A/B testing compares gameplay outcomes by assigning different balance versions to separate groups of users. Comparing retention between groups helps assess whether a balance change made a difference. section 9.6
Testing a gold reward
Give group A 50 gold for completing a level and group B 100 gold, then compare retention. Keeping other outside factors similar helps attribute a difference to the reward change. Multiple balance versions can also test several changes at once. section 9.6
Figure 9.9 We can split our users into two groups and have one group load balance A, and the other can load balance B. Then we can measure the retention of each group to determine which balance performs better.

The diagram shows assigning balance A and B to separate user groups and comparing their retention. section 9.6




