The cheapest listing usually sits at the median, and wide spreads mark busy markets
When you open the auction house, the first price you see is the cheapest listing, but the fair price is closer to the median of everything listed. We asked how far apart those two numbers usually are, and what it means when the gap is wide. The answer changes how you read a cheap listing: is it a bargain in a dead market, or the front line of an undercutting war?
The question
Section titled “The question”For every item we scan, there is a cheapest listing and a median listed price. The spread is the gap between them. We wanted to know two things. First, how far under the median the cheapest listing usually sits. Second, whether a wide spread marks a thin market, meaning a handful of listings where one stray price skews everything, or a pushed market, meaning a deep, busy book where sellers undercut each other to be first in line. A useful answer gives the typical gap, says how often the gap gets dramatic, and ties wide gaps to how much an item actually trades.
The short answer
Section titled “The short answer”- The cheapest listing usually does not sit under the median at all. On the newest Horde scan, the cheapest listing for the typical item sits at 100% of its median, right on it. Only for one item in four does the cheapest fall below 87.2% of the median.
- Deep discounts are rare: the cheapest listing is under half the median for just 6.8% of Horde items and 4.9% of Alliance items.
- A wide spread is a busy-market sign, not a thin-market one. Among busy Horde items, 13.5% have a cheapest listing more than a quarter under the median, against 7% of quiet items. Busy and quiet are defined in the method below.
Method
Section titled “Method”For each priced item we took the cheapest listing as a percent of the same scan’s median listed price. We call this the cheapest percent: 100% means the cheapest listing is the median, 50% means it is half the median.
The headline numbers come from the newest scan of each market on 8 November 2026: 3,751 priced items on PvP Horde and 3,879 on PvP Alliance. To check that one scan is not a fluke, we repeated the measure on every day of our own scans, 18 Horde days and 20 Alliance days between 8 October and 8 November, keeping only items priced on at least 3 of those days: 3,738 items on Horde and 3,821 on Alliance. The picture is the same on every single day: the daily median of the cheapest percent is 100% on all 38 market days.
Thresholds, and why. Wide means the cheapest listing is under 75% of the median, more than a quarter below it; that cut lands in the bottom sixth of items, far from normal undercutting of a few percent. Busy means at least 20 estimated sales a day, which is roughly the top quarter of Horde items by the sold-a-day estimate. Quiet means at most 1 sale a day, roughly the bottom quarter. Sold a day exists only in the newest scan, so the busy and quiet labels come from that one day.
The thin-versus-pushed verdict rests on three checks, all pointing the same way. By trading rate: wide spreads are about twice as common among busy items as quiet ones (13.5% against 7% on Horde, 13.2% against 9.9% on Alliance). By depth of the book: among Horde items with 200 or more units listed, 17.5% are wide, against 7% of items with 9 or fewer units. And by persistence: the items that stay wide on at least half their scan days sell about twice as fast and carry about twice the listings as items that are never wide (medians of 6.9 against 3.6 sold a day, and 20 against 10 units listed, on Horde).
Part of this is mechanical, and it is worth saying plainly: when an item has one or two listings, the cheapest listing often is the median, so a thin market cannot show a wide spread. Bow of Searing Arrows is typical of the quiet end: two listings, both at 250g, no sales seen. The wide end looks like Wool Cloth, which sells around twelve thousand units a day and had its cheapest listing at 68c against a 1s 34c median, or Light Feather, cheapest 7c against a 28c median. Those are undercutting wars in the deepest books on the server, the same items you see churning in The Terminal.
Charts
Section titled “Charts”Each bar is a share of the 3,751 priced Horde items in the newest scan. For 56% of items the cheapest listing is the median, and about seven items in ten sit at 90% of the median or better. The tail below 75% holds roughly one item in six.
Items are bucketed by estimated sales a day. Wide spreads are rarest where nothing sells (6.4%) and most common in the middle of the market, items selling 5 to 20 a day (23.3%). The very busiest items tighten up again, because the gap between the cheapest and the median gets bought and relisted constantly.
Day by day, the share of wide-spread items stayed between about 8% and 18% on both factions. It was highest on 5 November, right after the scan gap, when the fewest items were priced and the book was still settling, and it eased every day after.
Among Horde items priced on at least 3 of our scan days, 43.2% were never wide, 51% were wide on some days but fewer than half, and only 5.9% were wide on at least half their days. A wide spread is usually a passing event, a dump or an undercut that gets bought, not a standing feature of the item.
What would change this
Section titled “What would change this”Three later results would overturn parts of this answer. First, if the quiet quarter of items started showing persistent wide spreads across many scan days, the thin-market story would come back; right now quiet items almost never can, since one seller’s listing is both the cheapest and the median. Second, the busy and quiet labels rest on a sold-a-day estimate from a single scan; if estimates from several more weeks of scans moved items between buckets and the two-to-one gap between busy and quiet wide shares closed, the third bullet would weaken. Third, the November days show the widest spreads of the whole window; if that share kept climbing instead of easing as the book deepens, the claim that spreads sit in a stable 8% to 18% band would be wrong. More days of full scans answer all three.
Limits of beta data
Section titled “Limits of beta data”This is beta data under a level 30 cap, with no raids and a small population, and the beta economy is wiped, so none of these prices are launch prices. What tends to carry over is the order of things, which items trade thick and which thin, not any number on this page. Sixteen of our history days are daily backfill covering only about 320 items with no sales figure, so every multi-day result here uses our own full scans only, 18 Horde days and 20 Alliance days. Each day is represented by its last scan, and a single scan is a snapshot: a wide spread that appears and vanishes within an afternoon can be missed or overcounted. Sold a day is an estimate, exists only in the newest scan, and should be read as a rank, not a count. There were no scans between 22 October and 4 November, so the November days describe a freshly restocked auction house. How we scan and what the fields mean is on the methodology page, and the earlier look at thin markets is in the thin market report.