Methodology · Three documents
How the analytics engine thinks.
Courtside is built for cited, inspectable sports research. These pages document the projection model, the bet verdict math, and the DFS optimizer behind every answer the chat produces.
- Inspectable
- Every projection, every probability, every lineup slot is paired with the inputs that produced it. The model is most useful when you can argue with it.
- Cited
- Statistical claims link back to the tool result behind them. The answer is an essay you can audit at the line level, not a single number to trust.
- Calibrated
- Ranges over point estimates. Floor and ceiling beat a misleading median. Edge bands beat false precision.
01
02
03
Index3 documents · last updated this season
The three sources that power the answer.
- Doc 01sports projection methodology
Sports projection methodology
Learn how Courtside approaches NBA and MLB projections with recent form, long-term baselines, uncertainty ranges, and market context.
Read the doc - Doc 02bet verdict methodology
Bet verdict methodology
Learn how Courtside handles sportsbook margin, model probability, edge, expected value, stake guidance, and betting verdict caveats.
Read the doc - Doc 03DFS lineup optimizer methodology
DFS lineup optimizer methodology
Learn how Courtside builds DFS lineups with projections, salary rules, roster rules, exposure caps, and strategy settings.
Read the doc
Stop dashboard-hopping.
Ask a question.
Join the Discord for pick notes, slate questions, product drops, and smarter NBA and MLB conversation.