Backtest Results - Monday, February 2, 2026

SPREAD PICKS SCORECARD

Spread Record: 1-0 (+1.0u)

OVER/UNDER PICKS SCORECARD

O/U Record: 2-1 (+0.9u)

PARLAY RESULT

Parlay Result: LOSS (-0.5u)

DAILY SUMMARY

QUALITATIVE ANALYSIS

WHAT WORKED

• New Orleans +7.5 covered by 0.5: The injury adjustment methodology correctly identified Murray's absence as significant, and the 4.4-point line value calculation was spot-on despite medium confidence
• NOP @ CHA UNDER 230.5 crushed (197 total): Properly weighted the offensive impact of missing key players on both sides, leading to a massive 33.5-point under win

WHAT MISSED

• MIN @ MEM UNDER 230.5 got torched (265 total): Model severely underestimated how depleted Memphis would create a pace-up, high-scoring environment. Missing 6+ rotation players led to chaos basketball, not defensive grinding
• HOU @ IND OVER 218.5 hit easily (232 total): This wasn't actually picked in the analysis shown, but the injury logic should have flagged Houston's missing defenders (KD, FVV) creating higher-scoring conditions

MODEL CALIBRATION NOTES

• Medium confidence was appropriately cautious on the spread pick that barely covered
• Systematic bias detected: Model appears to overweight injuries as pace/scoring dampeners rather than considering how depleted rosters can create chaotic, high-scoring games
• Missing any HIGH confidence picks to evaluate calibration accuracy

METHODOLOGY ADJUSTMENTS TO CONSIDER

1. Implement "chaos factor" for heavily depleted teams: When 4+ rotation players are out, adjust for increased pace and defensive breakdowns rather than just offensive production loss
2. Separate injury impact by game phase: Missing defensive players (like Houston's KD/FVV) should trigger over leans, while missing offensive stars should trigger unders - current model doesn't distinguish impact direction effectively

Overall: 2-1 on identified picks, but missed the systematic bias around injury impact on game flow.

CUMULATIVE SEASON RECORD

View original analysis