Sunday, 18 October 2026 · Europe/Berlin
Every call, accounted for.
Measured direction, magnitude and the reasoning behind both successes and failures.
Last 7 days: 0 calls, direction not measured, full success not measured. Last 30 days: 0 calls, direction not measured, full success not measured.
What counts as success?
A pick earns 1 when its direction is correct and the estimated open-to-close move is within the greater of 1 percentage point or 50% of the estimate. Otherwise it earns 0. A flat return fails the direction test.
For example, a +2% forecast succeeds between +1% and +3%. Close-to-close return is recorded separately. Price returns exclude fees, dividends and trading execution. Confidence is an uncalibrated research judgement. The 7 to 8 out of 10 target is an ambition, never a claim about results.
Daily record and calibration
A trend needs measured observations. No synthetic history has been added.
By category, sector and catalyst
Category statistics start when the first calls are scored.
Full forecast log
Evidence first. Picks follow.
No completed research batch for this session. Sources are collected before any forecast is published. Missed deadlines stay visible.
Open researcher accessResearch knowledge base
Original reasoning is preserved. Failures become lessons to review, never instructions to blindly reverse the call.
No completed patterns yet.
Pipeline verification
AI usage snapshot · 18 Oct, 21:38 CEST
AI usage snapshot · 18 Oct, 21:37 CEST
AI usage snapshot · 18 Oct, 21:36 CEST
Research operations and coverage
Daily plan: two pick batches and one watch batch. On-demand watch research is capped at three batches per day. Source feeds are cached for one hour, quotes for 15 minutes. AI usage is billed in energy. A measured estimate appears after completed automatic research is recorded.
Usage observed 18 Oct, 21:38 CEST: 8.672 energy today. Key admission cap 50.0 energy per billing day, resets 19 Oct, 02:00 CEST. Active work can still accrue usage.
Daily planning scenario: 8.26 energy for 10 watched stocks; 51.63 for 100. Two pick batches plus one watch refresh. The 100-stock figure scales the measured sample and has not been load tested.
Provider-reported USD planning scenario: $0.83 per day for 10 stocks; $5.16 for 100. Energy is the enforced budget unit. The 100-stock scenario exceeds the requested $5 daily budget.
- Watch · failed · 18 Oct, 22:57 CESTDaily paid execution cap reached, including retries
- Watch · done · 18 Oct, 22:57 CEST
- Watchlist · done · 18 Oct, 22:57 CEST
- Usage · queued · 18 Oct, 21:38 CEST
- Usage · queued · 18 Oct, 21:38 CEST
- Picks · done · 18 Oct, 21:33 CEST
- Picks · done · 18 Oct, 21:33 CEST
- Score · queued · 18 Oct, 21:33 CEST
Research requests: #8 watch pending, #7 watch completed, #6 picks completed, #5 picks completed, #4 watch completed, #3 watch completed, #2 picks completed, #1 picks completed.
Some source feeds are unavailable. Coverage is incomplete:
https://www.sec.gov/Archives/edgar/usgaap.rss.xml: HTTPError: 403 Client Error: Forbidden for url: https://www.sec.gov/Archives/edgar/usgaap.rss.xml