MoyoGo knows 17 million invariant patterns and sorts them by order of Statistical Move likelihood: How many moves, on average, a player waited to play there. Pattern info includes Ko status, edge proximity, chain lengths and -liberties.

This system is unique and is able to predict pro moves better than any other Go software. It is less suitable for complex tactical situations but it is very useful deep into the opening. It represents a milestone in computer Go research and three Microsoft researchers are still trying to reproduce it.

The patterns are divided into the shapes shown here. During a pattern search, scrolling the mouse wheel down moves through the smaller patterns.

Most Fuseki, all Joseki and a huge, collection of "Good Shape" and Tesuji patterns are immediately available by simply pressing Ctrl+Alt and moving the mouse over the board.

Press AltGr or Ctrl + Alt and move the mouse over the board to see common patterns and their Statistical Move Likelihood.

The search results list immediately shows all games in the enabled databases that contain a move on that pattern.

The most notable points on the board are automatically annotated in blue letters:

How many of them are shown can be changed in Settings -General Settings - Patterns - Maximum number of shown pattern annotations. Their color can be changed in Settings - General Settings - Example Patterns (Statistical Move Urgency Analysis).

- The lighter blue the pattern annotation, the larger the pattern perimeter (area size).

- "A" has the highest statistical move likelihood, "Z" the lowest.

- You can turn off board annotation with the toolbuttons.

- You can make the annotation colors less "visible" with the annotation settings.

- You can get rid of the gradient by making both colors the same in annotation settings.

Press AltGr or Ctrl + Alt and move the mouse over a point to see its pattern perimeter. Use the mouse wheel to change pattern size. The statistical move likelihood is displayed as a tooltip, being:

    How often the pattern (with empty center point) occurred in all games

------------------------------------------------------------------------------ x 100%

How many turns the average player waited to play on the pattern's center point

Also see Understanding Statistical Move Likelihood and a detailed description in this PDF file (needs Adobe Acrobat Reader).

The list of pattern results includes the average rank of the player that played that move and the average winning percentage of games in which that move was played.

Games can be shown on a new Go board, at the significant cost of extra memory (RAM) use. Games are added as a new board tab with Ctrl + Double-Click or pressing Enter.

A simple Double-click swaps the current (focused tab) game with the selected game.

Fastest scrolling is done by Ctrl + Page Up/Down, which prevents the diagram from updating whilst scrolling.

Scrolling through that list, the moves are highlighted in a diagram and it is possible to play through a game in the diagram.

Another way to see a found game position is to double-click on a result to show the game position on the board. Ctrl + Z (undo) restores the original position.

Ctrl + Right and Ctrl + Left moves through the game in the diagram.

Pattern examples are automatically rotated, mirrored and color-reversed to match the pattern on the board. Also when a game result position is loaded, the board position is matched to the pattern. Use the rotate/mirror/color reverse toolbar to get a loaded search result game's colors back to original.

Any of Moyo Go's 17 million pre-analyzed (most common) patterns can be given a name by right clicking and choosing Comment shown pattern. This name will appear in the pattern results window and in the status bar, next time that pattern is encountered in another game. Also see Pattern descriptions.

Pattern names are stored in a database, not in the SGF game record itself. Comments on the game position should be done in the Editor.

Pattern names can be looked up in Sensei's Library by right-clicking and selecting Sensei search pattern.