**Model variations.** Does adding a new feature to your model improve performance? Run both versions in parallel for a defined sample size and compare. - **Staking strategies.** Does switching from flat staking to Kelly-based staking improve risk-adjusted returns? Run both approaches on paper for a