Reward-Adaptive Iterative Discovery: A Case Study on Automated Game Testing for NHL26 | Digital Library | PAMCET | PAMCET