Toggle menu
Toggle preferences menu
Toggle personal menu
Not logged in
Your IP address will be publicly visible if you make any edits.

Maia Engine

From Repertoire Builder
Subscribers Only: This functionality is available for Basic and higher subscription tiers.

Available in: Builder, Library, Practice, Tools

The Maia Engine card provides a human-like evaluation of the current position by showing the move probabilities predicted by Maia, a neural-network model trained on millions of real human games at different skill levels. Unlike traditional engines that aim for perfect play, Maia focuses on how humans actually respond, making it a powerful tool for practical opening preparation, training, and pattern recognition inside the Repertoire Builder.

Selecting Maia Level: You can change the Maia Engine level at any time by clicking the Maia XXXX label in the top-left corner of the card. Levels range from 1100 to 1900, each representing how human players at that rating typically behave.

For the current position, this section displays:

  • Move: Maia’s most probable human-style move for the selected rating level.
  • Probability: The likelihood (percentage) that a human player at this level would choose the move, based on Maia’s prediction.
  • Level Differences: A comparison showing how the move’s probability changes across all Maia rating tiers relative to your selected level. Each tier is represented as a small indicator: green arrow up for higher likelihood, red arrow down for lower likelihood, a dash for negligible difference, and an empty circle if the move does not appear at that level. Hover over any indicator to view the exact percentage difference.

At the bottom of the card, you can also see whether the move source was Opening Book or Maia Generated. Maia begins making predictions after ply 10; before that point, the moves are drawn from the Lichess Opening Book at the corresponding rating level to ensure accurate early-game data.

You can click on any move to add it directly to your opened item, or right-click to add the move together with its associated probability data as a comment for immediate reference.

This feature helps you:

  • + Study positions using realistic, human-style move predictions instead of perfect engine lines
  • + Understand the most natural moves and common mistakes made by players at different ratings
  • + Compare theoretical engine choices with realistic human tendencies for better preparation
  • + Add Maia-recommended moves directly into your repertoire or study material
Note: This feature leverages the open-source Maia project: https://github.com/CSSLab/maia-chess
Prev: K-Maps