Battle School isn't a military academy—it's a feedback loop designed to extract maximum performance by making surveillance invisible and advancement the only visible truth.
Ender never sees his scores, yet the ranking becomes the only information available to him about whether he matters.
The teachers watch him constantly—Graff and the others observing from above, taking notes, adjusting the game difficulty when Ender plateaus—and he competes anyway, climbing the rankings with no choice but to chase the metric he cannot see. He interprets this silence as evidence of his own inadequacy and pushes harder.
His friends watch him climb and compete to match him, so the competition becomes contagious not because anyone explained why it matters, but because the visibility of ranking creates the appearance of objective truth. Graff's design principle is pure engagement mechanics: isolation plus transparent ranking plus continuous nudging of difficulty equals voluntary, unsupervised escalation. The system doesn't tell Ender why he's losing or winning increasingly difficult battles—it just shows him the next opponent, and he optimizes himself without knowing he's being optimized.
The prediction is that you can make a child destroy himself by making him believe destruction is ascension, and he'll thank you for the opportunity to compete.
He never consents because he never sees the architecture, and the prediction is that you can make a child destroy himself by making him believe destruction is ascension, and he'll thank you for the opportunity to compete.
Read Tristan Harris's essay 'How Technology Hijacks Your Brain's Reward System' and notice that every mechanic he describes—the infinite scroll, the variable reward, the phantom notification—exists in Battle School's command hierarchy first.
Look for Andrew Critch's technical analysis 'Security Concerns with AI Trained on Human Feedback' which uses Ender's Game as a test case for how systems that optimize for human preference can produce outcomes that horrify the humans being optimized for.