Scroll to end for video review…
WEEKTICK™ Weekly Predictive Intelligence • 1,839 Consecutive Weeks • May 1991 – September 2026
Labor Day Split Decision: Absorption Improves, Demand-Quality Wobbles, and the Q4 Upgrade Trigger Moves Out
Week Ending Sunday, September 6, 2026 • Predictive indicators: #P/#N.L. and (#C+#P)/XK vs. Ave. $ of Closing • Palos Verdes Homes & South Bay
Ave. $ of Closing (26-wk MA)
$1.574M
new series high; raw week $1.431M
Price YoY Momentum
+2.2%
4th straight positive week; +1.7% last wk
#P/#N.L. (26-wk MA)
0.579
43rd pctile • raw jumped to 0.641
(#C+#P)/XK (26-wk MA)
3.23
down from 3.29 • raw fell to 2.36
Expired/Cancelled (XK)
20.0/wk
up from 12.0 • same week 2025: 20.3
Base-Case 12-Mo Projection
+4% to +8%
held, with lower conviction
Executive Summary
Two weeks ago this series identified a positive convergence between the demand ratios and the average closing price. Last week it firmed up. This week — now 1,839 consecutive weeks of CRMLS data through Sunday, September 6, 2026 — delivers a split decision, and I would rather report the split than paper over it.
The good: the absorption ratio #P/#N.L. printed a raw 0.641, the strongest week since mid-July, as new listings fell to 42.6 per week (lowest in four weeks) while pendings held at 27.3. Price momentum on the 26-week moving average accelerated for a fourth consecutive week, +1.5% → +1.7% → +2.2% year-over-year, and the MA itself set another series high at $1.574M.
The wobble: the demand-quality ratio (#C+#P)/XK fell from a raw 3.74 to 2.36 as expired and cancelled listings jumped from 12 to 20 per week. That dropped its 26-week MA from 3.29 to 3.23 — the first decline in this smoothed series in four weeks — and moved it further from, not closer to, the 3.54 Q4 threshold I flagged last week as the upgrade trigger.
The verdict: the positive convergence is intact but narrower. The stronger of the two predictors (#P/#N.L., r = 0.61 at a 39-week lead) improved; the weaker one (r = 0.58 at a 13–26 week lead) gave ground on what looks, on the evidence below, like the ordinary post-Labor-Day purge. The base case of +4% to +8% over the next 26–52 weeks stands, but conviction is lower, and the Q4 crossing I projected for late October–November now looks more like December–Q1 2027 unless XK collapses. Details and the arithmetic follow.
The Evidence
1. Lead-lag structure: unchanged on the updated series
Re-running the full lead-lag test on all 1,839 weeks: the 26-week MA of #P/#N.L. correlates with price momentum 26–39 weeks later at r = 0.61 (peak at 39 weeks); (#C+#P)/XK peaks at r = 0.58 with a 13–26 week lead. Correcting for the heavy autocorrelation in smoothed weekly data (Bartlett effective-N ≈ 23 independent observations), the t-statistics are 3.5 and 3.3 respectively — both clear the 99% confidence bar. These are statistically valid leading relationships, not artifacts of smoothing. The caveat, as always: r ≈ 0.6 explains roughly a third of the variance in forward price momentum. Two-thirds is something else — rates, the stock market, seasonality, and luck.
2. Quintile map: both indicators remain in Q3, but (#C+#P)/XK is drifting toward the middle of it
| (#C+#P)/XK 26-wk MA quintile | Range | Avg next 26 wks | Avg next 52 wks | Weeks |
|---|---|---|---|---|
| Q1 (weakest) | 0.56 – 1.78 | +0.2% | –1.8% | 353 |
| Q2 | 1.78 – 2.77 | +0.2% | +1.9% | 352 |
| Q3 ← current 3.23 | 2.77 – 3.54 | +4.1% | +8.0% | 352 |
| Q4 | 3.54 – 4.29 | +4.3% | +8.8% | 352 |
| Q5 (strongest) | 4.29 – 6.39 | +5.2% | +11.1% | 353 |
| #P/#N.L. 26-wk MA quintile | Range | Avg next 26 wks | Avg next 52 wks | Weeks |
|---|---|---|---|---|
| Q1 (weakest) | 0.19 – 0.41 | –1.3% | –3.2% | 353 |
| Q2 | 0.41 – 0.56 | +0.6% | +2.4% | 352 |
| Q3 ← current 0.579 | 0.56 – 0.69 | +3.4% | +7.2% | 352 |
| Q4 | 0.69 – 0.82 | +3.9% | +9.1% | 352 |
| Q5 (strongest) | 0.82 – 1.15 | +7.3% | +12.4% | 353 |
The step from Q2 to Q3 is the one that matters historically (+1.9% → +8.0% and +2.4% → +7.2% forward 52-week returns). Both indicators sit above that step. But note that the Q3 floor for (#C+#P)/XK is 2.77 and we are at 3.23 with the MA falling. The margin of safety is roughly 0.46 — about eight more weeks like this one before the demand-quality ratio would be arguing for the muted Q2 outcome.
