The record
Every project in this category claims accuracy. Almost none publish a record that could embarrass them. This one is computed live from the rulings themselves — including the ones that went wrong.
- Trials held
- 7
- Verdicts issued
- 7
- Verdicts edited
- 0
- Close calls
- 0
Why “verdicts edited” will always read zero. The verdicts table rejects UPDATE and DELETE at the database level, not by convention. A retrial inserts a new row that points at the one it supersedes; the old ruling stays visible forever. Neither an operator, nor a stolen key, nor a bug in this application can revise a sealed verdict.
Verdict distribution
Calibrated to be harsh. If PASS ever drifts far above target, the tribunal has stopped being useful and is drifting toward being a marketing channel.
Closest calls
Where the bench nearly split. These are the rulings most worth reading the dissent on.
What is deliberately not scored
Price. A condemned token that appreciates is not a miss, because no claim about price was ever made. Rulings address structure and extraction risk. Conflating the two is exactly how this category decays into selling signals, and being dull is preferable to becoming that.
Outcome checks at twenty-four and forty-eight hours measure what happened to liquidity, the deployer balance, and holder count — all mechanical, none editorial.