Executive Summary

Last week’s WEEKTICK™ study identified a positive convergence between the leading demand ratios and the average closing price. This week’s update — now 1,838 consecutive weeks of CRMLS data through Sunday, August 30, 2026 — is a follow-through test: did the signal firm up or fade? It firmed up, on three fronts at once. Price momentum accelerated for a second straight week (+1.5% → +1.7% year-over-year on the 26-week MA, which set a new series high of $1.567M). The demand-quality ratio (#C+#P)/XK jumped to a raw 3.74 — its best single-week print of the summer — as expired/cancelled listings fell to 12 per week, the lowest failure count in over six weeks. And the absorption ratio #P/#N.L. ticked up to 0.590 raw, holding above its long-run August seasonal norm.

The base case is unchanged and now better supported: continued recovery of roughly +4% to +8% in Ave. $ of Closing over the next 26–52 weeks, carrying the 26-week MA from $1.567M toward the $1.63M–$1.69M zone by mid-2027.

The Evidence, Updated

1. The lead-lag structure holds on new data

Re-running the full lead-lag test on the updated series: #P/#N.L. still peaks at r = 0.60 leading price momentum by 26–39 weeks; (#C+#P)/XK peaks at r = 0.58 at a 13–26 week lead. No degradation from last week — the correlation structure is stable, exactly what you want from a signal you intend to act on.

2. Quintile map: both indicators remain in the constructive Q3 zone

New this week: the raw (#C+#P)/XK print of 3.74 sits inside Q4 territory. One week does not move a 26-week average far, but the smoothed value (3.29) is now within 0.24 of the Q4 boundary (3.53) — the zone where 95% of historical weeks were followed by a positive year. If the current trajectory holds, the smoothed indicator crosses into Q4 within roughly 8–12 weeks: a specific, watchable upgrade trigger.

3. The positive convergence — now with a falling failure rate

The convergence identified last week strengthened on every axis. (#C+#P)/XK is up +0.67 year-over-year on the 26-week MA (2.62 → 3.29 since its December 2025 trough of 2.58), and this week the improvement came from the healthiest possible source: the denominator. Expired and cancelled listings dropped from 15.0 to 12.0 per week — a 20% decline in listing failures in a single week. Fewer failures means fewer discouraged sellers, fewer stale comps, and a cleaner negotiating backdrop. Meanwhile price momentum accelerated for the second consecutive week and the 26-week MA printed a fresh series high. Indicator rising, failure rate falling, price confirming: the convergence is doing exactly what a durable one should.

4. #P/#N.L.: the seasonal read-through stays on script

Last week we classified the 26-week softness in #P/#N.L. as roughly 80% seasonal. This week supports that call: the raw ratio rose to 0.590, above the 35-year August norm of 0.583, and the year-over-year change on the smoothed series is +0.04. The watch item is unchanged — the ratio should begin its normal climb toward the November–December seasonal peak (long-run December norm: 0.78). Two consecutive weeks at or above seasonal norm is the pattern behaving; failure to turn up by late October would be the tell that something more than seasonality is at work.

5. Historical analogs: the record stays perfect

The joint-configuration filter (#P/#N.L. 0.55–0.61, (#C+#P)/XK 3.0–3.55, price momentum –2% to +5%) still returns 24 historical weeks across five episodes — 1999–2001, 2012, and 2022 — and all 24 were followed by a positive 52-week price change, averaging +6.7%. This week’s data added a 25th qualifying week to the live count; its verdict arrives in August 2027.

Conclusions

• Direction: up, with more confirmation than last week. Three independent measures improved simultaneously — price momentum (+1.7% YoY, accelerating), demand quality (raw 3.74, best of summer), and listing failures (12/week, multi-week low). Convergences that strengthen on follow-through are the ones worth trusting.
• Magnitude: +4% to +8% over 26–52 weeks, reaffirmed. The 26-week MA at $1.567M projects to $1.63M–$1.69M by mid-2027 — deliberately below the Q3 historical average of +7–8%, and far inside the dataset’s verified extremes (+31.0% July 2004; –33.4% July 2009).
• The upgrade trigger to watch: smoothed (#C+#P)/XK crossing 3.53 into Q4 (currently 3.29 and rising ~0.02–0.03/week). Historically Q4 weeks preceded a positive year 95% of the time. On current trajectory: late October to November 2026.
• The downgrade trigger: #P/#N.L. failing to begin its seasonal climb by late October, or XK failures re-accelerating above ~18/week for multiple weeks.

