Backtest Results - Sunday, February 1, 2026
SPREAD PICKS SCORECARD
| Pick | Result | Outcome |
|---|---|---|
| LA Clippers +2.5 | Clippers 117 - Suns 93 | Covered by 26.5 |
| Denver Nuggets +7.5 | Thunder 121 - Nuggets 111 | Failed to cover by 2.5 |
| New York Knicks -5.5 | Lakers 100 - Knicks 112 | Covered by 6.5 |
Spread Record: 2-1 (+0.9u)
OVER/UNDER PICKS SCORECARD
| Pick | Result | Outcome |
|---|---|---|
| LAL @ NYK UNDER 227.5 | Lakers 100 - Knicks 112 (Total: 212) | Under by 15.5 |
| OKC @ DEN UNDER 224.5 | Thunder 121 - Nuggets 111 (Total: 232) | Over by 7.5 |
| LAC @ PHX UNDER 208.5 | Clippers 117 - Suns 93 (Total: 210) | Over by 1.5 |
O/U Record: 1-2 (-1.2u)
PARLAY RESULT
| Leg | Pick | Outcome |
|---|---|---|
| 1 | LAL @ NYK UNDER 227.5 | WIN |
| 2 | OKC @ DEN UNDER 224.5 | LOSS |
Parlay Result: LOSS (-0.5u)
DAILY SUMMARY
| Category | Record | Units |
|---|---|---|
| Spreads | 2-1 | +0.9 |
| Over/Under | 1-2 | -1.2 |
| Parlay | LOSS | -0.5 |
| Daily Total | -0.8 |
QUALITATIVE ANALYSIS
WHAT WORKED
Knicks -5.5 (WIN): The model correctly identified Austin Reaves as a massive loss (26.6 PPG) and properly valued his primary playmaking role. The 6.5-point cover validated the injury impact assessment.
Clippers +2.5 (WIN): Despite missing James Harden, the model identified Phoenix without Booker as equally compromised. The 26.5-point blowout suggests both teams were properly evaluated as depleted.
WHAT MISSED
OKC @ DEN Under 224.5 (LOSS): The model underestimated offensive output despite key injuries to both teams. Thunder still put up 121 points without Williams/Mitchell, suggesting injury impact calculations were inflated.
LAC @ PHX Under 208.5 (LOSS): Missed by 1.5 points - essentially a coin flip. The Clippers' 117-point explosion without Harden wasn't foreseeable, indicating depth evaluation was insufficient.
MODEL CALIBRATION NOTES
Confidence ratings need work: The Knicks pick was marked "LOW confidence" despite a 10.6-point edge and successful outcome. This suggests the confidence algorithm is poorly calibrated.
Injury impact appears overvalued: Multiple games went Over despite significant player absences, indicating the model assigns too much defensive value to missing offensive players.
METHODOLOGY ADJUSTMENTS TO CONSIDER
Refine confidence scoring: Large edges (8+ points) shouldn't automatically trigger low confidence. Separate edge size from uncertainty factors.
Adjust injury impact on totals: Reduce the defensive impact assumption when key offensive players are out. Missing scorers often leads to higher pace/desperation, not lower scoring.
Overall: 3-3 record with solid reasoning on spreads but totals methodology needs refinement.
CUMULATIVE SEASON RECORD
| Category | Record | Win % | Units |
|---|---|---|---|
| ATS | 19-14 | 57.6% | +9.0 |
| O/U | 15-15 | 50.0% | |
| Parlays | 4-7 | ||
| Season Total | +9.0u |