NBA Over/Under Analysis - 2026-02-07
Summary
| Metric | Value |
|---|---|
| Total Picks | 3 |
| Correct | 0 |
| Accuracy | 0% |
| Skipped | 7 |
Worst day in system history. All 3 picks (including 1 HIGH confidence) were wrong. Every pick was UNDER and every game went OVER.
Results
| Game | Pick | Vegas | Actual | Result | Margin | |
|---|---|---|---|---|---|---|
| Houston Rockets @ Oklahoma City Thunder | UNDER 211 | 211.4 | 218 | OVER +6.6 | +6.6 | |
| Cleveland Cavaliers @ Sacramento Kings | UNDER 231 | 234.1 | 258 | OVER +23.9 | +23.9 | |
| Memphis Grizzlies @ Portland Trail Blazers | UNDER 229 | 234.9 | 237 | OVER +2.1 | +2.1 | Correct Picks (0)_No correct picks_ Wrong Picks (3)Houston Rockets @ Oklahoma City Thunder
|
Cleveland Cavaliers @ Sacramento Kings
Memphis Grizzlies @ Portland Trail Blazers
Deep Analysis: What Went Wrong
1. UNDER Bias is Killing Us
The system has a massive UNDER bias that is not supported by results:
The entire system is built around finding UNDER signals (B2B, slow pace, injuries). We rarely if ever pick OVER. This is a fundamental design flaw.
2. Home B2B is Not an UNDER Signal
| Date | Game | Home B2B Team | Reg Total | vs Vegas | Result |
|---|---|---|---|---|---|
| 2026-02-04 | NOP @ MIL | MIL | 256 | +33.4 | OVER |
| 2026-02-07 | CLE @ SAC | SAC | 258 | +23.9 | OVER |
Home B2B: 0/2 with average miss of +28.7 pts. This is not a marginal signal failure — it's catastrophic. Home B2B should be an AUTO-SKIP, not an UNDER factor.
3. Dual B2B is No Longer Automatic
| Date | Game | Reg Total | vs Vegas | Final | Result |
|---|---|---|---|---|---|
| 2026-02-04 | DEN @ NYK | 216 | -10.3 | 261 | UNDER (reg) |
| 2026-02-07 | MEM @ POR | 237 | +2.1 | 237 | OVER |
Dual B2B: 1/2 overall, 1/1 in regulation only. The MEM@POR result shows that dual B2B doesn't guarantee UNDER, especially when both teams have strong OVER tendencies. The DEN@NYK game worked in regulation but went to 2OT. Dual B2B should be downgraded from "auto-UNDER HIGH" to "MEDIUM at best, check team OVER trends first."
4. Team OVER Trends Override Fatigue
Both MEM (100% OVER, 3/3) and POR (100% OVER, 2/2) went OVER despite dual B2B. Both SAC (67% OVER) and OKC (75% OVER) went OVER despite our UNDER picks. When teams with strong OVER tendencies play, fatigue factors may not be enough to suppress scoring.
New rule proposal: If both teams have >60% OVER rate, do NOT pick UNDER regardless of B2B status. SKIP or lean OVER instead.
5. Vegas Already Prices In Our Factors
All three picks tried to go below Vegas lines that already accounted for B2B and injuries:
When our factors are obvious enough for us to spot, Vegas has already priced them in. Going further UNDER is fighting the market.
Updated System Stats
| Metric | Before Feb 7 | After Feb 7 | Change |
|---|---|---|---|
| Overall Accuracy | 53% (16/30) | 48% (16/33) | -5% |
| HIGH Confidence | 71% (5/7) | 63% (5/8) | -8% |
| MEDIUM Confidence | 48% (10/21) | 43% (10/23) | -5% |
| UNDER Accuracy | 50% (13/26) | 45% (13/29) | -5% |
| OVER Accuracy | 75% (3/4) | 75% (3/4) | -- |
| Dual B2B | 1/1 (100%) | 1/2 (50%) | -50% |
| Home B2B | 0/1 (0%) | 0/2 (0%) | -- |
Cumulative Stats
Rule Adjustments Needed
1. Home B2B = AUTO-SKIP (0/2, avg miss +28.7 pts — never pick UNDER on home B2B again) 2. Dual B2B downgrade — No longer auto-HIGH. Check team OVER trends first. If both teams >60% OVER rate, SKIP instead of UNDER 3. UNDER picks need B2B away + slow pace — This is the only proven UNDER pattern (2/2). All other UNDER setups are coin flips or worse 4. Prioritize OVER picks — 75% vs 45% accuracy. System should actively seek OVER setups (Fast vs Fast, high OVER trend teams) 5. Stop fading Vegas on lowest-total games — When Vegas already sets a low total, they've priced in the defense/pace factors 6. MEDIUM UNDER confidence is broken — 43% (10/23). Must require away B2B as primary factor or SKIP