Applied Statistics / Quantitative Backtesting

What to call it: walk-forward validation, out-of-sample testing, real-data backtesting.

Real evidence

• NFL 2024 backtest: every claimed metric (MAE, winner accuracy, spread-beat rate) independently recomputed and matched to full floating-point precision against real season data.
• MLB backtest: recomputed and matched to three decimal places.
• Honest negative result included: NBA backtest showed the market baseline beating the model — a real, un-cherry-picked finding, which is itself evidence of rigor (a system that "wins" on every single test is the one to be suspicious of).

Honest calibration

The backtesting infrastructure and statistical discipline are real and verified. The live-trading/market-edge layer built on top of it (CSTEProbe, portfolio optimization) is still mostly conceptual scaffolding — a real distinction to keep clear with a quant-finance audience specifically.