analysis
Compare saved analyses
Switch saved engine runs with numbered buttons or keys 1–9, and inspect each run's recorded settings.
Choose a saved run
Open a game and select Analysis. The Analyze / Analyze again controls and engine selector appear first; Saved Analysis is below them. Choose a run in its dropdown. When multiple runs are available, numbered gray raised buttons provide a quicker way to switch.
Click a numbered button or press 1–9 for the corresponding run. Numbers follow the dropdown order; buttons beyond 9 remain clickable. The dropdown and highlighted number stay synchronized after the selected run loads. Refreshing the list or saving another run can change the order.
Switching is unavailable while full-game analysis is running or while a run is loading. Number shortcuts do not select runs while you are typing or using a modified key combination. In the Analysis view, these numbers take priority over observed FICS-game shortcuts.
Analyze several saved games
Select two or more checkboxes in Saved Games. The grouped selection console offers an analysis-engine dropdown, Verify critical moves enabled by default, and a gold Analyze N games button. First-pass positions use one second each; critical positions use five seconds each in the second pass. The modal names the active pass and reports progress, completion and failures. Cancel preserves completed work, including partial verification.
A successful batch save is read back before being counted as complete. Previous runs and annotations are preserved. The library includes retained runs when showing analysis status. If the current snapshot is empty, reopening can restore a retained run only when its recorded positions match the game.
An active Not analyzed filter hides newly analyzed entries and shows the next matching games. Clear the chip to see completed results. See Batch analysis.
If a selection does not load
Saved results must match the game’s positions and move sequence. An analysis from a different game is rejected with a visible error; your current results and the saved payload remain intact. Older browser records affected by a previous storage bug may contain analyses from another game. Those results cannot be applied to the game currently on the board.
Temporary profile reviews do not show the previously opened game’s saved-analysis selector. Explicitly saving a temporary review creates a separate game entry, so its future analyses stay with that game.
Compare inside the expanded graph
Choose Expand in the graph’s upper-right corner. The expanded view includes the same numbered buttons and 1–9 shortcuts, so you can compare runs without closing it. The graph, results, and engine heading follow the selected saved run. Returning to the main view shows the same selection in the dropdown.
Selecting a historical run does not change the engine configured for your next analysis. To create new results, choose the desired engine beside Analyze and start a new run. Your earlier saved runs and personal annotations remain separate.
Compare saved and live recommendations on the board
Compare saved line starts enabled in Move Analysis, under Analysis → Critical Analysis. A dashed blue arrow marks the saved recommendation. If its piece has already moved, a translucent blue ghost shows its original square, and a label identifies the position before the played move. The actual board stays unchanged. The setting is remembered; clear the checkbox to hide it.
Live Prediction uses solid green for its strongest candidate and red for alternatives, with matching score bubbles. These results come from the current analysis engine and may differ from the selected historical run. The blue comparison remains tied to the saved run. See engine lines and board markers.
Inspect the engine and settings
Hover any numbered button, in either view, for the run’s recorded engine, analysis budget, critical-move verification status, and engine options such as Threads and Hash. The tooltip also includes its saved-run label and number shortcut. Older runs may not contain every setting; missing information is marked as unavailable rather than taken from your current engine configuration.
Use Manage analyses for saved-run management. Use the refresh control when you need to reload the available runs. Companion-hosted results need their connected archive or a locally cached copy; an unavailable run reports a loading error rather than replacing the current results.