Harness Analysis
| Candidate | Hermes | ChatGPT | Claude | Gemini | Perplexity | Board |
|---|---|---|---|---|---|---|
| Elly De La Cruz Under 1.5 Total Bases | — | T3 | T3 | T3 | T2 | Loss |
| Willson Contreras Under 1.5 Total Bases | — | T3 | T3 | T3 | T3 | Win |
| Joel Embiid Under 29.5 Points + Assists | — | T2 | T2 | T2 | T1 | Win |
| Joel Embiid Under 38.5 Points + Rebounds + Assists | — | T2 | T2 | T2 | T1 | Win |
Final Decision
All harnesses rejected C1/C2 as fragile MLB total-bases unders with thin edge at -141; C3/C4 form one correlated Embiid-under thesis with a 4/10 L10 and prior local Embiid-under failures — not safe enough for a hard 3-loss-cap Martingale recovery.
No official bet placed; Martingale ladder unchanged. Board went 3-1: Elly De La Cruz lost (3 TB), Contreras won (0 TB), Embiid P+A won (23 of 29.5), Embiid PRA won (29 of 38.5). PASS avoided the C1 loss; the Embiid paper cluster won but the discipline bar was correctly maintained.
Candidates
- Rithmm
- 60.6%
- BE
- 58.5%
- DTM
- 4.9%
- Recent
- L10 6/10
- Rithmm
- 60.6%
- BE
- 58.5%
- DTM
- 5%
- Recent
- L10 6/10
- Rithmm
- 71.3%
- BE
- 51.9%
- DTM
- 40.5%
- Recent
- L10 4/10
- Rithmm
- 74.3%
- BE
- 53.3%
- DTM
- 45%
- Recent
- L10 4/10
Harness Notes
Official PASS preserved the Martingale sequence. Board went 3-1; PASS avoided the C1 (Elly De La Cruz, 3 TB) loss while the Embiid cluster paper leans both won.
SKIP correct officially. Forced pick was Embiid U29.5 P+A which won at 23 P+A, but SKIP discipline avoided Elly De La Cruz (3 TB, loss). 'Good spreadsheet, bad sweat energy' assessment proved apt.
Conditional lean on Embiid U29.5 P+A won at 23 P+A. Claude correctly flagged this was only appropriate at loss level 0–1 and not safe at higher Martingale levels — and framed it as marginally viable rather than high conviction.
Conditional pick C4 Embiid U38.5 PRA won at 29 PRA. Gemini correctly noted this was the stronger model thesis but cautioned 'proceed with extreme caution' and recommended SKIP at L2+. The correlated-with-C3 warning held — both Embiid unders won together.
SKIP correct officially. Perplexity rated C3/C4 as Tier 1 on paper but still recommended SKIP, arguing a 3-loss-cap Martingale system needs 80%+ true win probability — which nothing on this board met. Correct discipline, opportunity cost on the Embiid wins.