checkpoint @ game=4000 steps=4000

training_log.txt

@@ -4299,3 +4299,13 @@
 [2026-06-23 10:19:24] game= 3988 | 1/2-1/2 | moves= 51 | buf=50,000 | steps=3,988 | loss=8.6245 (p=8.4493 v=0.1752) | game=5.2s train=90.5s
 [2026-06-23 10:21:01] game= 3989 | 1/2-1/2 | moves= 61 | buf=50,000 | steps=3,989 | loss=8.6384 (p=8.4493 v=0.1891) | game=6.1s train=90.5s
 [2026-06-23 10:22:37] game= 3990 | 1/2-1/2 | moves= 51 | buf=50,000 | steps=3,990 | loss=8.6048 (p=8.4493 v=0.1555) | game=5.2s train=90.7s
+[2026-06-23 10:24:17] game= 3991 | 1/2-1/2 | moves= 68 | buf=50,000 | steps=3,991 | loss=8.6262 (p=8.4493 v=0.1769) | game=6.8s train=90.5s
+[2026-06-23 10:25:52] game= 3992 | 1/2-1/2 | moves= 48 | buf=50,000 | steps=3,992 | loss=8.6521 (p=8.4493 v=0.2028) | game=4.8s train=90.3s
+[2026-06-23 10:27:27] game= 3993 | 1/2-1/2 | moves= 48 | buf=50,000 | steps=3,993 | loss=8.6219 (p=8.4493 v=0.1726) | game=4.8s train=90.9s
+[2026-06-23 10:29:08] game= 3994 | 1/2-1/2 | moves= 69 | buf=50,000 | steps=3,994 | loss=9.1417 (p=8.4493 v=0.6925) | game=9.6s train=90.8s
+[2026-06-23 10:30:51] game= 3995 | 1/2-1/2 | moves= 69 | buf=50,000 | steps=3,995 | loss=9.0791 (p=8.4493 v=0.6298) | game=12.1s train=91.1s
+[2026-06-23 10:32:27] game= 3996 | 1/2-1/2 | moves= 51 | buf=50,000 | steps=3,996 | loss=8.9128 (p=8.4493 v=0.4635) | game=5.1s train=91.1s
+[2026-06-23 10:34:04] game= 3997 | 1/2-1/2 | moves= 52 | buf=50,000 | steps=3,997 | loss=8.8539 (p=8.4493 v=0.4046) | game=5.3s train=91.0s
+[2026-06-23 10:35:39] game= 3998 | 1/2-1/2 | moves= 48 | buf=50,000 | steps=3,998 | loss=8.8067 (p=8.4493 v=0.3574) | game=4.9s train=90.8s
+[2026-06-23 10:37:16] game= 3999 | 1/2-1/2 | moves= 56 | buf=50,000 | steps=3,999 | loss=8.7946 (p=8.4493 v=0.3453) | game=5.7s train=90.8s
+[2026-06-23 10:38:53] game= 4000 | 1/2-1/2 | moves= 58 | buf=50,000 | steps=4,000 | loss=8.7273 (p=8.4493 v=0.2780) | game=5.9s train=91.1s