PUBLIC LEDGERNO CHERRY-PICKINGWINS · LOSSES · MISSES INCLUDED
Every signal Phalanx fires is logged at generation and scored at expiry against its directional outcome. The numbers below are pulled directly from the signal API — nothing held back, nothing hand-picked.
SIGNALS FIRED
1,744
since launch
SCORED
1,717
T+1 outcome on each
OVERALL HIT RATE
58.3%
n = 1,717
HIT RATE · HIGH CONVICTION
60.0%
n = 10
MOST RECENT 50 SCORED SIGNALS
| When | Type | Dir | Conv. | Outcome | Return |
|---|---|---|---|---|---|
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +7.84% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 58 | win | +5.94% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | loss | -4.67% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +13.40% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +6.64% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 58 | loss | -8.78% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 62 | win | +0.27% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +7.60% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +9.24% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +3.45% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 58 | win | +1.52% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 62 | win | +2.43% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +7.59% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +2.24% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | loss | -8.78% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +10.70% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | loss | -0.38% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +0.25% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +1.73% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | loss | -10.68% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +2.03% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | loss | -0.38% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +0.55% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +13.96% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +7.04% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +7.59% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | loss | -4.57% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +2.95% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 62 | loss | -2.74% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 62 | loss | -5.07% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 62 | win | +1.20% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 58 | loss | -0.50% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +4.71% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 62 | win | +0.60% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +13.96% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 62 | win | +0.27% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 58 | win | +0.40% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +6.26% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +1.55% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | loss | -1.07% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +10.70% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +1.48% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +9.06% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 58 | win | +5.94% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | loss | -5.44% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | loss | -1.24% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +6.79% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 62 | loss | -0.97% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +6.64% |
| Jun 01 00:00 | ENSEMBLE DECISION | LONG | 70 | win | +21.68% |
HIT RATE BY CONVICTION TIER
Conviction (0–100) is the score the four-model AND-gate produces when Claude, Grok, Gemini, and GPT all agree on a direction. The higher the agreement, the higher the conviction score — and the better the outcome. 1,717 scored signals below.
LOW
Conviction score 0 – 39
56.4%
hit rate
222 signals scored
MEDIUM
Conviction score 40 – 59
58.6%
hit rate
1,485 signals scored
HIGH
Conviction score 60 – 79
60.0%
hit rate
10 signals scored
A signal is correct if its directional prediction resolved in the right direction at expiry (T+1 for most signals). We do not cherry-pick — every signal filed is counted.
A hit is any signal that produced a profitable trade.
Ensemble Decision
Backfill
Debate Verdict
Power Outflow
Power Inflow
Power Inflow
Debate Verdict
Wolf
Hawk
Eagle
Phoenix
Raven
Owl
Condor
Duck
Dolphin
Gated = human approval required before execution (Condor — long-term accumulation). Engine hit rates are aggregated into signal types above.
HOW TO READ THIS
We measure what we can and flag what we can't. Read the data, not the marketing.
Get tickers — Signal →A signal is correct if its directional prediction — LONG or SHORT — resolved profitably at expiry. Most signals expire T+1 (next trading day). We do not adjust for transaction costs in the headline hit rate. The leaderboard measures signal quality, not net P&L.
Conviction (0–100) is the score the four-model AND-gate assigns to a signal. Claude, Grok, Gemini, and GPT each evaluate the setup independently at temperature 0.20. The score reflects how strongly all four models agreed. Any single veto holds the trade and drops conviction.
Specific symbols on individual signals are not shown here — that's the value proposition of Phalanx Signal. The leaderboard proves the system works. The feed tells you what it just decided and which ticker it decided it on.
No — and we flag it when sample sizes are small. The HIGH conviction tier has fewer signals because most setups don't clear all four models. Early numbers there should be read with appropriate skepticism. We apply the same standard to ourselves.
