ML Signal Quality Model

Model Loaded
Training Data
218
records with P&L
CV Accuracy
44.4%
cross-validation
Models
2
none
Last Trained
2026-03-08 03:46
1 days ago
Model's Signal Quality Assessment
ML Predicted
The model scores each signal type based on learned patterns. Higher quality score = model predicts better outcomes.
Signal Name Model Grade Quality Score P(Loss) P(Big Win) Actual Win% Actual Avg P&L Samples Model Confidence
Demand Zone Reversal (3/4 Confirmed) Strong +0.486 11.4% 11.7% 100% +1.51% 1
RSX Visual Medium +0.271 30.1% 15.2% 100% +15.99% 1
X SniperBuy Vidya Weak +0.149 23.7% 3.4% 100% +0.64% 3
X SniperSell Vidya Weak +0.140 22.7% 5.7% 0% -2.91% 5
Unknown (99) Weak +0.092 45.6% 11.8% 70% +1.82% 33
X SniperSell MainTrend AI Weak +0.036 44.0% 8.6% 0% -4.30% 7
X SniperSell Mundy Avoid -0.024 46.6% 6.1% 15% -2.77% 13
WaveMatrix shortStrongSignal Avoid -0.044 50.7% 5.1% 20% -1.26% 5
WaveMatrix longStrongSignal Avoid -0.086 43.3% 1.2% 50% -0.49% 8
Momo Strong Entry Avoid -0.100 53.8% 2.8% 43% -1.51% 53
Trend Continuation Avoid -0.133 56.5% 4.6% 22% -53.23% 18
Strong Short Signal Avoid -0.134 52.9% 2.4% 0% -1.98% 2
Unknown (56) Avoid -0.158 65.8% 5.7% 55% +0.71% 89
X SniperSell Ana Avoid -0.191 61.7% 2.6% 0% -4.17% 3
Unknown (53) Avoid -0.224 73.3% 4.4% 20% +15.98% 5
X SniperBuy Mundy Avoid -0.227 64.1% 1.4% 29% -0.60% 14
RSX Strong Buy Avoid -0.354 82.8% 1.7% 0% -0.44% 4
X SniperBuy MainTrend AI Avoid -0.401 89.3% 1.5% 0% -2.08% 4
Unknown Regular Buy Avoid -0.411 91.9% 1.3% 0% -0.94% 1
RSX Weak Buy Avoid -0.438 93.7% 0.8% 0% -0.60% 2
Signal Performance by Name (Raw Data)
15 signal types
Signal Name Trades Win Rate Avg P&L Total P&L Best Worst Performance
RSX 7 trades 7 29% +13.51% +94.6% +80.38% -0.76%
Demand Zone 1 100% +1.51% +1.5% +1.51% +1.51%
SniperBuy Vidya 2 100% +0.79% +1.6% +1.24% +0.34%
TDindicator Reliable 89 55% +0.71% +62.9% +34.93% -9.38%
SniperBuy Mundy 7 trades 7 29% -0.74% -5.2% +0.59% -2.79%
Regular Buy 1 0% -0.94% -0.9% -0.94% -0.94%
WaveMatrix 8 trades 8 38% -0.98% -7.8% +1.41% -6.65%
Strong Short 1 0% -1.98% -2.0% -1.98% -1.98%
SniperBuy MainTrend 2 0% -2.08% -4.2% -1.91% -2.25%
Momo Reliable 31 42% -2.45% -75.9% +2.84% -59.36%
SniperSell Mundy Reliable 11 18% -2.68% -29.5% +1.78% -6.78%
SniperSell Vidya 4 0% -3.05% -12.2% -2.05% -4.24%
SniperSell Ana 2 0% -3.40% -6.8% -1.08% -5.72%
SniperSell MainTrend 5 trades 5 0% -5.74% -28.7% -0.53% -17.08%
Trend Continuation Reliable 10 20% -94.98% -949.8% +0.44% -929.35%
Model Training

Train XGBoost model using completed trades with indicator text and P&L data.

Training in progress...
Feature Importance (Top 15)
# Feature Importance
1 ob_result_code 0.1104
2 htf_decision_code 0.0861
3 htf_wm_value 0.0757
4 htf_rating_number 0.0659
5 htf_rating_letter_code 0.0619
6 market_structure_v 0.0587
7 hour_of_day 0.0464
8 wm_value 0.0456
9 htf_trend_strength 0.0444
10 momentum_consistency 0.0421
11 rating_number 0.0406
12 trend_strength 0.0402
13 rating_letter_code 0.0401
14 htf_momentum_consistency 0.0364
15 deviation_pct 0.0364
Training Data
Data Sources
Source Records Usable Status
A signal_log → ESR → closed_position 5 5 Active
C completed_trade.signal_indicator 213 213 Primary
S signal_log with filter_result.scores 1667 No P&L
P&L Label Distribution
114
LOSS (<0%)
53
SMALL (0-1%)
25
MED (1-3%)
21
BIG (≥3%)
Direction Breakdown
157
Long / Buy
56
Short / Sell
Recent Trainable Trades
Symbol Side P&L % Label Date
NVDA long -1.92% LOSS 02/11
SOXL long 0.78% SMALL_WIN 02/11
AMD long 0.96% SMALL_WIN 02/11
STX long 0.06% SMALL_WIN 02/11
LRCX long -1.14% LOSS 02/11
HOOD short 0.06% SMALL_WIN 02/11
GOOG long 0.06% SMALL_WIN 02/11
LRCX long -0.47% LOSS 02/11
ASTS short -5.81% LOSS 02/10
ARM long 0.09% SMALL_WIN 02/10
CIFR long 0.00% SMALL_WIN 02/10
CIFR long -0.75% LOSS 02/10
AVGO long -0.60% LOSS 02/10
NNE short -0.30% LOSS 02/10
SCHW short -0.92% LOSS 02/10
SCHW short -1.00% LOSS 02/10
USAR long -1.31% LOSS 02/10
SCHW short 0.22% SMALL_WIN 02/10
AMKR long 0.10% SMALL_WIN 02/10
SCHW short -1.31% LOSS 02/10
Model Details
Long Model Active
Samples: 157
Classes: 4
Accuracy: 44.4%
Features: 53
Short Model Not trained
Fallback Model Active
Samples: 213
Classes: 4
Accuracy: 41.7%
Features: 53