Backtest Results - Friday, March 20, 2026
SPREAD PICKS SCORECARD
| Pick |
Result |
Outcome |
| MEM +15.5 |
Celtics 117 - Grizzlies 112 |
Covered by 10.5 |
| MIN -1.5 |
Trail Blazers 108 - Timberwolves 104 |
Failed to cover by 5.5 |
| None qualify |
Could not match team: None qualify |
NO_RESULT |
Spread Record: 1-1 (-0.1u)
OVER/UNDER PICKS SCORECARD
| Pick |
Result |
Outcome |
| GSW @ DET UNDER 217.5 |
Warriors 101 - Pistons 115 (Total: 216) |
Under by 1.5 |
| NYK @ BKN UNDER 213.5 |
Knicks 93 - Nets 92 (Total: 185) |
Under by 28.5 |
| POR @ MIN UNDER 231.5 |
Trail Blazers 108 - Timberwolves 104 (Total: 212) |
Under by 19.5 |
O/U Record: 3-0 (+3.0u)
PARLAY RESULT
| Leg |
Pick |
Outcome |
| 1 |
GSW @ DET UNDER 217.5 |
WIN |
| 2 |
POR @ MIN UNDER 231.5 |
WIN |
Parlay Result: WIN (+2.6u)
DAILY SUMMARY
| Category |
Record |
Units |
| Spreads |
1-1 |
-0.1 |
| Over/Under |
3-0 |
+3.0 |
| Parlay |
WIN |
+2.6 |
| Daily Total |
|
+5.5 |
QUALITATIVE ANALYSIS
WHAT WORKED
- Under betting was exceptional (3/3): Model correctly identified low-scoring environments, particularly the NYK/BKN under that crushed by 28.5 points
- Memphis spread (+15.5): Despite heavy injuries, correctly predicted Grizzlies would keep it competitive at home against Boston's road fatigue
WHAT MISSED
- Minnesota -1.5: Model underestimated impact of Anthony Edwards absence (-8.9 pts) on a thin 1.5-point spread - this should have been flagged as too close given the star player out
- "None qualify" entry: Analysis shows systematic issue where model identified no qualifying spread bets but results suggest missed opportunities
MODEL CALIBRATION NOTES
- Confidence ratings need review: LOW confidence Memphis pick actually hit while model avoided other spreads that may have had value
- Under bias worked perfectly: All three totals went under, suggesting model may have systematic edge in identifying pace/scoring environments better than oddsmakers
METHODOLOGY ADJUSTMENTS TO CONSIDER
- Tier 1 injury threshold: Any spread under 3 points with a Tier 1 player (8+ point impact) out should trigger automatic fade of that team
- Expand qualifying criteria: The "none qualify" result suggests overly conservative thresholds - review minimum edge requirements for spread bets when totals analysis is clearly sharp
Bottom line: Strong totals methodology, but spread analysis too conservative and needs better integration of star player injury impact on close lines.
CUMULATIVE SEASON RECORD
| Category |
Record |
Win % |
Units |
| ATS |
51-49 |
51.0% |
-4.1 |
| O/U |
57-54 |
51.4% |
|
| Parlays |
7-34 |
|
|
| Season Total |
|
|
-4.1u |
View original analysis