[Study] Used DDR4 Aegis 16GB: Ask to Cut to SOLD — Case File #1
Decision details
Fits if: Readers who want to understand how used clearing prices actually move from ask to sale — and are willing to read a study's limits.
Main drawback: n=1: one kit, one board, one period. The SOLD marker is the seller's own marker, not a verified transaction receipt. No market-wide conclusion can be drawn from this file alone.
Compare: two dated readings of the same object; every reading proposed to the fact registry; the SOLD marker recognised as a seller marker; the comparison band carried with its status (relay n=1) separately; arithmetic done on the named readings.
In this guide
Welcome to the first case file of the [Study] series — betulkex’s format for dated first-party observation, not advice. The first object: one G.Skill Aegis 16GB (8x2) DDR4-2666 kit on Lowyat Garage Sales. Why this one? Because it is rarely possible to record in full: from asking price, to cut, to SOLD — all visible in a public thread.
What is the study object and why does it qualify?
The object: a Lowyat thread titled “WTS DDR4 Gskill Aegis 2666 8gb x 2” (forum.lowyat.net/topic/5576558) — a used 16GB DDR4-2666 kit, COD deal, PJ location. It qualifies as the first object because its movement is fully readable from outside: an asking price that held, a cut in the seller’s own words, then the SOLD marker in the thread. Most sellers delete prices once sold — this one left a trail.
What counts as one ‘reading’ in our protocol?
The series’ observation protocol (named so others can check it):
- One reading = one thread-visible price row OR one SOLD marker, captured together with its date.
- Who observes: our observation lane (not the seller, not the buyer).
- Where: a named public thread — for this file, topic 5576558 on Lowyat Garage Sales.
- How: board-index readings (diff between captures) + readings of the thread itself.
Two readings of this object: baseline 2026-10-01 and the SOLD reading 2026-10-02. One object, two lenses — only then does it earn the word study.
Timeline: how did this price move?
| Reading | What was visible | Date |
|---|---|---|
| Baseline | "Aegis 16GB kit cut RM250->RM225 with 1.8k views in 22 days unsold" (mining-dr D2 FINDINGS, 2026-10-01: original RM250 ask held 22 days, 1.8k views, unsold; the cut to RM225 happened before baseline) | 2026-10-01 |
| Cut (in thread) | "Price reduce to RM225 for both" | in-thread, before baseline |
| SOLD | "SOLD......." (reply in the thread) — answered 2026-09-30 per the index, captured at the 2026-10-02 reading | 2026-10-02 |
[source] [source] [source] [source]
Arithmetic: where does the clearing price land versus the ask?
The arithmetic below is not our own figure — it is an observation bank row quoted byte-exact (every input named):
“Clearing price RM225/16GB kit = RM14.06/GB used DDR4-2666 vs same-day ask band RM18.75/GB (plain 16GB) — clearing sits ~25% below ask, inside the projected asking-minus-15-40% band.” (vein4_ram_window_c1.jsonl movement row, 2026-10-02)
Read that arithmetic with each input’s source: the RM225 clearing price (seller marker, above), the same-day comparison ask strip (bank row), and the 15-40% projection band (status below). For cross-reference, the REGISTERED ask band: the DDR4 asking median, snapshot 2026-10-01 [source].
What is the 15-40% band, and why do we label it n=1 too?
Layered honesty: the “asking-minus-15-40%” band inside that quote is a projection from DR analysis — and the related relay record in the fact registry (the ask-vs-clearing gap on one premium B-Die case) is carried with its exact label: relay, n=1 — one case confrontation, a friction indicator, not a market-wide discount rate [source]. A relay row cannot carry [Study] weight without a second read — that is the doorway to this study too: the second reading of this Aegis object is what made it qualify.
What does this study NOT prove?
The series’ mandatory section — written before the conclusion, not as a footnote:
- One kit only — no generalisation to other kits, other speeds, other sellers.
- One board only — Lowyat; Carousell, Shopee and shops have their own dynamics.
- The SOLD marker is the seller’s, not a verified transaction — we know the final printed price and the seller’s word that it sold; we do not see the receipt.
- No market-wide discount rate claimed — “~25% below ask” is arithmetic on ONE object; the relay band is carried at n=1.
- No price-direction forecast — this file is rearview design, not a front lens.
The DR used-DDR4 guidance band stays as context with its status [source] — DR-sourced guidance, not transactions we measured ourselves.
What is the series’ second object?
The object already named in the series list: the WTB-below-WTS spread on the 64GB kit class — the gap between what buyers’ ads want to pay and what sellers’ ads want, in the same class, on the same day. When both halves meet, we get a true band; when they do not, the gap itself is the story. The 64GB price page now carries that field. If you see a candidate object (a thread whose price moves are readable from outside), send it via WhatsApp below — the entry criteria: two readings must be possible, the public thread must be nameable, no seller personal information lifted.
The neighbouring pages own the other questions: today’s used price band at used DDR4 RAM Malaysia (band owner); the wait-or-buy decision at wait or buy RAM 2026 (BM) (decision owner); the 32GB class lens at 32GB DDR4 kit used (BM). This file owns exactly one thing: dated observation discipline.
Read also: DDR4 vs DDR5 · used RAM prices · used DDR4 RAM Malaysia
Frequently asked questions
How is this [Study] page different from a normal used RAM price page?
A price page (like used DDR4 RAM Malaysia) holds the market band on one day. A [Study] follows ONE object across time with dated readings — it answers 'how did this one item's price move', not 'what is the market price today'. The two link to each other; they do not overlap.
How many readings before a [Study] can publish?
At least two readings of the same object — one reading is an anecdote, not a study. This file has two: the 2026-10-01 baseline (from the mining-dr D2 finding: a kit asking RM250 held with 1.8k views over 22 days unsold, then cut) and the 2026-10-02 SOLD reading (thread named in the page).
Is the SOLD price a verified transaction?
No. The 'SOLD.......' marker in the thread is the SELLER's own marker with the final price printed — not a verified sale receipt (same method as the other sold-basis rows in the fact registry [source]). We write the limit in the page because it is the true limit of the reading.
So if I sell DDR4, should I cut 25%?
No — that is a misreading of this file. The comparison figures in this study come from ONE object only; the wider relay comparison band (ask-vs-clearing gap of 63-70% on one premium B-Die case) is also n=1 status and DR-sourced [source]. No market-wide discount rate is claimed anywhere on this page.
What is the series' next object?
The next object named in the series list: the WTB-below-WTS spread on the 64GB kit class — when a buyer's 'want to buy' ad and a seller's 'want to sell' ad do not meet on price, that gap is itself an interesting object. Other object proposals are welcome via WhatsApp below.
Why isn't the Carousell 18.75/GB ask band printed as our number?
Because of its status: that comparison was made in the bank row (quoted on the page) against a reading not yet in the fact registry — until admission, it is displayed as a dated quote with its source, not as our own figure. The REGISTERED Carousell ask band (median RM450, snapshot 2026-10-01) is printed with its key [source].