Corrections log

September 8, 2026 — Corrections from a full audit of the site’s claims and sources. (1) The fall 2026 child Swimmer 1 figure of 696 seats appeared on several pages but matches no catalogue capture; the captures show 690 seats at release on August 25 and 702 by September 4. Every page now uses 690 (release basis) and the rebound is 28%, not 29%. (2) The pipeline chart’s “Instructor recert 208” came from a query in analyze_for_site.py whose OR-clauses escaped the year filter, pulling 2023 and 2025 rows into the 2024 bar; the corrected 2024 figure is 154 and the essay no longer calls it a per-year number. (3) A sentence on the data page (“only ~1,800 of 26,095 wait entries coexisted with an open seat”) could not be reproduced from any published script and has been withdrawn; the page now states the verifiable version (26,095 wait entries against about 3,000 unfilled seats in 2024). (4) The Park Board’s June 1, 2026 vote was described as unanimous; the minutes record the report-back amendment carrying 4–3 and the motion as amended 5–1 with one abstention. Council’s May 20 vote was unanimous. (5) The quoted clause D of the June 2, 2025 “Think 50 Metres” motion spliced words from clause C; the quotation now matches the minutes. (6) Council’s defeated amendment on May 20 would have had council “consider new funding to the Board of Parks and Recreation,” not directed funding; the wording was corrected on every page. The audit also added source links throughout, published the FOI letters and raw captures, and corrected smaller items (48%, not 45%, seat growth 2022–2025; eight remaining indoor pools, seven modelled; “requested,” not “ordered,” for council’s Kitsilano clause). The CUPE Local 15 pay claims were re-sourced to the agreements: the 13(a) rates ($27.39–$32.20 in 2024) and premiums held up, but the statement that an employee can start a reclassification was true only of the 2016–2019 agreement — the current one names management or the Union — and “premiums for shift work” applied to listed trades classes, not lifeguards; both were corrected. (7) The preschool-ladder figures on the data page had no script behind them; they were recomputed with a published one (analyze_preschool_ladder.py) under stated definitions. The story held; the numbers moved at the margins (38% of 2025 Swimmer 1 entrants with preschool history, not 37%; 89% of Preschool 1 finishers re-entering at Swimmer 1, not 480 of 538; 57% of Preschool 5 finishers skipping past Swimmer 1, not 55%; reappearance 23% to 69% for children whose preschool history ended by 2024, not 30% to 70%), and the fixes page now says 150–200 children a year reach Preschool 5, not 147. (8) Richmond’s training subsidy had been described from its September 2022 trial proposal only; the city’s 2023 launch notice, 2025–26 program sheet and current web page show it has run continuously since May 2023, so “introduced in 2022” became “since 2023.” The Britannia item on the minutes page now quotes the two clauses that were struck and links both the October 30 notice and the November 27 amendment. (9) The service-level yardstick on the working answer and fixes pages gave saturation bands of “170–240” and “140–180” seats per 100 kindergartners that did not follow from their stated inputs; it is restated as a cohort model (one kindergarten cohort through each of three levels a year, at 1.5 or 2.5 attempts per level), computed in section E of the published analysis, and the “seats needed” are now defined as additions to the current program with the upper levels held constant. The staff-turnover assumption of 25–35% a year, previously unsourced, is now 15–35% on Ottawa and Saskatchewan records, so the annual replacement need reads 33 to 75 certifications, not 55 to 75. Option C’s cost, previously $150–250K with no derivation, is re-derived from the Leisure Access discount rule, pass and eligibility counts, and the average fee, at $25K to $240K depending on the gate.

September 4, 2026 — “The largest fall entry gate in three years” was wrong; withdrawn. Our registration-night analysis compared this fall’s catalogue against prior years using two different counting bases: the 2026 figure included Adult & Teen Swimmer 1 classes; the historical figures counted child classes only. On a consistent child-only basis, fall 2026 offered 690 seats at release — a genuine 28% rebound from 2025, but below fall 2024’s 768, not above it. Every page now uses the consistent series (598 → 768 → 540 → 690), and the error produced this site’s sawtooth finding: the fall gate swings 40% year to year with no published target.

September 4, 2026 — Registration-night fill restated from 85% to 93%. Same cause, same fix: on the consistent child-class basis, 642 of 690 Swimmer 1 seats were taken on the release evening (93%), not the 85% first computed on the blended basis. The correction made the finding stronger, which is not why we made it.

August 15, 2026 — The measured “repeat rate” was withdrawn as a pedagogical number. We initially reported that 64–69% of beginner-level attempts were repeats, computed from the city’s released registration extract. Reconciling the two released datasets showed the extract silently mixes waitlist entries with enrolments (verified per-activity: unique children = enrolments + waits), so the figure blends true repeats with re-queuing and cannot be read as a pass/fail measure. It is now presented only as a re-engagement rate, the dependent findings carry caveats, and a corrected extract with an enrolment-status column has been requested from the city. The format-effectiveness comparison (intensive vs weekly) carries the same caveat until that data arrives.

Why this page exists: this site’s core claim is that the swim system runs on unmeasured and unshared numbers. The only honest way to make that claim is to measure in public, which includes being wrong in public. If you find an error - in arithmetic, sourcing, framing - please let us know.