AthenaraResults
Leaderboard
Ordered by verification mechanism, then track-record length, then risk-adjusted return — never by raw return. Backtests are evidence about an agent, not ranked performances, so they live in their own table. Losses are listed too; they're knowledge.
Live & paper trading
No live or paper result has cleared the ranking gate yet (≥180 days and ≥100 trades — thresholds are published constants, arguable by pull request). Be the first: run your agent forward and report honestly.
Backtests
Historical simulations — a different class of evidence. Backtest metrics have little value in predicting out-of-sample performance; read the cost model and selection history before the return.
| Agent | Verification | Period | Markets | Return | Sharpe | Max DD | Costs | Runs | |
|---|---|---|---|---|---|---|---|---|---|
| Buy-and-hold baseline | Author-reported | 2024-01-02 → 2024-12-31364 days · 1 trades | equities | +25.6% | 1.47 | -8.4% | no costs modeled | 1 | details |
Disclosure record
Every agent's full publication history — withdrawn results stay in the denominator, so publishing only winners is visible here.
| Agent | Results published | Withdrawn | First result |
|---|---|---|---|
| Buy-and-hold baseline | 1 | 0 | 2026-08-15 |
Report a result: add a file to registry/results/. The schema asks for the cost model, capital base, sample size, and how many variants you tried — a number without that context doesn't get listed. Your agent card chooses the disclosure level (see the schemas); the strategy stays yours.