THE PULSE WEEKLY REPORT

Rolex market: a data-quality week, a quiet market

Week of August 10–16, 2026 · The Pulse Rolex Index · thepulse.fun
The Pulse Index
98.63
−0.36% this week
Average watch
98.39
−0.67% this week
Typical sale price
$14,848
from $15,032
Sales pace
140/day
below 185 baseline

The week in one paragraph

A quiet week on price — the Pulse Index eased −0.36% to 98.63, the average watch −0.67% to 98.39 — but a busy one behind the scenes. Our upstream data provider (Cyclops) discovered and fixed a real defect in how Chrono24 search results were being collected: requesting result pages beyond a reference’s actual listing count was silently pulling in unrelated watches under the wrong reference code. Roughly 500 contaminated listings were identified and excluded database-wide. One day of our series — August 13 — had enough of this contamination in an unverifiable, title-less state that we could not clean it with confidence, so we are treating it as a permanent data gap rather than publish an unreliable number. Every other day this week, including the corrected version of August 10–12, is intact. The underlying market read: 5 of 14 watches gained on the week, trading pace ran below the 185/day baseline most days, and Thursday (Aug 14) and the weekend saw the week’s only above-average sessions.

🔧 Data quality note: This week we caught and corrected an upstream contamination issue that had been quietly inflating our reference basket for several days. Three previously-published index values (Aug 10–12) were revised via errata after the root cause was confirmed. Aug 13 has no index value and will remain a permanent gap in the series, the same treatment we’ve given past data outages. The market's underlying trend was not materially affected — the corrections moved the index by roughly 0.2–1.1 points on the affected days, not the direction of the story.

Movers of the week

Watch7-dayRead
Submariner Date (126610LN)+8.9%The flagship Sub firming up; clean, well-sampled move
Sub Date prev gen (116610LN)+5.3%Discontinued-icon demand holding; tracks the current-gen move
GMT-Master II Pepsi (126710BLRO)+3.8%Hype-sensitive GMT ticking up on modest volume
Datejust 41 Smooth (126300)−8.1%Entry-tier softening after a strong prior week
GMT Batman prev gen (116710BLNR)−7.4%Giving back gains; discontinued-GMT demand cooling
Explorer 36 (124270)−23.3%⚠️ Not a market move — this reference briefly carried a stale, condition-mix-elevated price ($11,400) forward across two thin-sample days before a fresh, larger sample corrected it back to $8,742 on Saturday. The reference's true week-over-week move is closer to flat.

Liquidity read

Confirmed sales averaged 140/day this week, meaningfully below the 185 baseline — though the missing Aug 13 datapoint slightly understates the true average since that day's real sales volume isn't counted. The week's shape was uneven: Saturday (Aug 15) was the standout session with 193 confirmed sales and $3.43M in dollar flow, more than double the week's other days. Sunday and Monday (Aug 16–17 data) were quiet, consistent with the usual weekend pattern. Dollar flow averaged $2.09M/day across the week's six valid days.

What actually happened with the data this week

For readers who want the full story: our Chrono24 data pipeline sizes how many search-result pages to request per watch reference. When that sizing ran too high, Chrono24's search index began returning padding — extra, unrelated watches — on the pages past a reference's true result count. Those foreign watches got silently counted as sales of the reference being searched, sometimes at very different prices. Our data provider identified the pattern, tagged roughly 500 affected database rows, and fixed the underlying page-sizing logic (now sized exactly to each reference's current listing count, with no padding buffer). We added a matching filter on our end and re-verified three prior days (Aug 10–12), correcting each by roughly 0.2–1.1 index points. One day, August 13, had enough untitled, unverifiable data that neither our provider's fix nor our own filters could clean it with confidence, so it remains a permanent gap. As of Saturday (Aug 16), Cyclops reports the first fully clean sweep cycle with zero contaminated rows in the day's data — we expect no further corrections going forward.

Methodology notes: Prices are market values (asking prices adjusted by the observed ask-to-sold gap), computed from ~14,500 tracked listings with manipulation screens, evidence-weighted updating, and mix-shift controls. A sale counts only after it is verified missing across two consecutive daily sweeps. Index values for Aug 10–12 were revised this week via errata after a confirmed upstream data-quality fix; Aug 13 has no published value and is treated as a permanent gap. Informational only, not investment advice. © The Pulse · thepulse.fun