3. Was this week signal or seasonal noise? The 2025 fingerprint says noise
| Week | New Listings | XK | Pendings | Closings | Raw Ave. $ of Closing | (#C+#P)/XK |
|---|---|---|---|---|---|---|
| Aug 31, 2025 | 36.6 | 19.9 | 26.7 | 18.0 | $1,644,610 | 2.25 |
| Sep 7, 2025 | 49.4 | 20.3 | 24.6 | 16.9 | $1,362,420 | 2.04 |
| Aug 30, 2026 | 46.3 | 12.0 | 27.3 | 17.6 | $1,615,274 | 3.74 |
| Sep 6, 2026 | 42.6 | 20.0 | 27.3 | 20.0 | $1,431,233 | 2.36 |
The Labor Day week of 2025 printed XK of 20.3, a raw closing average that fell $282K week-over-week, and a (#C+#P)/XK of 2.04. This year: XK 20.0, a $184K raw drop, and 2.36. The fingerprint is nearly identical, and this year's version is milder on every line — more pendings, more closings, fewer new listings. Sellers who did not get their price over the summer pull listings the week after Labor Day; it is a calendar event, not a demand event. Across 35 years, September XK averages 27 per week; 20 is still well below the long-run norm.
I also tested every week since 1991 where XK spiked 40% or more above its prior four-week average while the 26-week price MA was rising and positive year-over-year: 49 instances. Average price change over the following 26 weeks was +3.5% and over 52 weeks +7.9%, with 76% of the 26-week outcomes positive. Single-week XK spikes inside a rising-price regime have not, historically, been the start of anything.
4. Accountability check: the Q4 crossing arithmetic has gotten harder
Last week I projected the smoothed (#C+#P)/XK would cross 3.54 into Q4 territory in late October–November and promised to tell you the week it happens. Here is the honest arithmetic. The 26-week window is now shedding its March–May 2026 readings, which were strong (they average 3.47 over the next 12 exits, including single weeks of 4.45, 4.24 and 4.13). To cross 3.54 by early November, raw prints would need to average about 4.5 for eight straight weeks. The raw 13-week average is 3.02. Scenario paths for the 26-week MA over the next 12 weeks:
| If raw (#C+#P)/XK averages | MA in 4 wks | MA in 8 wks | MA in 12 wks (late Nov) | Verdict |
|---|---|---|---|---|
| 3.0 (recent run-rate) | 3.16 | 3.08 | 3.01 | Drifts to mid-Q3; no crossing |
| 3.5 | 3.24 | 3.24 | 3.24 | Flat; no crossing |
| 4.0 | 3.32 | 3.39 | 3.47 | Approaches, crosses ~Dec/Jan |
| 4.5 (2025 fall pace) | 3.39 | 3.55 | 3.70 | Crosses week 8 (early Nov) |
So the trigger stands, the timing moves. Revised expectation: a Q4 crossing is a December 2026–Q1 2027 event under the most plausible path (raw settling near 4.0 as fall XK normalizes), not a late-October one. If raw stays near 3.0, the crossing does not happen and I will say so. The downgrade trigger — MA below 2.77 — is roughly eight bad weeks away and I will report it the week it prints.
5. The price MA has a base effect of its own
Full disclosure on the “new series high” headline: the 26-week price MA is about to shed some very strong spring closings (the next 12 exiting weeks average $1.59M raw, including one $1.95M week and two $1.89M weeks). If raw weekly closings average $1.55M this fall, the MA drifts to about $1.554M by late November; at $1.65M it edges to $1.60M. A modest MA dip this fall would be mechanical, not a reversal — and I am telling you now so it does not look like one later.
Conclusions
- Positive convergence: intact, narrower. Price momentum +2.2% YoY and accelerating; the higher-conviction predictor (#P/#N.L.) improving; both indicators in Q3, whose historical forward-52-week average is +7–8%.
- Negative divergence: none confirmed, one on watch. (#C+#P)/XK declined this week against a rising price MA. One week inside a known seasonal purge is not a divergence. Three consecutive weekly declines in the 26-week MA, or a break below 3.0, would be.
- Base case: +4% to +8% in Ave. $ of Closing over 26–52 weeks, held. Probability-weighted, I would now put roughly 60% on that band, 25% on flat to +4%, 15% on a decline, versus about 65/25/10 last week.
- Statistical validity: yes at the 99% level after effective-N correction, with the standing reminder that a valid signal is a tilt in the odds, not a guarantee.
Actionable Advice for Sellers
- Do not be the XK. Twenty listings a week are now expiring or cancelling, and the analog from 2025 says that number rises through September. Every one of those sellers priced to a hope. With #P/#N.L. at 0.64 — buyers absorbing nearly two of every three new listings — a correctly priced home is not the one that fails.
- The fall list window is now, not October. New listings dipped to 42.6 per week; you have less competition than at any point since early August, while pendings are steady at 27. Listing in the next three weeks puts you in front of the buyers who are transacting rather than the listings that are quitting.
