Starknet (STRK) Backtest

We ran a full walk-forward backtest using the AlgoZilla Omniscius v6.2 Pythia model for STRK - Starknet. The results below cover the complete out-of-sample window from Sep 2024 → May 2026 across 3 walk-forward cycles (= train on the past, test exclusively on unseen future bars, then roll the window). The model executed 670 trades in total, compounding an initial $10,000 base to a final value of $92,150. All results are net of trading fees based on Bitvavo rates (0.15% entry + 0.25% exit). The model uses a 15‑minute delay on every signal to simulate live execution latency — the same decision logic runs in production.