Shift adaptive learning to growth sizing

tests/conftest.py

@@ -38,6 +38,7 @@ def make_settings():
             learning_lookback_trades=120,
             learning_min_samples=3,
             learning_max_adjustment=0.12,
+            learning_max_position_multiplier=1.6,
             llm_advisor_enabled=False,
             ollama_base_url="http://192.168.0.210:11434",
             ollama_model="gemma4:e4b",

tests/test_learning.py

@@ -60,16 +60,18 @@ def test_trade_learner_builds_adaptive_rules_for_losing_ema_exit(make_settings,
     state = learner.refresh()
     rules = learner.rules_for("BTCUSDT", "нейтрально")
 
-    assert state.adaptive_rules["risk_mode"] == "defensive"
+    assert state.adaptive_rules["risk_mode"] == "selective"
+    assert state.adaptive_rules["trade_permission"] == "selective_growth"
+    assert state.adaptive_rules["reduce_exposure"] is False
     assert rules["ema_exit_mode"] == "profit_only"
     assert rules["effective_entry_threshold_adjustment"] > 0
-    assert rules["min_hold_seconds"] > settings.min_hold_seconds
+    assert rules["min_hold_seconds"] == settings.min_hold_seconds
 
 
-def test_trade_learner_enters_capital_protection_and_validates_rules(make_settings, tmp_path) -> None:
+def test_trade_learner_blocks_only_bad_symbol_pattern(make_settings, tmp_path) -> None:
     settings = make_settings(tmp_path, learning_min_samples=3, starting_balance_usdt=100, max_total_exposure_usdt=80)
     storage = Storage(settings.database_path)
-    for value in (-0.12, -0.10, -0.08):
+    for _ in range(9):
         storage.insert_trade(
             Trade(
                 id=None,
@@ -78,7 +80,7 @@ def test_trade_learner_enters_capital_protection_and_validates_rules(make_settin
                 qty=1,
                 entry_price=100,
                 exit_price=99,
-                net_pnl=value,
+                net_pnl=-0.10,
                 reason="краткосрочный тренд ослаб ниже EMA50",
                 entry_pattern="нейтрально",
                 entry_confidence=0.7,
@@ -91,9 +93,44 @@ def test_trade_learner_enters_capital_protection_and_validates_rules(make_settin
     state = learner.refresh()
     rules = state.adaptive_rules
 
-    assert rules["trade_permission"] == "capital_protection"
-    assert rules["reduce_exposure"] is True
-    assert rules["bad_market_entry_block"] is True
-    assert rules["target_total_exposure_usdt"] == 35.0
+    assert rules["trade_permission"] == "selective_growth"
+    assert rules["reduce_exposure"] is False
+    assert rules["bad_market_entry_block"] is False
+    assert rules["target_total_exposure_usdt"] == settings.max_total_exposure_usdt
+    assert rules["blocked_symbols"] == ["ETHUSDT"]
+    assert rules["blocked_patterns"] == ["нейтрально"]
     assert rules["validation"]["status"] == "accepted"
     assert rules["validation"]["avoided_loss_usdt"] > 0
+
+
+def test_trade_learner_scales_positive_expectancy_setups(make_settings, tmp_path) -> None:
+    settings = make_settings(tmp_path, learning_min_samples=3, learning_max_position_multiplier=1.5)
+    storage = Storage(settings.database_path)
+    for _ in range(4):
+        storage.insert_trade(
+            Trade(
+                id=None,
+                symbol="SOLUSDT",
+                side="SELL",
+                qty=1,
+                entry_price=100,
+                exit_price=101,
+                net_pnl=0.08,
+                reason="test",
+                entry_pattern="пробой вверх",
+                entry_confidence=0.8,
+                opened_at=utc_now(),
+                closed_at=utc_now(),
+            )
+        )
+
+    learner = TradeLearner(settings, storage)
+    state = learner.refresh()
+    rules = learner.rules_for("SOLUSDT", "пробой вверх")
+
+    assert state.adaptive_rules["trade_permission"] == "positive_expectancy"
+    assert state.adaptive_rules["risk_mode"] == "growth"
+    assert rules["effective_entry_threshold_adjustment"] < 0
+    assert rules["symbol_position_size_multiplier"] > 1.0
+    assert rules["pattern_position_size_multiplier"] > 1.0
+    assert 1.0 < rules["effective_position_size_multiplier"] <= settings.learning_max_position_multiplier

tests/test_strategy.py

@@ -229,6 +229,40 @@ def test_strategy_uses_fractional_kelly_position_size(make_settings, tmp_path) -
     assert signal.diagnostics["position_notional_usdt"] == settings.max_position_usdt
 
 
+def test_strategy_scales_kelly_with_positive_learning_multiplier(make_settings, tmp_path) -> None:
+    settings = make_settings(tmp_path, max_position_usdt=50, kelly_fraction=0.25, kelly_max_fraction=0.20)
+    strategy = SpotStrategy(settings)
+    ticker = Ticker(
+        symbol="BTCUSDT",
+        last_price=101,
+        bid=100.99,
+        ask=101.01,
+        turnover_24h=10_000_000,
+        volume_24h=1000,
+        change_24h=1.0,
+    )
+    common = dict(
+        symbol="BTCUSDT",
+        candles=_ready_candles(),
+        ticker=ticker,
+        open_positions_for_symbol=0,
+        forecast={"usable": True, "probability_up": 0.62, "volatility_percent": 0.2},
+        account={"equity": 200.0},
+    )
+
+    neutral = strategy.entry_signal(**common)
+    scaled = strategy.entry_signal(
+        **common,
+        learning={"adaptive_rules": {"effective_position_size_multiplier": 1.5}},
+    )
+
+    assert neutral.action == "BUY"
+    assert scaled.action == "BUY"
+    assert scaled.diagnostics["position_sizing"]["risk_multiplier"] > neutral.diagnostics["position_sizing"]["risk_multiplier"]
+    assert scaled.diagnostics["position_notional_usdt"] > neutral.diagnostics["position_notional_usdt"]
+    assert scaled.diagnostics["position_notional_usdt"] <= settings.max_position_usdt
+
+
 def test_strategy_buys_probabilistic_rebound_after_stabilized_drop(make_settings, tmp_path) -> None:
     settings = make_settings(tmp_path, rebound_entry_confidence=0.58, rebound_min_probability=0.58)
     strategy = SpotStrategy(settings)