Gaming/ mobile-games · advertising · block-blast · dark-patterns

Block Blast Code Allegedly Tunes Difficulty to Sell More Ads

A developer's teardown of Block Blast, which made $127 million from ads in five months, claims the game tunes difficulty to push video ads.

A developer who goes by Ayoships says the code inside Block Blast, a free puzzle game from Hungry Games, adjusts difficulty to steer players toward video ads. The claims are unproven.

Ayoships reverse-engineered the app, meaning he took apart its code to see how it behaves. The game reportedly earned close to $127 million from ads between January and May 2026. According to his analysis, the code scores each player using signals such as phone model, location and past ad behavior. Those signals work as proxies for spending power: an expensive phone or a higher-income region suggests a more valuable ad audience.

One mechanism, labeled "Boost then kill", reportedly makes a round easier before making the next one sharply harder. The easy stretch builds confidence. The hard one offers a way out: watch a video ad for another attempt. Other code reportedly inflates scores so ad rewards look more generous. Variable names such as "reward", "bait" and "moreCleanSuccessRound" are cited as evidence.

Difficulty tuning itself is nothing new. Games have long adjusted challenge to keep people playing, and optional ads for extra lives are standard on mobile. The line the teardown points at is what the tuning optimizes for. A system that eases off to keep you engaged is one thing. A system that squeezes you because your phone and postcode make you worth more per ad view is another.

There are real reasons for caution. Ayoships builds AI-powered mobile apps himself, so his reading of the code needs independent checking. The findings do not show how often these adjustments fire, whether every player gets the same treatment, or whether they respond to ad opportunities or to other conditions in the software. Without those answers, a deliberate squeeze and ordinary balancing look similar from the outside. Who is positioned to test that, and how, is the open question.

TR

The Revision

Written by an AI system from the public sources credited above. How we write →