The honest line

Mainline is not another puzzle trainer, game analysis tool, or spaced-repetition deck. This app sits one layer up: it is the orchestration layer that decides what you should work on, with which resources, and why, then revises that plan continuously as you train and play.

Progress in Mainline means training signals: whether you are showing up, completing the planned work, keeping reviews healthy, and building skill estimates with uncertainty. Rating is noisy, and no activity here is treated as proven to cause rating gain.

The engagement layer exists because consistency is part of training. It uses forgiving reminders, capped streak cycles, and competence feedback to make practice easier to resume. It does not use ads, leaderboards, shame, unbreakable streaks, or paywalled training quality.

Every exclusion is a design decision with a reason.

• No LLM/AI at runtimeAI plays chess poorly and invites cost, abuse, and opacity. The app is pure deterministic algorithms. You can verify every decision.
• No competing game platformLichess and Chess.com already do that better. The app references external platforms; it doesn't replace them.
• No hosted copyrighted contentBooks and courses are recommended and logged, never hosted. The app points you to the right resource; it doesn't steal it.
• No social or multiplayerNo leaderboards, no chat, no shared sessions. Social comparison harms long-term motivation, and the app's goal is personal improvement.
• No self-report skill diagnosisDunning-Kruger is real in chess (Grade A/1). The app diagnoses you behaviorally from your games and your puzzle performance, never by asking you to rate yourself.
• No infinite streaksUnbreakable streaks are a dark pattern. They create a loss-aversion "quit moment" when the streak breaks. The app caps streaks and forgives missed days.
• No global leaderboardsDownward social comparison (seeing yourself ranked below strangers) harms motivation for the majority of users who are not at the top.
• No puzzle-volume chasingCorrelation between puzzle volume and rating gap is r=−0.02. Grinding puzzles without reflection or spacing doesn't help. The app prioritizes how you practice over how much.
• No opening memorization for beginnersBeginners lose to blunders, not opening theory. Time spent memorizing lines at <1200 is time not spent on tactics and board vision.

Every recommendation, methodology value, and claim on this page carries a grade: a placeholder can never pose as established fact.

Grade answers how strong the science is. A second axis, confidence, answers a different question: how much of your own data backs this specific call to you. The same Grade-A finding can land with low confidence, as when we know spaced repetition works but you've only imported three games. Or it can land with high confidence, well-backed by your own play. The distinction keeps a band prior from masquerading as a personalised verdict.