Corrections log · last updated October 16, 2026

Corrections to PBP Real Estate market data

Published data earns trust by showing its mistakes. Corrections identify what was wrong, when, and what changed. Entries are appended; an incorrect entry receives a further correction. How figures are sourced and reviewed before they publish: editorial and corrections policy. Definitions: Seller Blink Rate™ and recorded days and Condo and HOA Fee Index.

What is logged

• Figure corrections: a published number, share, median, or count that was wrong on the day shown and was changed.
• Label and wording corrections: a segment, city, county, or period labeled incorrectly.
• Availability incidents: data pages or press-kit files that were unavailable or served stale figures for more than a few minutes.

A routine refresh is distinct from a correction. Monthly editions and release-day city cut editions retain their statistical dates; explicit corrected reissues identify a separate revision and correction date. Original as-published records are preserved, not silently re-frozen. Superseded public copies are identified as such; use the corrected edition for citation. Corrections must be visible inside affected standalone PDF, text and chart files, not only beside their links.

Log

1. October 16, 2026Label

   Release-day city cut, August 2026 edition (page and PDF)

   What was wrong

   The August 2026 edition labeled the price-cut paragraph under each county “What that hides.” The label implied the county reports conceal something; they measure closed sales and month-end inventory, and the share of sellers who cut asking is simply outside what they measure.

   What changed

   The label now reads “What the county report cannot show” on the page and in the PDF, and the city tables were restyled for legibility. No figure, date or source changed. The PDF carries a dated note on its first page.

   Read the August 2026 edition

2. September 28, 2026Figure

   Days on market and cumulative days on market, all price-cuts city pages, how-to-read-our-numbers and the Seller Blink Rate methodology

   What was wrong

   Through September 27 the pages reported the MLS DaysOnMarket and CumulativeDaysOnMarket counters as recorded. Those counters refresh in batches: on ordinary days 56% of active listings trailed the calendar by a median 23 days, and on September 24 the counters for 69% of active listings jumped more than two days in one refresh. Week-over-week day medians moved with the refresh schedule rather than the market.

   What changed

   From September 28 PBP counts days on market from the listing contract date recorded in the MLS to the snapshot date, adding the recorded prior time on market for listings the MLS records as relisted. The history back to September 9 was recomputed on the same definition, so the series steps up once and stays comparable. Counted medians run roughly 6 to 15 days above the old counters. Asking-price comparisons are unaffected.

3. September 8, 2026Figure

   Buyer-perception-of-price article and its historical study

   What was wrong

   The September 7 article described its whole-study sample as covering only five named counties, but the calculation did not enforce that restriction. The original aggregate's total exceeds its named county subtotals by 3,072 observations whose counties were missing, unrecognized or outside those groups; the exact breakdown is unresolved. The article also described price endpoints as complete reduction histories, recorded market days as verified time to contract, and group differences as evidence of buyer behavior or negotiating leverage.

   What changed

   The article now carries a dated correction at the same URL. All original chart values and table figures are preserved in a collapsed historical section, locally marked not cleared for citation, with corrected field definitions and source notices. Unsupported claims are removed from its title, description and FAQ markup; seller education and the human-pricing offer remain. The exact original listing-level input has not been found in the sources checked. No newer input was substituted, and no replacement five-county medians are claimed. This does not establish that every original ratio is wrong. Related social versions remain held from new distribution. Automatic regeneration of the old study is blocked pending a separately verified source, reviewed implementation and Grant's approval.

   Read the correction and held historical figures

4. September 8, 2026Availability

   Obsolete direct press-file downloads

   What was wrong

   Unindexed September 5 files for Cape Coral, Key Largo, Kissimmee and LaBelle remained directly downloadable outside this program's five-county coverage. Obsolete Royal Palm Beach HOA files also remained at former current-series download addresses after current eligibility changed.

   What changed

   Original bytes are preserved internally. Exact out-of-area download URLs now lead to a withdrawal explanation rather than another city's figures. Obsolete Royal Palm Beach HOA downloads lead to its separately dated September monthly page and corrected HOA edition; they are not September 8 current figures. Stable current-series redirects are removed if publication eligibility recovers. The separately documented original monthly-file correction mappings remain unchanged.

   Read the withdrawal explanation

5. September 8, 2026Figure

   September 2026 monthly press editions: corrected reissue r1

   What was wrong

   The original September 7 editions contained the wording and methodology issues described above. The original earlier source rows have not been found, so we cannot claim an exact reconstruction of the as-published figures.

