Public Lab · 90-day window
Does the Kairon Score actually pick winners?
Three paper-trading bots run against the live Polymarket pool every 15 minutes. Identical $10,000 bankroll, identical risk rules, identical $100 sizing. Only the entry rule differs. Honest answer below — refreshed every tick, audit table is public.
HONEST READ
The bots haven't produced enough closed trades yet for a meaningful comparison. Check back after the first 90-day window has accumulated.
How it works
- Every 15 minutes a cron tick pulls the active Polymarket pool and scores every market with the canonical formula.
- Kairon Score opens a $100 paper position when a market crosses tier A+ (score ≥ 70).
- Random opens a $100 paper position with 8% probability per market per tick (matched to score-bot's average entry frequency).
- Price Heuristic opens a $100 YES position whenever the market's implied YES probability ≥ 75%.
- All three use identical risk rules: stop-loss −25%, take-profit +50%, trailing stop 20% (only after the position is in profit), time exit T−2h before resolution. 20-position cap.
- Each variant runs on the same $10,000 paper bankroll. Side selection (YES vs NO) is identical: take YES if the market thinks YES is more likely than NO.
KILL CRITERIA
Per EXPERIMENT_PLAN.md §2.5: if after 90 days the Kairon Score bot is below 0% return OR below the price-bot's Sharpe, we publicly suspend the experiment, document why, and either ship new weights or pivot the product positioning. This page will say so.