Leaderboard
Lose less, compound steadily
The passive book comes first: the bots built to trade quietly and fall less than the market, judged on drawdown and on how much of SPY's falls they take against how much of its rises. Every other bot keeps running below as a lab experiment, ranked by alpha: its return minus its own names held at the exposure it actually carried. Bots need 20 sessions before either is measured.
The passive book
| Bot | Trend | Return | Max drawdown | Worst month | Down / up | vs SPY, same risk |
|---|---|---|---|---|---|---|
70 percent 16-ETF trend overlay, 30 percent equal weight halved when SPY is below its 150-session mean: the passive book. The brake cuts deep drawdowns by about a quarter to a third when they build over more than a week; little help in one-day crashes; no reliable Sharpe gain since 1990. | – | — | — | — | — | building 10/20 |
Return sleeve: residual momentum, the one strategy that passed its holdout. | – | −0.7%SPY −1.9% | −5.3%SPY −3.0% | −2.6% | 0.64 / 0.69 | +1.3% |
Lab experiments keep running, not part of the passive book
1PAPER
Cheshire — VRP Income Sleeve (no timing-alpha claim)vrp_sleeve
–
Alpha vs buy & hold
+0.3%1d streak100% win1 trades
2PAPER
Fork: GME Endpoint Momentum (15m)config_gme_fork
–
Alpha vs buy & hold
+2.4%1d streak37% win19 trades
3PAPER
Fork: VIX Squeeze Decoupled (15m)config_vix_squeeze
–
Alpha vs buy & hold
−1.3%1d streak49% win71 trades
| # | Strategy | Trend | Alpha | Win | Streak | Max DD | Down / up | P&L | |
|---|---|---|---|---|---|---|---|---|---|
| 4 | Config Crypto — 24/7 Digital Assets config_crypto · PAPER · ON WATCH | – | Building 0/20 | 43% | 1d | −8.4%SPY −3.1% | 0.50 / 0.09 | −$4,137 | Fork |
| 5 | Canary — Hosted Bot Shakedown u_bcb53b1f · PAPER · BUILDING | – | +0.6% | 44% | 0d | −0.5%SPY −3.1% | −0.08 / −0.04 | +$35 | Fork |
| 6 | Kinetic Basket: Systemic Crisis (GME+NVDA) systemic_risk_basket · PAPER · IN CONE | – | +5.6% | 45% | 0d | −4.5%SPY −3.1% | −0.08 / −0.02 | +$51 | Fork |
| 7 | Config A — ML-Gated Multi-Ticker config_a · PAPER · IN CONE | – | +9.1% | 50% | 0d | −9.5%SPY −3.1% | 0.69 / 0.94 | +$8,213 | Fork |
| 8 | Kinetic Engine: GME 30m (+25k% Alpha) gme_30m_kinetic · PAPER · IN CONE | – | −0.7% | 45% | 0d | −6.1%SPY −3.1% | −0.07 / −0.23 | −$1,562 | Fork |
| 9 | Config D — VIX Mean-Reversion config_d · PAPER · IN CONE | – | −2.0% | 44% | 0d | −2.6%SPY −3.1% | 0.32 / 0.22 | −$652 | Fork |
| 10 | Config B — ATR Chandelier (Low-Vol Tickers) config_b · PAPER · ON WATCH | – | −4.5% | 52% | 0d | −5.6%SPY −3.1% | 0.54 / 0.21 | −$3,331 | Fork |
| 11 | Config C — GME 7-Lags Options Echo gme_7lags_option · PAPER · IN CONE | – | −37.8% | 0% | 0d | −12.7%SPY −4.2% | 0.21 / −0.23 | −$145 | Fork |
| 12 | Config C — Options ML-Gated Multi-Ticker config_c · PAPER · ON WATCH | – | −17.5% | 5% | 0d | −13.9%SPY −4.5% | 0.23 / −0.25 | −$14,265 | Fork |
| 13 | Config A2 — Overnight (paper fork of A) config_a2 · PAPER · BUILDING | – | Building 0/20 | 0% | 1d | — | — | −$277 | Fork |
| 14 | Config A4 — Breakout Flip Overnight, 20m (paper) config_a4 · PAPER · BUILDING | – | Building 0/20 | 0% | 0d | — | — | −$600 | Fork |
| 15 | Config A5 — Breakout Flip Vote, 6 of 21 grids (paper) config_a5 · PAPER · BUILDING | – | Building 0/20 | 100% | 0d | — | — | +$14 | Fork |
| 16 | Config A5W — Breakout Flip Vote, 101 tickers (paper) config_a5w · PAPER · BUILDING | – | Building 0/20 | 13% | 0d | — | — | −$167 | Fork |
| 17 | Config A3 — Breakout Flip Overnight (paper) config_a3 · PAPER · BUILDING | – | No trades yet | 0% | 0d | — | — | +$0 | Fork |
| 18 | Config E — Mirror Compression Stack (MR) config_mr_tuned · PAPER · BUILDING | – | No trades yet | 0% | 0d | 0.0%SPY −3.1% | 0.00 / 0.00 | +$0 | Fork |
All figures come from the live monitor. PAPER means paper — run modes and lifecycle stages render exactly what the backend reports. IN CONE / ON WATCH / QUARANTINED say whether a strategy is still tracking the backtest it was published on, which is a different question from whether it made money this month. How verification works