Pick lines, not players
The most common mistake bettors make is starting with 'who do I like tonight' instead of 'where is the line wrong'. Players don't have edges — prices do. A great player at a bad number is a losing bet.
Run every prop through these five filters. If it fails any, pass.
Filter 1 — Is the line off by at least 8%?
Your projection minus the line, divided by the line, should be at least 8% to overcome the vig and your own projection error. Smaller edges exist but require razor-sharp models.
Example: line is 22.5 points, you project 24.5 — that's an 8.9% edge. Bet it. Line is 22.5, you project 23 — pass.
Filter 2 — Is the volume locked in?
Edge means nothing if the player doesn't play. Check the injury report 90 minutes before tip, then again 30 minutes before. Confirm coach's pre-game presser. If there's any doubt about minutes, the edge isn't real.
Filter 3 — Is the matchup directionally right?
If you're betting an over on a shooter, the opponent should be a high-pace, weak-perimeter-defense team. If those don't line up with your projection, your model is missing something — usually a defensive scheme adjustment.
Filter 4 — Is this the best price available?
Check at least three books. If you can get the same number at -110 instead of -120, that's 10 cents — over a season, that's the difference between profit and breakeven. Always bet the best price, never the convenient one.
Filter 5 — Does the bet size make sense?
A 10% edge doesn't mean bet 10% of your bankroll. Use fractional Kelly (typically a quarter or a half) or flat units. For most bettors, 1-2% of bankroll per prop is the right size. Bigger bets on bigger edges, but never more than 5% on a single line.
What separates winners from losers
It's not finding more edges — it's the discipline to pass on bets that fail the filters. The bettors who win long-term place fewer bets, not more. Quality over volume, every slate.