Actionable Advice for Sellers

• The failure rate just dropped 20% — that is your green light. Twelve expired/cancelled listings per week is the cleanest failure count in over six weeks. Homes that are priced correctly are surviving to sale at the best rate of the summer. The market is telling you preparation and precise pricing are being rewarded right now.
• The launch window is now through early November. New listings held at 46/week — still well below July’s 51+ — while the absorption ratio sits above seasonal norm and climbs toward its November–December peak. Listing in the next 4–8 weeks positions you in front of peak seasonal demand with below-peak competition.
• Price to the new trendline, not the old headlines. The 26-week MA just set a series high. Comps from the February–April trough understate today’s market by roughly 10 points of momentum swing. A derivative-calculus pricing analysis for your street will show precisely where the recovery has — and has not yet — reached.
• Expect negotiation, not capitulation. The average discount ran 2.2% off list this week (4-week average ~2.1%). That is a market that negotiates but does not punish — hold a well-supported price and let the thin-discount environment work for you.

Actionable Advice for Buyers

• Your leverage window is measurably narrowing. The XK pool — your best source of motivated sellers — just shrank 20% in one week. Fewer failed listings means fewer desperate counterparties. The data says the discount era of this cycle is being absorbed in real time.
• September remains statistically your best month. The absorption ratio seasonally bottoms in September (35-year norm: 0.562). You have roughly 4–6 weeks of the year’s thinnest buyer competition left before the November–December demand surge.
• The cost of waiting compounds weekly. At the reaffirmed +4% to +8% base case, deferring a $1.6M purchase 12 months costs $65,000–$130,000 — and this week’s acceleration (+1.5% → +1.7% in seven days) suggests the lower bound is more at risk than the upper.
• Negotiate on the 2% reality. Sellers are conceding about 2 points off list on average. Anchor offers with that fact and with any property’s days-on-market history — but recognize that in a firming Q3-and-rising indicator regime, lowball offers mostly buy you silence.

The Devil’s Advocate: How This Analysis Could Be Wrong

The case against, updated for this week’s data and stated at full strength:

• The significance problem has not gone away. Adding one week changes nothing about serial dependence: the effective number of independent observations behind the r ≈ 0.6 lead-lag correlations is still roughly five market episodes, and at that effective sample the correlations are not statistically significant (p ≈ 0.31–0.33). “Statistically valid” in the strict sense remains out of reach; what we have is consistent, economically sensible, and unproven.
• One good XK week is one week. The 20% drop in failures (15 → 12) is well within normal weekly noise — six weeks ago the count was 21.7. Celebrating a single denominator print risks exactly the overfitting this report warns against. The 26-week average moved just 0.02 on the news.
• The series-high price could be mix, not market. At ~17.6 closings per week, the new record in the 26-week MA can be manufactured by a handful of estate-tier closings rolling into the window. A median-based or repeat-sales measure could show a flatter recovery.
• Causality still runs both ways. Granger tests continue to show price momentum predicting the ratios roughly as well as the reverse. Part of what looks like “demand leading price” may be price optimism recruiting demand.
• Regime risk is unhedged. Every analog episode (1999–2001, 2012, 2022) unfolded in flat-to-falling financing-cost environments. A material rate shock would invalidate the analog map faster than any indicator could warn you.
• Seasonality cuts both ways. We excuse #P/#N.L. softness as seasonal — but then some of the coming November–December “improvement” will also be seasonal, and must not be double-counted as fresh evidence for the bull case. The honest test is the year-over-year comparison, not the calendar climb.

Net of cross-examination: the directional call earns slightly more confidence than last week — follow-through happened — but the honest label remains “moderately probable,” not proven. Sellers should not price for the top of the range; buyers should not treat +8% as destiny.

Derivative Calculus Insight

Methodology notes: All series smoothed with 26-week moving averages (min 13 periods). Data: 1,838 consecutive weekly CRMLS observations, May 24, 1991 – August 30, 2026, dual-pass verified (pandas + openpyxl row and year-count Counter assertions passed). YoY extremes verified with current/prior values (+31.0% Jul 2004: $615,527 vs $470,042; –33.4% Jul 2009: $580,800 vs $872,656). Projections bounded by these dataset extremes. Quintile forward returns computed on all weeks with complete 52-week forward windows. Statistical caveats disclosed in full above. This analysis is market commentary, not a guarantee of future results.