How we test Knos
Every number Knos shows comes from a run you can open.
What each Qloo signal changes
65 removals on 6 recorded samples: 24 change the city, the target or the verdict.
The table
| Sample | Removed | Launch city | Target | Verdict |
|---|---|---|---|---|
| a crypto podcast for beginners | search (the matched products) | none → none | none → none | REFUSE → REFUSE |
| a crypto podcast for beginners | demographics | none → none | none → none | REFUSE → REFUSE |
| a crypto podcast for beginners | audiences | none → none | none → none | REFUSE → REFUSE |
| a crypto podcast for beginners | trending | none → none | none → none | REFUSE → REFUSE |
| a crypto podcast for beginners | recommendations | none → none | none → none | REFUSE → REFUSE |
| a crypto podcast for beginners | destinations | none → none | none → none | REFUSE → REFUSE |
| a crypto podcast for beginners | places | none → none | none → none | REFUSE → REFUSE |
| a crypto podcast for beginners | tags (persona) | none → none | none → none | REFUSE → REFUSE |
| a crypto podcast for beginners | every Qloo answer | none → none | none → none | REFUSE → CANNOT_DECIDE |
| a probiotic kombucha brand | search (the matched products) | none → none | none → none | OPEN → REFUSE |
| a probiotic kombucha brand | demographics | none → none | none → none | OPEN → REFUSE |
| a probiotic kombucha brand | audiences | none → none | none → none | OPEN → REFUSE |
| a probiotic kombucha brand | compare (the leader) | none → none | none → none | OPEN → REFUSE |
| a probiotic kombucha brand | trending | none → none | none → none | OPEN → OPEN |
| a probiotic kombucha brand | recommendations | none → none | none → none | OPEN → OPEN |
| a probiotic kombucha brand | destinations | none → none | none → none | OPEN → OPEN |
| a probiotic kombucha brand | places | none → none | none → none | OPEN → OPEN |
| a probiotic kombucha brand | tags (persona) | none → none | none → none | OPEN → OPEN |
| a probiotic kombucha brand | brand ladder | none → none | none → none | OPEN → OPEN |
| a probiotic kombucha brand | every Qloo answer | none → none | none → none | OPEN → CANNOT_DECIDE |
| a quantum physics book for kids | search (the matched products) | none → none | none → none | REFUSE → REFUSE |
| a quantum physics book for kids | demographics | none → none | none → none | REFUSE → REFUSE |
| a quantum physics book for kids | audiences | none → none | none → none | REFUSE → REFUSE |
| a quantum physics book for kids | compare (the leader) | none → none | none → none | REFUSE → REFUSE |
| a quantum physics book for kids | trending | none → none | none → none | REFUSE → REFUSE |
| a quantum physics book for kids | destinations | none → none | none → none | REFUSE → REFUSE |
| a quantum physics book for kids | places | none → none | none → none | REFUSE → REFUSE |
| a quantum physics book for kids | tags (persona) | none → none | none → none | REFUSE → REFUSE |
| a quantum physics book for kids | brand ladder | none → none | none → none | REFUSE → REFUSE |
| a quantum physics book for kids | every Qloo answer | none → none | none → none | REFUSE → CANNOT_DECIDE |
| a recycled ocean plastic swimwear brand | search (the matched products) | none → none | none → none | OPEN → REFUSE |
| a recycled ocean plastic swimwear brand | demographics | none → none | none → none | OPEN → REFUSE |
| a recycled ocean plastic swimwear brand | audiences | none → none | none → none | OPEN → REFUSE |
| a recycled ocean plastic swimwear brand | compare (the leader) | none → none | none → none | OPEN → REFUSE |
| a recycled ocean plastic swimwear brand | trending | none → none | none → none | OPEN → OPEN |
| a recycled ocean plastic swimwear brand | recommendations | none → none | none → none | OPEN → OPEN |
