How a verdict is made.
IdeaTribunal runs your idea through the v19.5 method: it gathers real evidence, checks the ten ways ideas die, runs nine hard gates, and hands back a brutal verdict with sources. The point is simple: kill a dead idea in a couple of minutes, before you sink 60 days and $5-10K into it.
How it works
One engine that runs through fixed stages in order. Demand is tested against real search numbers. Competitors, open source, pricing and regulators come from live web search, and every claim has a source link. Ten detectors check the idea from different angles, and the final judge runs on AI. A source the engine didn't open itself doesn't make the report.
Parse → demand → 10 killer detectors → web evidence (competitors / open source / pricing / regulators) → hard gates → score → verdict → final judge.
Ten killer patterns
Most ideas die from one of ten recurring patterns. Each is checked independently with cited evidence and a falsifier. Open the full catalog →
- K1Nothing unique
- K2No money in consumers
- K3The state already runs it
- K4Built into the platform
- K5Free open-source exists
- K6A free product owns the market
- K7High competition
- K8ChatGPT already does it
- K9No real pain
- K10Graveyard idea
Nine hard gates
These are strict pass/fail checks over the collected evidence. A failed check caps the verdict, no matter how high the score is. The score never overrides a failed check.
- G1Demand is strong enough to clear the bar (enough search volume, or a close match to a real market)
- G2There are at least some traces of the problem - people search for it or talk about it
- G3Payment & pricing demand
- G4No open-source or official incumbent owns the core job
- G5No foundation-model or platform replacement risk
- G6You can actually get the data the product needs
- G7Monetization fit
- G8No regulatory red flag
- G9The analysis rests on independent sources, not just the founder's word
A falsifier on every claim
Every piece of evidence carries a falsifier - one concrete fact that would cancel it out if it turned up. Each falsifier is graded STRONG / WEAK / REJECT. Lazy ones like "if the data shows otherwise" get thrown out on the spot; weak ones lower the report's confidence. That way the verdict is something you can check and push back on, not just take on faith.
Scoring
The score out of 100 is built from two parts. Strength (up to 54): how the idea does on the 8 hard checks. Each check it passes adds points (about 7 each), a failed one adds zero, an unknown one only a little. Evidence (up to 46): how solid the outside proof is (up to 22), whether the idea can survive AI and big platforms doing it themselves (up to 12), and how sharp the pain is for real people (up to 12). The score does not pick the verdict on its own: first the hard checks set the ceiling (the best verdict the idea can even reach), and only then does the score show how strong it is inside that ceiling.
The verdict
- NO-GOAlready dead. A hard gate failed - building it wastes 60 days.
- WATCHLISTNot now. Real blockers; revisit if the market shifts.
- TESTWorth a cheap experiment before you write code.
- GODemand + monetization + no killers. Rare.
- REFORMULATEToo vague to judge - the engine refuses to score air.