Most desks bolt weather onto the fundamentals. We build the fundamentals on a better weather forecast — European power & gas, Nordics to TTF to Henry Hub. Back-tested, source-cited, defendable.
One daily picture across the markets you actually trade — weather, hydrology and prices for the Nordics, Germany, France and Spain, wired to the TTF gas curve and the Henry Hub · LNG chain. The edge is in the extremes that move the book: Atlantic hurricanes, cold-snap freeze-offs like Winter Storm Fern (Jan 2026), heatwaves and hydro drought — the weeks the EC ensemble smooths to climatology while our 45-year analog atlas still finds structure. That read converges with the fundamentals, the energy & geopolitical news flow, and the street’s price consensus against our own view. The Nordics is our deepest book; Germany, France, Spain and the US gas chain are live and expanding. Every figure back-tested and source-cited — numbers, not adjectives. Private beta, a small institutional client base.
Three things you can’t get on Bloomberg, Refinitiv, or any other AI-news layer
01
A proprietary weather engine — not a news feed
DIANA's own 45-year atmospheric atlas + analog-cohort matcher produces W+0..W+5 forecasts for NP / DE / FR / ES that beat ECMWF at every lead from W+1. Where EC's ensemble has converged on climatology, the atlas still sees structure — back-tested, not asserted.
Hydrology + fundamentals, every claim source-cited
Reservoirs, water value, river-flow risk, cross-border spot and gas → TTF → power — the physical supply/demand drivers behind the price, each traced to a named source (EIA · IEA · Wood Mac · NVE · ENTSO-E), never a black box.
28 years of empirical research built in-house — the Select / NoGo analog cohorts, cross-regional correlation structure, climate-drift findings, and an engine atlas nobody else can reproduce. The all-sides news layer rides on top as context — never the headline.
Six trader-shaped use cases. Each card maps to a daily-refreshed surface inside the portal — not a marketing promise. We don’t compete with Bloomberg on intraday execution; we compete on defensibility: every claim traces to a named source or a backtested cell.
Directional
Multi-week edge over ECMWF
Where EC's ensemble has converged on climatology, DIANA's analog atlas still sees structure.
W+0..W+5 calls on Wind, Solar, Precip, Temp across NP / DE / FR / ES — daily refresh.
Trust Verdict (layer agreement) + Risk Check (cone width + engine drift) gate the trade before you size.
Hit-Score per month × per lead — see the proof strip above for the actual delta on the 2020-2024 walk-forward backtest.
Where natural hedges actually live across the four markets, by season — and which folklore broke under climate drift.
Pearson-r correlation matrices per variable per season (Wind / Precip / Solar / Temp × DJF / MAM / JJA / SON).
Lead → follower conditional probability: when NP+DE both go bullish/bearish, what does FR / ES historically do?
Folklore audit on 28 yrs (1997-2024) names which trader-rules HOLD vs which were broken by the post-2018 climate-shift era.
Production
Production & dispatch planning
Zone-level energy forecasts down to the 9 Nordic price areas — usable for reservoir, dispatch, hedging on the production side.
Hydro precipitation energy + wind + solar at NO1-NO5 + SE1-SE4 sub-areas plus continental DE / FR / ES.
Reservoir 6-month forward + cooling-season demand windows + heat & drought risk cards (Pre / Peak / Late Summer per market).
Climate-drift-corrected normals — Era-A (1997-2010) vs Era-C (2018+) stratification so 2025+ dispatch isn't anchored to the wrong baseline.
Geopolitical
Geopolitical & news intelligence
Empirical rhetoric × price linkage. All-sides perspective tracked side-by-side because tone shifts move prices regardless of factual reliability.
~63 approved sources across wire, official, specialty, analyst, research, and all-perspective state media (Fox / MSNBC / RT / Al Jazeera / PressTV / Global Times).
Per-source perspective chips + mandatory dissent line — observation never asserted as fact.
