Prediction evidence

incorrectPolicy v1.0

Source post

Published 4 Sept 2026, 20:04 UTC · fetched 3 Nov 2026, 11:01 UTC

Cheds Trading

@BigCheds

In reply to · 4 Sept 2026, 16:22 UTC · context given to the AI

$ETH with its longest range ever. Think bigger than just $10k. https://t.co/2BFAPNkDFl

@MaeliusCrypto Problem is the longer the sideways the more likely it is to actually drop

Deterministic: the same prices and policy always give the same result

Baseline: first ETH price at or after publication, $2,451 at 4 Sept 2026, 21:00 UTC.
Evaluation: price at the end of the 24 hours horizon, $2,478 at 5 Sept 2026, 21:00 UTC.
r = 2477.86 / 2451.47 − 1 = +1.1%
Direction d = −1 (bearish), so d × r = -1.1% (neutral only if the price is unchanged) → incorrect

Only the end-of-horizon price counts. A brief move followed by a reversal is not a success. Read the methodology

Uses only text available at publication time

Asset: Ethereum (ETH)
Direction: ▼ Bearish
Intent: inferred
Horizon: 24 hours (default, none stated)
Evidence span: “@MaeliusCrypto Problem is the longer the sideways the more likely it is to actually drop”
Confidence: 98%
Review state: auto accepted
Model / prompt: gpt-6-luna+decisions · decisions-v3