- Price to the 26-week MA, not to last week's raw print. Raw weekly averages swung from $1.615M to $1.431M in seven days. That is composition, not value. Use the smoothed series and your street-level comps.
- If you relist after a cancel, change something. Price, presentation, or terms. Relisting into a 20-per-week XK cohort with the same package is how a home becomes a statistic twice.
Actionable Advice for Buyers
- Mine the expired and cancelled pool. Twenty listings came off this week; a meaningful fraction are motivated sellers who learned the market's price the hard way. That is where negotiated deals live this fall.
- The window is real but not wide. The predictors say prices are more likely up than down in 6–12 months. Seasonal softness this fall plus the mechanical MA dip in Section 5 is your best entry pricing before the spring 2027 cycle.
- Watch the two triggers I publish. A Q4 crossing of (#C+#P)/XK means demand quality is upgrading and leverage shifts back to sellers. A break below 2.77 means the opposite. You will read either one here first.
- Get pre-approved and pre-underwritten now. With 27 pendings a week against 43 new listings, the correctly priced home still gets multiple offers. Speed and certainty of close are your negotiating currency.
Devil’s Advocate: How This Could Be Wrong
- I may be explaining away a real signal. Calling the XK spike “seasonal” is comfortable. But the 2026 pre-Labor-Day XK (12.0) was unusually low, so the jump to 20 may be catch-up from deferred cancellations, and the 26-week MA of XK has now risen four straight weeks (13.9 → 14.4). If it keeps rising through October, the demand-quality ratio deteriorates regardless of calendar.
- The convergence is one-legged. #P/#N.L. improved partly because new listings fell, not because pendings rose (27.3, unchanged). A ratio can rise on a shrinking denominator; that is thinner supply, not stronger demand. If listings rebound to 48–50 in October and pendings stay at 27, the ratio drops back to 0.55 and the improvement evaporates.
- Effective-N of 23 is small. Thirty-five years of weekly data sounds enormous, but after correcting for autocorrelation it is closer to two dozen independent observations. A t-stat of 3.3 on 23 degrees of freedom is significant, but the confidence interval on the true correlation is wide — roughly 0.25 to 0.80. The lower end would make these indicators nearly useless.
- The 99.7th-percentile price level is not a neutral starting point. Every quintile forward return in the tables above was computed from lower price levels. Prices at an all-time high with mortgage rates well above the 2010s norm face an affordability ceiling that none of the historical Q3 analogs did.
- The 2025 analog cuts both ways. Yes, last September looked like this one. But from September 7, 2025, the 26-week price MA fell 9.4% over the following 26 weeks before recovering. The analog supports “this week is seasonal,” and also supports “a real winter dip follows.” I lean on the 52-week outcome (+2.2%); a skeptic would lean on the 26-week one.
- Raw week-to-week volatility is 14% of the mean. A single $9.75M Palos Verdes Estates closing (there was one on September 4) or its absence moves the raw average by $150K+. Averaging helps but cannot fully separate value from mix in a market this thin.
- Exogenous shocks are not in the model. A rate spike, an equity-market drawdown, or a local insurance or tax change can overwhelm any demand ratio. The indicators have a 26–39 week lead; policy can move in a day.
Net of all that: I still take the bet, but at a smaller size than last week. The right posture is to act on the base case and keep both triggers in view.
Methodology Notes
Dataset: 1,839 consecutive weekly observations of CRMLS daily activity averages (Monday–Sunday), May 24, 1991 through September 6, 2026, Palos Verdes Peninsula and South Bay, dual-pass verified (pandas + openpyxl, row and year-count assertions, zero duplicate weeks). All ratios recomputed from the raw counts and matched to the spreadsheet columns. Smoothing: 26-week simple moving averages. Momentum: 52-week percent change of the smoothed closing average. Lead-lag: Pearson correlation of indicator MA level with forward price momentum at 0/13/26/39/52-week leads; Bartlett effective-N correction for autocorrelation. Quintiles: equal-count bins of the indicator MA over the full history with mean forward 26- and 52-week price change. Projections are bounded by the dataset's own historical YoY extremes.
George Fotion • REALTOR®, Call Realty • California DRE# 00785373
Call/Text: (424) 722-9136 • george.fotion@homeispalosverdes.com
PalosVerdesHomesBest.com • SearchHomesInPrivate.com • Book a call: calendly.com/george-fotion
This video analyzes Palos Verdes and broader South Bay real estate absorption versus listing failures to determine whether the market is strengthening or weakening. The area saw 20 expired or canceled listings, the highest in five weeks, yet buyers absorbed 64% of new inventory, the best absorption since July, while the 26 week average closing price hit another series high at 1.574 million with momentum up about 2.2% year over year. Indicators show demand quality fell as failed listings rose, and the author moved a previously optimistic forecast trigger from late October or November to December or first quarter of 2027. The base case remains price growth of roughly plus 4 to plus 8% over the next 6 to 12 months, with lower conviction than last week.
