The monthly publication: exit-validated clearing rents, band-first, phantom-exits corrected. The method, in full: /paper.
Downtown Toronto, 3 km of Bay and Queen. Exit-validated clearing rents, not asking hopes. Methodology and caveats: /methodology. Numbers regenerate from the canonical corpus; rows carry engine provenance.
· Asking median $2,150 vs exit-validated $1,850: the market cleared 14.0% under ask (conformal 90% band plus or minus 15.5)
· Liquidity: median time on market 18 days (still-live censored)
· Exit labels this corpus: 81 page-confirmed, 407 unverified feed disappearances. Lifetime page checks: 82 confirmed gone, 741 still live of 1159 checked (the phantom-exit correction in action).
| cohort | n asking | asking | n exits | exit | gap % |
|---|---|---|---|---|---|
| 1 | 333 | $2,250 | 0 | n/a | n/a |
| 2 | 1 | $2,399 | 0 | n/a | n/a |
| ? | 118 | $1,970 | 1 | $1,850 | 6.1 |
| studio | 5 | $1,950 | 0 | n/a | n/a |
| scan | asking | exit | n ask | n exits | |
|---|---|---|---|---|---|
| 2026-07-02 | $2,292 | $2,250 | 38 | 349 | |
| 2026-07-03 | $2,300 | $2,250 | 27 | 442 | |
| 2026-07-04 | $2,300 | $2,250 | 27 | 442 | |
| 2026-07-05 | $2,370 | $2,250 | 22 | 545 | |
| 2026-07-06 | $2,350 | $2,250 | 13 | 586 | |
| 2026-07-07 | $2,300 | $2,300 | 7 | 643 | |
| 2026-07-08 | $2,208 | $2,250 | 16 | 688 | |
| 2026-07-09 | $2,225 | $2,295 | 32 | 763 | |
| 2026-07-10 | $2,225 | $1,790 | 24 | 3 | **break** |
| 2026-07-11 | $2,100 | $1,790 | 29 | 3 | |
| 2026-07-12 | $2,200 | $1,790 | 26 | 3 | |
| 2026-07-13 | $2,100 | $1,790 | 27 | 3 | |
| 2026-07-14 | $2,100 | $1,790 | 31 | 3 | |
| 2026-07-15 | $2,050 | $1,790 | 27 | 3 | |
| 2026-07-16 | $2,148 | $1,790 | 26 | 3 | |
| 2026-07-17 | $2,175 | $1,790 | 24 | 3 | |
| 2026-07-18 | $2,121 | $1,790 | 27 | 3 | |
| 2026-07-19 | $2,150 | $1,790 | 449 | 3 | **break** |
| 2026-07-20 | $2,162 | $1,790 | 514 | 3 | |
| 2026-07-21 | $2,175 | $1,790 | 508 | 3 | |
| 2026-07-22 | $2,009 | $1,702 | 6 | 12 | |
| 2026-07-23 | $2,224 | $1,758 | 12 | 8 | |
| 2026-07-24 | $1,900 | $1,795 | 205 | 6 | |
| 2026-07-25 | $2,195 | $1,758 | 490 | 8 | |
| 2026-07-26 | $2,195 | $1,680 | 492 | 5 | |
| 2026-07-27 | $2,185 | $1,702 | 478 | 4 | |
| 2026-07-28 | $1,900 | $1,650 | 208 | 2 | |
| 2026-07-29 | $2,200 | $1,650 | 457 | 2 | |
| 2026-07-30 | $2,178 | $1,850 | 470 | 1 | |
| 2026-07-31 | $2,150 | $1,850 | 457 | 1 |
### Breaks in this series
· 2026-07-10: Exit validation switched from naive feed-disappearance to page-confirmed exits with still-live censoring. Exit counts fall from the hundreds to single digits because the old counts were inflated by phantom exits (152 of 227 units page-checked were still live). Counts before and after this date are not comparable.
· 2026-07-19: Two changes land on this date. (1) Realtor.ca (CREA MLS) enabled as a source, adding roughly 600 downtown records to a corpus that previously qualified about 20 per day, so asking medians shift because the sampled population widened, not because landlords repriced. (2) An unobservable scoring pillar now renormalizes instead of scoring 0, so composite scores rise across the board and are not comparable to composites recorded before this date; 39 units cross 85 where none did the day before.
· Exit counts are small; treat the gap as directional and lean on the band, never the point alone.
· An exit is a page-confirmed departure where possible; unverified feed disappearances are labeled, and relisted units (new id, same unit) never count as exits.
· One geography, one product class. Other markets need their own corpus (the organism factory exists for exactly that).