Skip to content

Level 4 — General manager

Level 4 · General manager

Identity. Roster construction, multi-year strategy, draft and free agency.

You stop running individual analyses and start building systems that produce predictions. Schema design matters. Feature engineering matters. Train/test discipline matters.

SQL / DataPython / ML
Schema design and normalizationscikit-learn pipelines
Indexing and query optimizationfeature engineering for tabular data
dbt basics (models, refs, tests)model evaluation: holdout, cross-validation, baselines

Draft pick value model. Predict career Approximate Value from college production, combine numbers, and draft position. Train on 2000-2020, test on 2021-2023, project the 2025 Lions class.

A predictive model with:

  • Documented features and the rationale behind each
  • A real held-out test set (no leakage)
  • A baseline comparison (e.g., draft-position-only model)
  • A one-page writeup of model performance, feature importance, and the 2025 projections
  1. Schema design — normalize, then index
  2. Indexes — when, what, why
  3. Query optimization with EXPLAIN
  4. scikit-learn basics
  5. Feature engineering
  6. Pipelines and cross-validation
  7. Model evaluation — baselines, metrics, calibration
  8. dbt basics