SolRadar track record

Methodology, in full: We snapshot every tracked token's trust score periodically while it's in our live window (see trust_score_history). For the low-score bucket, we take every token whose worst-ever recorded score was below the threshold, then check whether it eventually crashed 90%+ from its all-time-high market cap (ATH must have been at least $2,000, to filter out tokens that never took off in the first place). Same logic in reverse for the high-score bucket, using each token's best-ever score. Tokens still in the live window (no resolved outcome yet) are excluded from both counts entirely -- we don't count an open case as a hit or a miss either way. This is the exact same looks_rugged() definition used everywhere else on the site (Rug Autopsy, Wall of Shame), not a special looser or stricter bar for this page.