty-second act for the voter, the permanent resident, the commuter from Brampton, the international student, and the diaspora cousin watching from abroad.

The signature rail is being built and switches on when its custodian line is real; until then, start at the door — and if the larger act is calling you, so is the deadline.

Receipts — where every number comes from

Every figure on this page traces to one of these sources, each checkable without asking our permission:

Per-candidate assessments are never published; aggregates only — committed before publication, held privately by design. No aggregate on this page is attached to a group so small it would describe one or two identifiable people.

Past editions — how this page has changed

This page publishes as one standing record, updated in place with each refresh rather than re-hosted at a new address every time. Earlier drafts are kept as dated internal editions, not separately web-published — described here rather than linked. In reverse date order:

Found a mistake?

That's the system working. Every figure on this page traces to a public source or a computation we show; if one is wrong, out of date, or reads differently than the source supports, we want to know before anyone else repeats it. Send it to the corrections page — it takes thirty seconds, and every correction is published, dated, and left visible.