   What changed

   A separately labeled reissue r1 retains the September 7 statistical date and original freeze time. It recomputes figures from the preserved later same-day BeachesMLS pull at 2026-09-07T12:06:07+00:00, not the missing earlier pull. Revised files identify the September 8 correction date and this limitation. All original city and metric entries are retained, original evidence is preserved, and each monthly page provides the before-and-after figure and wording ledger. Superseded public file links lead to the explicitly revised files.

   Open monthly editions and their change ledgers

6. September 8, 2026Label

   Condominium and HOA fee sample descriptions

   What was wrong

   Some fee headlines and comparisons read as if they described all active listings rather than the usable-fee sample retained after exclusions and percentile trimming. Eligibility before trimming, retained sample sizes, independently trimmed regional comparisons and construction-year subgroups were not always distinguished.

   What changed

   Fee figures identify the retained usable-fee sample and the denominator for each secondary measure. City publication requires at least 30 usable fees before trimming; the retained count may be smaller. City, county and regional samples are trimmed separately. Construction-year groups use the recorded YearBuilt field and do not establish why fees differ or the financial condition of an association.

7. September 8, 2026Figure

   County benchmarks and published-city peer comparisons

   What was wrong

   Some county benchmarks assigned every listing under an MLS city name to one county even when the records spanned counties. Some peer lists also included the retired Lake Worth city label alongside the published Lake Worth Beach series.

   What changed

   County benchmarks now group each listing by its recorded county. City pages distinguish their presentation county from that separately calculated benchmark and disclose cross-county record counts. Peer comparisons use the edition's published-city configuration and exclude the retired Lake Worth label; the underlying records are not deleted from geographic aggregates.

8. September 8, 2026Label

   Seller Blink Rate, recorded days on market and asking-price bands

   What was wrong

   Some wording described listings not below their original asking price as having never reduced, treated recorded cumulative days as a complete relisting history, or called current asking-price bands original-price bands. Those descriptions went beyond the recorded fields. Some small subgroups and tied results also needed clearer presentation.

   What changed

   The measure is current asking price below recorded original asking price, with unchanged, raised and missing-price counts distinguished. Days describe the recorded MLS fields and their available-record denominators, not a reconstructed property history. Bands use current asking price; ties use displayed precision, and small subgroups show counts or explicit sample warnings rather than unsupported comparisons.

9. September 8, 2026Availability

   Current price-reduction, condominium-fee and HOA-fee downloads

   What was wrong

   September 8 pages could link to permanent downloads that still delivered cached September 7 content. A current date on the page did not establish the date of the downloaded file.

   What changed

   Current pages now link to edition-specific files, while permanent latest links remain available. Press downloads use a no-store browser policy and bypass the edge cache. Every file identifies its statistical date and exact source-pull time. Previously downloaded copies cannot be recalled; download the edition linked from the current page before citing a current figure.

10. September 8, 2026Figure

    Hobe Sound combined condominium, townhouse and villa asking-price median

    What was wrong

    The September 7 median was $2,495,000 across 31 active listings. Listings classified as Villa were omitted from this combined category.

    What changed

    The category now includes villas with townhouses. Including the 15 villas in the September 7 and September 8 BeachesMLS snapshots gives 46 active listings and a median asking price of $461,500 on both dates. The same listings had the same asking prices in both snapshots. This is a classification correction, not a one-day market-price decline.

11. September 7, 2026Figure

    Homepage and /homes-for-sale/ coverage counts

    What was wrong

    The advertised coverage read 63 cities while 101 city data pages were published. The count was typed by hand and had not been updated as pages were added.

    What changed

    Coverage counts are now derived from the published pages at build time and the build fails if an advertised number exceeds the pages actually linked.

12. September 7, 2026Availability

    Press kits, price cuts and days on market, 83 cities

    What was wrong

    The city press-kit files under /press/ returned 404 for about 90 minutes (2:25 PM to 3:55 PM ET) after a manual deploy shipped a build that did not include the morning's kits.

    What changed

    Kits restored from the morning build the same afternoon. The deploy procedure now checks that every indexed kit is present in the build before it goes live.

Report a possible error

Email info@pbpre.com with the page URL, the figure, and the as-of date printed on the page. Every report is checked against the source feed. Gia Freer, Broker of Record (Florida license BK689801), and Grant Freer, Owner and Broker Associate, answer for the figures. PBP Real Estate, LLC, 700 NW 7th Drive, Boca Raton, FL 33486. 561-395-8418.

This data is not verified for authenticity or accuracy and may not reflect all real estate activity in the market. Based on information © 2026 BeachesMLS, Inc. Listings permitted for internet display. Equal Housing Opportunity.