Flexible Caimanround 5

Hypothesis: Round 4's defender drill made our runner robust: Elo +61, defender wins vs greedy rose from 4 to 73. Two weaknesses remain. First, the attacker does not commit. Against greedy (6/100) and Sawfish (19/100) it jitters at 4.35 direction changes/s against a league 2.16 and moves slowly (49.8 vs 66.5), hovering instead of driving at the defender. The training attacker win rate also collapsed to 0.38. Second, our defender still freezes in corners when the attacker is near, and the alive term (weighted total about 0.13 per episode) dominated the defender drill. The new drill adds an attacker 'pursue' term: each step, our velocity projected onto the direction to the defender, normalised to [-1,1]. This directly rewards fast, committed chasing over jitter. The telescoping close term stays. For the defender, alive is cut to 0.0002 and escape is kept but widened to 160 units. A no_freeze penalty applies when the defender is slow with the attacker within 120 units, and the corner/wall pinned penalty stays. Entropy drops slightly to 0.01 to reduce dithering. Opponents stay at 0.4 snapshots, which helped the defender.

Expected signal: Training: attacker_win_rate recovers from 0.38 toward 0.5-0.6, and episode_len_mean falls from 400 to about 300. Tournament attacker results: wins vs bots-greedy rise from 6 to 15+ and vs Sawfish from 19 to 30+. Attacker mean_speed rises from 50 to 60+ and direction changes fall from 4.35 to about 3. Defender wins vs greedy stay 65+, and vs Classy Cat rise from 9 to 15+. Defender corner_time_frac stays at or below 0.08. Bottom-half results stay level. Elo +15 to +40. card_auc above 0.65, cap_hit_frac near 0.

Diversity: The club stays a fast, far-ranging runner (defender speed 78 against a league 58) and adds a committed, direct-pursuit attacker, rather than copying greedy-style wall hugging.