diff --git a/core/orchestrator.py b/core/orchestrator.py
index cc7b8a9..cb4299b 100644
--- a/core/orchestrator.py
+++ b/core/orchestrator.py
@@ -1204,10 +1204,12 @@ class TradingOrchestrator:
         if not self.realtime_processing:
             return
         try:
-            # This is where you would feed the state to the DQN model for prediction
-            # or store them for training. For now, we just log and store the latest.
-            # self.latest_cob_state[symbol] = cob_data['state']
-            # logger.debug(f"COB DQN state updated for {symbol}: {cob_data['state'][:5]}...")
+            # Store the COB state for DQN model access
+            if 'state' in cob_data and cob_data['state'] is not None:
+                self.latest_cob_state[symbol] = cob_data['state']
+                logger.debug(f"COB DQN state updated for {symbol}: shape {np.array(cob_data['state']).shape}")
+            else:
+                logger.warning(f"COB data for {symbol} missing 'state' field: {list(cob_data.keys())}")
             
             # If training is enabled, add to training data
             if self.training_enabled and self.enhanced_training_system:
@@ -1574,10 +1576,12 @@ class TradingOrchestrator:
             # Log statistics periodically (every 10 inferences)
             stats = self.model_statistics[model_name]
             if stats.total_inferences % 10 == 0:
+                last_prediction_str = stats.last_prediction if stats.last_prediction is not None else "None"
+                last_confidence_str = f"{stats.last_confidence:.3f}" if stats.last_confidence is not None else "N/A"
                 logger.debug(f"Model {model_name} stats: {stats.total_inferences} inferences, "
                            f"{stats.inference_rate_per_minute:.1f}/min, "
                            f"avg: {stats.average_inference_time_ms:.1f}ms, "
-                           f"last: {stats.last_prediction} ({stats.last_confidence:.3f})")
+                           f"last: {last_prediction_str} ({last_confidence_str})")
                 
         except Exception as e:
             logger.error(f"Error updating statistics for {model_name}: {e}")
@@ -2148,12 +2152,26 @@ class TradingOrchestrator:
             best_loss = model_stats.best_loss if model_stats else None
             avg_loss = model_stats.average_loss if model_stats else None
             
+            # Calculate reward for logging
+            reward, _ = self._calculate_sophisticated_reward(
+                predicted_action, 
+                predicted_confidence, 
+                actual_price_change_pct, 
+                time_diff_seconds / 60,  # Convert to minutes
+                has_price_prediction=predicted_price is not None
+            )
+            
             # Enhanced logging with detailed information
             logger.info(f"Completed immediate training for {model_name} - {outcome_status}")
             logger.info(f"  Prediction: {predicted_action} (confidence: {predicted_confidence:.3f})")
             logger.info(f"  {price_outcome}")
             logger.info(f"  Reward: {reward:.4f} | Time: {time_diff_seconds:.1f}s")
-            logger.info(f"  Loss: {current_loss:.4f} | Best: {best_loss:.4f} | Avg: {avg_loss:.4f}")
+            
+            # Safe formatting for loss values
+            current_loss_str = f"{current_loss:.4f}" if current_loss is not None else "N/A"
+            best_loss_str = f"{best_loss:.4f}" if best_loss is not None else "N/A"
+            avg_loss_str = f"{avg_loss:.4f}" if avg_loss is not None else "N/A"
+            logger.info(f"  Loss: {current_loss_str} | Best: {best_loss_str} | Avg: {avg_loss_str}")
             logger.info(f"  Outcome: {outcome_status}")
             
             # Add performance summary
diff --git a/web/layout_manager.py b/web/layout_manager.py
index 90e8615..7e64931 100644
--- a/web/layout_manager.py
+++ b/web/layout_manager.py
@@ -256,11 +256,12 @@ class DashboardLayoutManager:
         return html.Div([
             html.Div([
                 html.Div([
-                    # Chart header with manual trading buttons
+                    # Chart header with manual trading buttons and live price
                     html.Div([
                         html.H6([
                             html.I(className="fas fa-chart-candlestick me-2"),
-                            "Live 1m Price Chart (3h) + 1s Mini Chart (5min) - Updated Every Second"
+                            "Live Price (WebSocket): ",
+                            html.Span(id="current-price", className="ms-2 text-primary", style={"fontWeight": "bold", "fontSize": "1.2em"})
                         ], className="card-title mb-0"),
                         html.Div([
                             html.Button([