| a recycled ocean plastic swimwear brand | destinations | none → none | none → none | OPEN → OPEN |
| a recycled ocean plastic swimwear brand | places | none → none | none → none | OPEN → OPEN |
| a recycled ocean plastic swimwear brand | tags (persona) | none → none | none → none | OPEN → OPEN |
| a recycled ocean plastic swimwear brand | brand ladder | none → none | none → none | OPEN → OPEN |
| a recycled ocean plastic swimwear brand | every Qloo answer | none → none | none → none | OPEN → CANNOT_DECIDE |
| a solar powered music festival | search (the matched products) | Las Vegas-Henderson-North Las Vegas, NV → Las Vegas-Henderson-North Las Vegas, NV | 23 → 23 | OPEN → REFUSE |
| a solar powered music festival | demographics | Las Vegas-Henderson-North Las Vegas, NV → Las Vegas-Henderson-North Las Vegas, NV | 23 → 23 | OPEN → REFUSE |
| a solar powered music festival | audiences | Las Vegas-Henderson-North Las Vegas, NV → Las Vegas-Henderson-North Las Vegas, NV | 23 → 23 | OPEN → REFUSE |
| a solar powered music festival | compare (the leader) | Las Vegas-Henderson-North Las Vegas, NV → Las Vegas-Henderson-North Las Vegas, NV | 23 → 23 | OPEN → REFUSE |
| a solar powered music festival | trending | Las Vegas-Henderson-North Las Vegas, NV → Las Vegas-Henderson-North Las Vegas, NV | 23 → 23 | OPEN → OPEN |
| a solar powered music festival | recommendations | Las Vegas-Henderson-North Las Vegas, NV → Las Vegas-Henderson-North Las Vegas, NV | 23 → 23 | OPEN → OPEN |
| a solar powered music festival | heatmap | Las Vegas-Henderson-North Las Vegas, NV → none | 23 → none | OPEN → OPEN |
| a solar powered music festival | destinations | Las Vegas-Henderson-North Las Vegas, NV → Las Vegas-Henderson-North Las Vegas, NV | 23 → 23 | OPEN → OPEN |
| a solar powered music festival | places | Las Vegas-Henderson-North Las Vegas, NV → Las Vegas-Henderson-North Las Vegas, NV | 23 → 23 | OPEN → OPEN |
| a solar powered music festival | tags (persona) | Las Vegas-Henderson-North Las Vegas, NV → Las Vegas-Henderson-North Las Vegas, NV | 23 → 23 | OPEN → OPEN |
| a solar powered music festival | brand ladder | Las Vegas-Henderson-North Las Vegas, NV → Las Vegas-Henderson-North Las Vegas, NV | 23 → 23 | OPEN → OPEN |
| a solar powered music festival | every Qloo answer | Las Vegas-Henderson-North Las Vegas, NV → none | 23 → none | OPEN → CANNOT_DECIDE |
| An AI recipe planner for families | search (the matched products) | Richmond, VA → Richmond, VA | 13 → 13 | OPEN → REFUSE |
| An AI recipe planner for families | demographics | Richmond, VA → Richmond, VA | 13 → 13 | OPEN → REFUSE |
| An AI recipe planner for families | audiences | Richmond, VA → Richmond, VA | 13 → 13 | OPEN → REFUSE |
| An AI recipe planner for families | compare (the leader) | Richmond, VA → Richmond, VA | 13 → 13 | OPEN → REFUSE |
| An AI recipe planner for families | trending | Richmond, VA → Richmond, VA | 13 → 13 | OPEN → OPEN |
| An AI recipe planner for families | recommendations | Richmond, VA → Richmond, VA | 13 → 13 | OPEN → OPEN |
| An AI recipe planner for families | heatmap | Richmond, VA → none | 13 → none | OPEN → OPEN |
| An AI recipe planner for families | destinations | Richmond, VA → Richmond, VA | 13 → 13 | OPEN → OPEN |
| An AI recipe planner for families | places | Richmond, VA → Richmond, VA | 13 → 13 | OPEN → OPEN |
| An AI recipe planner for families | tags (persona) | Richmond, VA → Richmond, VA | 13 → 13 | OPEN → OPEN |
| An AI recipe planner for families | brand ladder | Richmond, VA → Richmond, VA | 13 → 13 | OPEN → OPEN |
| An AI recipe planner for families | every Qloo answer | Richmond, VA → none | 13 → none | OPEN → CANNOT_DECIDE |
Recomputed by the check's own rules from cards recorded on 22 Oct 2026; live, the later questions are asked from the search's matches (the values).