Research
Source-cited fundamentals + price view
Every claim traces to a named institutional source. Black-box assertions don't survive review.
Synthesis across EIA · IEA · Wood Mac · RBC · OIES · Argus · Montel — dozens of named institutional references per document on Watch → Fundamentals.
TradeWPower variant view sits ABOVE the live spot tiles — opinion first, commodity context second. Day-ahead spot stack across NP / Baltic / DE.
Calibration trail per estimate — every revision is timestamped with the reasoning, so you can audit how the view evolved.
Expertise
Analyst expertise · Library · knowledge base
Two named founders + a weekly video Library + the deepest empirical knowledge base in EU power.
Daily morning brief authored by a named analyst (not AI synthesis), with a mandatory 'what we're wrong about' dissent line.
A weekly video Library — Ivan's market views (daily + weekly), DIANA methodology walkthroughs, and a curated knowledge archive of the framework.
The /strategy page consolidates 28 years of empirical patterns — correlation matrices, persistence highlights, climate-drift findings, and the trader-folklore audit — the analytical authority backdrop competitors can't replicate.
US gas & LNG
US weather risk → Henry Hub · LNG · TTF
The same engine, pointed at the US gas complex: cold freeze-offs, summer cooling demand, and Atlantic hurricane risk translated to explicit energy calls.
Cold / heat / hurricane risk per US energy sector (Gulf-LNG, Permian, Haynesville, ERCOT, …) with per-cell leave-one-year-out skill — no made-up numbers.
DIANA's read placed beside NOAA · CSU · ECMWF, plus the institutions' own live maps (NESDIS tracker · NHC outlook · NWS HeatRisk) — the disagreement is the trade.
Freeze-off → HH spike analogs (Winter Storm Fern · Uri · Elliott) and the EU↔US link: a pure-US freeze-off the market may misprice as a European cold spillover.
TradeWPower's expertise covers every aspect of each regime — hurricane (genesis, steering, rapid intensification, landfall, LNG exposure), winter cold-air outbreaks and summer heat; the on-page skill numbers come from an automated engine that keeps developing and learning.
Atlantic track history (HURDAT2) × US LNG export clusters — one of the DIANA-US maps.
Navigate the data
Ask DIANA — talk to the data, not a brochure
15+ reports, 100+ packet fields, 28 years of empirical patterns — that’s a lot to navigate on a tight trading day. Ask DIANA is a Claude-powered analyst chat that reads directly from today’s JSON packets and answers in seconds, with the exact source field cited on every reply.
What it does
Cross-references every report
One question pulls from R1 / R2 / R4 / R3 / Info / Strategy / News / Fundamental / Price as needed. No need to remember which packet holds which number.
Every answer cites the source field
Replies carry a citation chip naming the exact packet location (e.g. r4.week_details[2].targets.NP_Wind). Auditable end-to-end.
Honest about gaps
If a number isn’t shipped in today’s packet, DIANA says so — no fabricated values, no hallucinated trends.
Hi — I'm DIANA's analyst chat. Ask me anything about today's reports. I read directly from the JSON packets and cite the exact field for every number. If a value isn't shipped in the packet, I'll tell you.
Try asking
What's W+2 NP_Wind looking like today?Why is W+3 DE_Solar LOW conviction?Compare today's Sel cohort to last year'sShow me the Reservoir Risk Trajectory for NO2Which sub-areas show a Precip deficit through Q4?Where does the engine disagree with EC this week?
Preview — live chat unlocks after access approval.
Request access
Private beta with a small institutional client base. Tell us who you are and what markets you trade — we’ll reply by email. Direct line: post@tradewpower.no.
DIANA exists for the analyst preparing a morning brief, defending a view to a client, or back-testing whether news bias actually moved a market on event days. We don’t compete with Bloomberg or Refinitiv on intraday execution. We compete on defensibility — every number on every page traces to a primary source URL, the methodology is open, and the backtest record is published with its limitations stated.
