fix(cuda): disable hard stop-loss — dones override creates state mismatch

Setting dones[b]=1 in rl_drawdown_stop without actually closing the
lobsim position created a lie: Q trained on false trade-closes while
the position stayed open. Result: done_count=0 from step 99 onward,
990 trades total in 2000 steps (should be ~80k), training collapsed.

The drawdown penalty (per-step min(0, unrealized_r) × rate) is kept —
it creates exit gradient without corrupting the lobsim state. Hard
stop-loss needs to override the ACTION (to FlatL/FlatS) BEFORE the
lobsim step, not set dones after. Tracked for future implementation.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

crates/ml-alpha/cuda/rl_drawdown_stop.cu

@@ -39,10 +39,11 @@ extern "C" __global__ void rl_drawdown_stop(
         rewards[b] += unrealized_r * penalty_rate;
     }
 
-    // Hard stop-loss: force-close when drawdown exceeds threshold.
-    // unrealized_r is normalized by initial_r (trade-level vol), so
-    // threshold is in units of "initial risk multiples."
-    if (unrealized_r < -stop_thresh && dones[b] < 0.5f) {
-        dones[b] = 1.0f;
-    }
+    // Hard stop-loss DISABLED: setting dones=1 here creates a state
+    // mismatch — the lobsim position stays open while Q sees a false
+    // close. The done signal must come from actual position changes
+    // in the lobsim, not synthetic overrides. To implement hard
+    // stop-loss properly, the action must be overridden to FlatL/FlatS
+    // BEFORE the lobsim step, not after.
+    (void) stop_thresh;
 }