Three‑day slide stops; tiny servers produce the day’s biggest shocks

The single most unusual move yesterday came from Aspen University (pt‑BR). It averaged 189 players after averaging 11 the day before and has a seven‑day norm of 3 — a jump so extreme (+1,561.5% vs yesterday; +5,552.4% vs its own norm) that it dominates the ‘most unusual’ ranking. That scale of change on a tiny baseline is where leaderboard surprises come from: a reopened server, an accidental listing, or a short‑lived event can all produce these numbers, so persistence over the next few days will tell whether this is a genuine re‑entry or a one‑off spike. BAD TOWN! (th‑TH) is the other headline mover — it climbed from single‑digit averages to 192, is up four days straight and vaulted into the leaderboard (rank 297) from well outside the top 30k. Its rise produced one of the day’s largest rank gains and is genuinely disruptive for a server with a small recent baseline. Beyond those extremes there were several regionally consistent gains: Arabic and Thai servers (GraveRP, เฮฟวี่ซิตี้มาเเล้วจ้า, مدينة الـ90) added double‑digit player counts and sit above their 7‑day norms, which matches the regional share shifts we track below. On the downside, Corleone City — a server we’ve named recently — continued a decline, dropping to 156 average (down three days straight and roughly half its 7‑day mean). That fall is meaningful because it is a sustained move for an established, high‑visibility server rather than a tiny baseline wobble.

Several of the ecosystem’s most widely installed resources showed a coordinated small decline yesterday. Top packages — ox_lib, pma‑voice, oxmysql, bob74_ipl, ox_inventory, ox_target and others — each lost on the order of 90–160 listings and are flagged as down four days straight. The absolute numbers are small relative to their installed bases (the drops are mostly sub‑1% changes), but the breadth and consistency across unrelated packages is notable and not what we expect from random one‑day noise. Some of those resources still show modest positive seven‑day nets (oxmysql +33, pma‑voice +5) while others have negative seven‑day nets; the mixed seven‑day context suggests this could be short‑term churn rather than a structural shift. For script authors the practical takeaway is to watch whether removals continue into a fifth day: sustained multi‑day declines across core libraries would be the signal that something operational (packaging, distribution, or a scanner change) is affecting installs; a reversion over the next 48–72 hours would point to ephemeral churn.

First‑time and returning servers yesterday leaned Brazilian: several new pt‑BR listings appeared (BAIXADA SANTISTA #01SP, QUEBRADA SP INAUGUROU HOJE, REAL ROLEPLAY ➜ INAUGURAÇÃO HOJE), and pt‑BR now accounts for 20.4% of players — up from 14.7% a week ago. en‑US and th‑TH also gained share relative to seven days prior, keeping the regional picture tilted toward Brazil, Thailand and the US for the last week. There were no single large new entrants that shifted global totals; instead the day’s flavour was regional depth and local openings. No new script breakout appeared among first‑time servers and nothing in the new listings list looks, by itself, like a long‑term addition — again, persistence will be the deciding factor.

Where yesterday sits in the larger picture is simple: it ended the short string of daily declines but did not produce recovery to the week’s mean. The last 14 daily averages show frequent reversals; yesterday’s +2% is within that volatility and therefore not meaningful systemically. What is meaningful are the outliers: Aspen University’s extreme jump and BAD TOWN!’s rapid climb are the kinds of events that move ranks and attention quickly; if either holds over several scans they will reframe local leaderboards and regional share. Conversely, the four‑day softening across many top resources is an early warning signal — if that pattern continues it will be visible in average servers and slot fill; if it ends quickly it was likely temporary churn. For operators: don’t over‑interpret single scans on tiny baselines, but monitor the persistence of these spikes and the resource declines over the next 48–72 hours. For developers: a continued steady fall in installs across infrastructure libraries would merit an operational check (packaging, manifests or distribution channels); absence of continuation will simply read as noise.