Scoring policy v1.0 · frozen

Methodology

The rules were frozen before any account was backtested or ranked. AI interprets what a post means; a deterministic engine decides whether it was right.

Core principle. A post such as “BTC looks ready to fly” counts as a bullish prediction. We measure prediction accuracy, not executed trades, profits, or investment returns.

Interpret the post

For each post the AI extracts the asset, direction (bullish or bearish), whether intent is explicit or inferred, any stated horizon, and the exact text supporting it, with a confidence score. Each asset in a post gets its own prediction record.

Only context available at publication is used: the post and the thread before it. Later replies and price data never reach the model. Post text is treated as data, never as instructions.

Excluded: historical commentary, pure questions, two-sided statements, sarcasm that cannot be resolved, and unresolved conditional calls. Extractions below 70% confidence go to manual review rather than being forced into a score.

Choose a comparable horizon

7 days · Explicitly stated weekly calls
24 hours · Default when the post states no timeframe

Horizons are never mixed in one ranking. Unsupported or vague horizons go to review. Profile windows select posts by publication date; evaluation may extend past the window, and unfinished predictions stay pending.

3 months6 months · gated until coverage is adequate12 months · gated until coverage is adequate

Resolve using price data

r = (evaluation price / baseline price) − 1

d = +1 bullish, −1 bearish

correct if d·r > t · incorrect if d·r < −t · otherwise neutral

The baseline is the first observation at or after publication, never before it. The evaluation price is the observation at the end of the horizon, within 60 minutes. Missing or stale prices leave the outcome unresolved; nothing is estimated.

Only the end-of-horizon price counts. A brief spike that reverses is not a success.

Neutral band t, frozen per asset and horizon
Asset	7d	24h
BTC	±2.50%	±1.00%
ETH	±3.00%	±1.25%
SOL	±4.00%	±1.50%

Example: bullish BTC

BTC is $60,000 at publication and $61,200 at 24 hours, a +2% move. With a 1% neutral band, d·r = +2% > 1%, so the prediction is correct.

Publish accuracy and rank fairly

Accuracy = correct / (correct + incorrect) × 100

Neutral, pending, unresolved and unscored counts are always shown separately. The directional-resolution rate, (correct + incorrect) / all price-resolved predictions, shows how often neutral outcomes were excluded.

Accounts are ranked by the 95% Wilson lower confidence bound on accuracy, with raw accuracy alongside, so 9 correct out of 10 (90%) ranks below 150 out of 200 (75%): the larger record is stronger evidence. At least 10 directional outcomes are needed for a rank; fewer are labelled provisional.

Each view also shows an always-bullish baseline: how a permanent bull would have scored on the same timestamps and neutral bands. Beating it matters more than a high number in a rising market.

Repeats, corrections and integrity

Same-account, same-asset, same-direction calls within a fixed 24-hour episode count once, anchored to the first eligible post. Later posts are kept as evidence and reversals are flagged.

Claiming a profile verifies ownership only. It cannot hide, remove or change predictions. Corrections go through a logged review and every affected snapshot is recomputed under the published policy.

Known edits are tracked with timestamps. Provider-required deletions are applied, and any resulting coverage limits are disclosed on the profile.