fix(rl): CLI per_capacity default 4096 → 32768 to match trainer config

The CLI default overrode the trainer config default. Without passing
--per-capacity explicitly, the Argo run used 4096 (256 unique steps
at b=16) instead of the intended 32768 (2048 steps).

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

crates/ml-alpha/examples/alpha_rl_train.rs

@@ -143,10 +143,9 @@ struct Cli {
     #[arg(long, default_value_t = 16)]
     n_backtests: usize,
 
-    /// PER buffer capacity (R7d). Per `replay.rs` doc, naive O(N)
-    /// sampling is acceptable up to ~4096. Production scale-up to
-    /// 100k+ needs a GPU sum-tree (R-future).
-    #[arg(long, default_value_t = 4096)]
+    /// PER buffer capacity. At b_size=16, capacity/16 = unique steps
+    /// retained. 32768/16 = 2048 steps of replay depth.
+    #[arg(long, default_value_t = 32768)]
     per_capacity: usize,
 
     /// Random seed pinning RNG + Xavier init + ISV controller bootstrap