Loading…

Prediction evidence

correctPolicy v1.0

Source post

Published 20 May 2026, 14:57 UTC · fetched 3 Nov 2026, 13:33 UTC

Byzantine General

@ByzGeneral

In reply to · 20 May 2026, 14:56 UTC · context given to the AI

$BTC Could see a retest of 80k before lower prices. Weekly chart still looks cooked short term imo. https://t.co/ie614FLqMZ

@theo_crypto99 I could see a retest of 8k before even lower prices.

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

Baseline: first BTC price at or after publication, $77,362 at 20 May 2026, 15:00 UTC.
Evaluation: price at the end of the 24 hours horizon, $76,892 at 21 May 2026, 15:00 UTC.
r = 76891.99 / 77362.34 − 1 = -0.6%
Direction d = −1 (bearish), so d × r = +0.6% (neutral only if the price is unchanged) → correct

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: Bitcoin (BTC)
Direction: ▼ Bearish
Intent: explicit
Horizon: 24 hours (default, none stated)
Evidence span: “@theo_crypto99 I could see a retest of 8k before even lower prices.”
Confidence: 99%
Review state: auto accepted
Model / prompt: gpt-6-luna+decisions · decisions-v3