Calibrated ESO forecasts — scored in public
Probabilities for what Update 51 and the addon ecosystem will actually do, each with a fixed resolution date, an objective source, and the reasoning shown. When they resolve, the results — and our Brier score — publish here, right or wrong. No one else in game coverage scores their own predictions; that's the point.
register updated 2026-07-23Open forecasts (10)
78% Update 51's PTS opens on August 3, 2026 — the date ZOS re-announced on July 21 — and not later.
Resolves: 2026-08-04 · How it resolves: Official ZOS post or PTS patch notes on forums.elderscrollsonline.com; PTS live = the esoui/esoui GitHub 'pts' branch receiving a 12.1 dump also counts.
Why 78%: The date was already pushed twice (Jul 6 → Jul 27 → Aug 3), which argues for caution, but re-announcements made 13 days out with a stated reason usually stick. The July ZeniMax layoffs (213 ESO staff) are the main residual risk.
Registered 2026-07-23 · id u51-pts-aug3
60% Update 51 is live on PC by September 22, 2026 (the roadmap targets September 21).
Resolves: 2026-09-22 · How it resolves: Official live patch notes for PC/Mac on forums.elderscrollsonline.com dated on or before Sept 22, 2026.
Why 60%: An Aug 3 PTS start leaves ~7 weeks to the target — more slack than the usual 5-week PTS cycle needs — but the schedule has already slipped twice and the team just absorbed major layoffs. Slip risk is real; a one-week slide costs little internally.
Registered 2026-07-23 · id u51-live-by-sep22
88% Update 51 is live on PC by October 6, 2026.
Resolves: 2026-10-06 · How it resolves: Same source as u51-live-by-sep22.
Why 88%: Covers a two-week slip on top of the target. ESO major updates have historically shipped within a couple of weeks of their announced quarter-cadence dates even in bad years; a slip past early October would be an outlier requiring a new public delay announcement.
Registered 2026-07-23 · id u51-live-by-oct6
97% Update 51's APIVersion is 101051.
Resolves: 2026-08-10 · How it resolves: The header line of ESOUIDocumentation.txt in the U51 PTS dump (esoui/esoui GitHub 'pts' branch).
Why 97%: The pattern 1010NN == Update NN has held for roughly twenty consecutive updates (we verified 101038 through 101050 directly from the doc headers). The 3% covers a numbering scheme change, which ZOS has done before — once, a decade ago.
Registered 2026-07-23 · id u51-api-101051
85% The U51 API diff removes at least 20 global symbols (functions + events + constants) relative to API 101050.
Resolves: 2026-08-10 · How it resolves: Our published diff of ESOUIDocumentation.txt 101050 vs the U51 dump, using the same parser as the Patch-Break Radar (counts published on /addons/patch-radar).
Why 85%: Base rate from our own graveyard data: the last five measured cycles removed 22, 37, 33, 41 and 38 symbols — every one cleared 20. U51's announced hybridization completion (Mundus, buffs, class passives) points to more churn than average, not less.
Registered 2026-07-23 · id u51-removes-20plus
40% LibAddonMenu-2.0 — the library 1,333 addons depend on — ships a new release within 14 days of U51's PTS opening.
Resolves: 2026-08-24 · How it resolves: The last-update date on LibAddonMenu-2.0's ESOUI page (the public MMOUI feed we snapshot for /addons) showing a release within 14 days after the actual PTS start date.
Why 40%: The naive prior says critical libraries update instantly — but the data disagrees: LAM's last release was May 12, BEFORE Update 50 went live, and it simply didn't need a U50 update. Well-built libraries often survive cycles untouched. It updates within 14 days only if U51 actually breaks it.
Registered 2026-07-23 · id lam-updates-pts14
70% Between 6% and 14% of all ESOUI addons ship an update in the 30 days after U51 goes live.
Resolves: u51-live +30 days · How it resolves: Our /addons feed snapshot: share of all listed addons whose last-update date falls in [live date, live date + 30d]. Denominator = addons listed on the feed on the resolve date.
Why 70%: Measured base rate: 8.1% of addons updated in the 30 days after Update 50 went live (252 of 3,115). U51 is a bigger-churn patch (more forced updates) but the ecosystem's active-author pool is the binding constraint and it doesn't grow with patch size. The window is deliberately wide enough to be informative, not safe.
Registered 2026-07-23 · id addon-updates-30d
45% Necromancer · Stamina DPS is still the #1 DPS entry on our /meta ESO Logs pull 30 days after U51 goes live.
Resolves: u51-live +30 days · How it resolves: The top overall_dps entry of our /meta page's ESO Logs top-parse pull, re-run on the resolve date with the same method as today's.
Why 45%: Necro Stam only just took #1 (our Jul 24 pull: 883 vs Arcanist's 402; in June Arcanist led). Top-parse counts are volatile, U51's class balance is unknown until PTS, and a newly-dominant spec is exactly what balance passes target. Holding #1 through a major patch is closer to a coin flip than it looks.
Registered 2026-07-23 · id necro-stam-holds-no1
80% The #1 DPS entry on our /meta pull 30 days after U51 live is either Necromancer · Stamina or Arcanist · Stamina.
Resolves: u51-live +30 days · How it resolves: Same source as necro-stam-holds-no1.
Why 80%: These two have traded #1 all year and currently hold 883 and 402 points vs 238 for third place — a huge gap for a challenger to close in one patch. The 20% tail is a deliberate U51 over-buff of a third class (the announced Warden refresh is the obvious candidate).
Registered 2026-07-23 · id top-dps-necro-or-arc
45% Warden · Healer is the #1 healer entry on our /meta pull 30 days after U51 goes live.
Resolves: u51-live +30 days · How it resolves: The top overall_healer entry of the same /meta pull.
Why 45%: It's already a statistical tie (Warden 584, Arcanist 580 on Jul 24) — a coin flip before considering U51. The announced Warden class refresh cuts both ways: refreshes usually buff, but they also rework the exact passives top healers are built around.
Registered 2026-07-23 · id warden-healer-holds-no1
Track record
Nothing has resolved yet — the register opened July 23, 2026. Every forecast above has a fixed resolution date and source; results and Brier scores will appear here whether we were right or wrong. A registered probability is never edited after the fact.
Method
Each forecast passes a resolvability gate (objective source, threshold, and date fixed at creation), is anchored on an explicit base rate where one exists, and is scored with the Brier score after resolution. Probabilities are our honest odds, not advice. Created probabilities are never edited; material news gets a NEW dated entry superseding the old one, so the track record can't be quietly rewritten.
Inputs we rely on: our own ESO Logs meta pulls (dated, reproducible), the ESOUI feed snapshots, the Patch-Break Radar's API diffs, and ZOS's official announcements. "u51-live +30 days" resolution dates are pinned to the actual live date once ZOS ships.
FAQ
What is a calibrated forecast?
A probability meant to be scored, not to persuade. If our 80% forecasts come true about 80% of the time over many forecasts, we're calibrated. The Brier score measures this: 0 is perfect, 0.25 is what you'd get by always saying 50%.
Why publish probabilities about a video game?
Because everyone in game media makes predictions and nobody scores them. 'The meta will shift' is unfalsifiable; 'Necromancer Stamina holds #1 on this exact data pull on this exact date — 45%' can be checked, and we will check it publicly.
Where do the probabilities come from?
Base rates first (e.g. 8.1% of addons updated within 30 days of Update 50 — measured from the feed, not guessed), then evidence-based adjustment, then an honest look at what we don't know. Each forecast's reasoning is shown in full.