Every outside test
| Test | Question | Arms | Numbers | 95% interval | Result |
|---|---|---|---|---|---|
| Where demand is | Where is a kind of business most common per person? | Knos with Qloo, Claude alone | 0.193 against 0.407 (n 30) | [-0.360, -0.060] | loss |
| Films fans also love | Which films do a film's fans also love? | Knos with Qloo, Claude alone | 0.320 against 0.307 (n 30) | [-0.060, +0.087] | tie |
| Kickstarter, 250 held out | Did a past project reach its goal? | open signals with Qloo, always success | 78.0% tuned (59.6% as one verdict) against 78.8% (n 250) | [-0.063, +0.039]* | tie |
| Kickstarter, 300 ranked | Does the score rank projects by money raised? | open signals with Qloo, open signals alone | Spearman 0.108 against 0.113 (n 300) | [-0.037, +0.025] | tie |
| 200 research awards | Did a past award succeed? | open signals, a coin | 47.5% against 50% (n 200) | [-0.093, +0.044]* | tie |
| Each Qloo signal removed | Does one Qloo answer change the launch city, the target or the verdict? | each recorded card with every Qloo answer, then with one call family removed | 18 of 59 removals change one; all removed, 6 of 6 (6 cards) | [+0.203, +0.431]† | win |
| Launch cities, pre-registered | Do Qloo's lifts find the metros where a kind of business is densest? | top 15 metros by people reranked by lift, people alone, Claude alone, shuffled lift | not run: Qloo has refused every call since 25 Oct (HTTP 429) | none | not run |
An interval above 0 is a win and one below 0 a loss. One that holds 0 is a tie. * Wilson interval of the accuracy, less the base rate, from the published counts. † Wilson interval of the share of removals that change one. A test that has not run has no result.
Where it fails
Census: Knos with Qloo 0.193, Claude alone 0.407 at naming where a business is dense. Knos loses. The result
Films: 0.320 with Qloo, 0.307 alone, interval [-0.060, +0.087]. A tie. The result
Kickstarter, tuned: 78.0% right with Qloo, below the 78.8% of always guessing success. The report
No map: 12 of 30 business types got no Qloo heatmap, so Knos named no city for them. The result
Qloo beside Claude alone
20 ideas: Qloo changed the answer on all 20. Claude alone named the same audience on 0.
Kickstarter, 250 held-out projects
Tuned: 76.8% right on open signals, 78.0% with Qloo. Always guessing success: 78.8%.
Kickstarter, 300 projects ranked
Spearman 0.113 on open signals, 0.108 with Qloo. Difference -0.006, 95% interval [-0.037, +0.025].
Where demand is, against the Census
30 kinds of business: overlap 0.407 for Claude alone, 0.193 for Knos with Qloo. Difference -0.213, 95% interval [-0.360, -0.060].
Which films a film's fans also love
30 films on MovieLens 32M: overlap 0.320 for Knos with Qloo, 0.307 for Claude alone. Difference +0.013, 95% interval [-0.060, +0.087].
The plan, fixed before the run · The result · The report · Data: Harper and Konstan (2015), MovieLens
200 past research awards
47.5% right where the base rate is 50%. No better than a coin.
An agent paid for verified work
5 test units on devnet, 23 Oct 2026.
The transaction · The newest payout, read from devnet · The check
Last updated