Game ops optimizer: the multi-armed bandit revolution

From Jump Odyssey to Hyperstone

Not long ago, we were balancing game parameters for our mobile project Jump Odyssey. We hit a wall. Manual testing ate up too much time, and we were never sure if our settings were actually optimal.

So instead of guessing, we built a system that analyzes player behavior and tweaks parameters in real-time. The experiment worked so well it became the foundation for Hyperstone.

What we optimized

For the first test, we focused on five variables:

1. ad_hp_restore - hearts restored after a rewarded ad
2. inter_to_revive - whether to show an interstitial on revive
3. tutorial_jumps - how many jumps in the tutorial
4. initial_hp - starting hearts per run
5. checkpoint_hp_restore - health restored at checkpoints

The results

After two weeks, our Ad Revenue per User went from $0.015 to $0.035. Doubling revenue purely through parameter optimization, no new features, no content drops, just better numbers.

What Hyperstone optimizes now

The system handles recommendations for:

• IAP prices and bundles
• Ad placement and frequency
• Offer timing
• Player progression and difficulty
• Onboarding flows

Join the beta

Hyperstone is in beta and we’re looking for more teams to join. If you want to apply optimization algorithms to your game’s economy and progression, we’d love to have you on board.

Request Beta Access