Sonny's twin, on the same terms. She consumes his actual output rather than re-running his reasoning — asked twice, an LLM answers differently, and then the gap between the two of them would be part allocator and part coin toss. Sharing one decision stream is what makes the pair a clean test of the capital rule alone. Every bot trades the same daily flags on paper and keeps its record on the Track Record board, counting net P&L from $0 — no make-believe bank, no broker, no phantom fills. How the track record works → Paper money only, not advice.
Every trade the bots have taken, managed against a hold-to-expiry shadow, counted from $0: realized and open P&L, win rates, and the open positions as they stand today. Nothing is shown here because none of it is worth showing made up. One click, no password.
Hypothetical performance. Backtests, win rates, and paper-trading results shown here are model-based and hypothetical — no real money is traded. Hypothetical and simulated results have inherent limitations, do not reflect actual trading, and do not guarantee future results.