The long-awaited Aquarium 4.0.6 patch was finally released at the end of December. This is a huge update, involving more than one thousand changes, some of them major.

There is plenty of information available on the Internet about the new version, including several of my columns from last year. I think it’s time to give a high-level overview of this new version based on what has been written about it since the first hints of expected improvements appeared more than six months ago. The changes in the new version fall into these major categories:

• New features, many of which have been described in previous ChessOK Cafe columns. I know that many users appreciated seeing their suggestions being implemented in this version.
• Usability improvements, including a more responsive user interface and a more streamlined implementation of several features. Again, these changes were often based on user feedback.
• Major redesign of the internal structure of important components of the program. These changes are not immediately obvious, but affect the ability of Aquarium to handle large amounts of data, improve performance of some time consuming operations, reduce resource requirements, decrease start up time when there are many open databases, games, IDeA projects, etc. Language support was also redesigned making, it easier to translate Aquarium to different languages.
• And last but not least, bug fixes.

I have written a lot about IDeA lately, but as you can see from the following list (even if it’s far from being complete), the changes affect almost all parts of Aquarium:

Engine competitions. A large number of improvements. The same options are now available for both matches and tournaments (as far as possible). Games can begin from the initial position, from a specific position (FEN string), an EPD file with a collection of positions, a PGN game file, or normal opening books (both CTG and Aquarium books are supported). Asymmetric time controls are now tied to engine instead of color (allows handicapping an engine). More information is displayed when watching games. Additional information is recorded in engine games, including tablebase hits for each move. More detailed information is displayed when watching engine games.
Gauntlets have been improved and new engines can also be added to roundrobin tournaments in progress.

Infinite analysis. Improved support for multiple engine analysis and analysis presets. “Exclude moves” affects all running engines in a window. Copy all lines in analysis window in PGN format. Send analysis straight to IDeA. Improvements to infinite analysis history. Insert/copy sampled search variations.

IDeA. Major improvements in all areas, including fast minimaxing, master trees and game linking. Use a mixture of different computers and still get the same analysis quality from every engine. IDeA projects can now be easily shared. Really, there are too many important improvements to list here. See previous ChessOK Cafe columns.

Engines. Optionally extend tablebase path to subdirectories. Browse buttons for easily specifying files and folders in engine options. Export/import engine settings. Advanced logging features and CPU affinity control (tested on a forty-eight core system). Aquarium now communicates with engines using FEN instead of EPD.

Databases. Additional columns and sorting of game list. Faster conversion of databases. PGN handling was completely rewritten and is now faster than before. ECO classification.

Sandbox and game view. IDeA integration via game linking makes it easier to combine infinite analysis and IDeA.

Correspondence play. Various improvements to XfccPlay that can be used to play correspondence games on any Xfcc compliant server (such as ICCF). Aquarium has direct support for working with XfccPlay.

Remote engines. Setting up a server with remote engines in easier than before with the improved RTHomeServer. More robust handling of communication between Aquarium and remote engines. Off-line infinite analysis: Analysis can continue on the remote server while you are disconnected.

Trees. Display move colors from several trees in the same configuration. Each tree configuration is now stored in a separate file that makes it easy to share them with other users. Folder structure for tree configurations. More flexible “Add moves to main tree.”

Aquarium Scripter. A powerful scripting language that allows users with programming skills to write their own scripts. Scripts can be used for processing databases, games and trees, create specialized engine analysis features, generate tasks for IDeA etc.

Customers who bought Aquarium 2010 have enjoyed free updates for over a year and the same updates apply to Aquarium 4. I think it’s fair to say that Convekta has been very generous and it’s even possible that one more free update will be released before the developers turn their attention to the next version. I wouldn’t be surprised if they slipped in a couple of improvements in addition to any bug fixes.

More New Features

There are a few new features in the 4.0.6 version that I haven’t described before, but deserve to be mentioned.

Compress infinite analysis trees. As most Aquarium users know, when you run infinite analysis, the results are automatically stored and can be viewed when you visit the same position at a later time, even if it occurs in a different game or database. The infinite analysis history is stored in a position database or a tree (infinite_analysis), and now you can see the size of the tree from within Aquarium. If the current position has been previously analyzed, you will see its evaluation in the status bar. When you hover over the evaluation with the mouse pointer you will see the analysis results.



Version 4.0.6 adds information about the infinite analysis trees (infinite_analysis and infinite_analysis$1), both the number of positions (kilonodes) and their size in MB. In this example the size of infinite_analysis is 165MB and infinite_analysis$1 is 296MB. Since users analyze many unrelated positions, the trees become fragmented over time. A new menu item has been added to the infinite analysis history to compress the trees.



When you right-click over the infinite analysis history in the status bar, the menu shown in the image appears. Select “Compress trees” and Aquarium will start compressing the trees. This may take some time depending on their size. When it’s done you can see the results by using the same method as was shown above. Note that you may need to move to a different position for the display to be updated.



The compression has reduced the size of infinite_analysis from 165MB to 13MB (less than 10% of its original size) and infinite_analysis$1 from 296MB to 54MB.

User selectable font size in tree window. The font size in the tree window can adjusted. Changes affect all columns and column headings. Right-click over the tree window and select “Font Size” from the menu.

The default size is ten, as shown in the screen-shot. If you want to increase the size, try eleven or twelve.

Quick restore of IDeA tree backup. Aquarium stores multiple backups of IDeA trees, so a serious loss of data should be very rare. Besides keeping backups of the trees themselves, the analysis results are also stored in EPD format. Aquarium 4.0.6 uses this EPD file if you need to rebuild the tree.



While in IDeA mode, click the “Manage” link in the “Stage Status” window and select “Restore Tree” from the menu. Aquarium will rebuild the tree.

Insert IDeA engine analysis in notation. While IDeA is running you can click on a task in the queue that is being analyzed and follow the analysis of the engine. Starting with version 4.0.6 you can also insert the engine variation into the notation.



Start by selecting a task in the queue that is being analyzed. In this example I have clicked the leftmost active task, which is highlighted with a black frame. Below the queue, you can see the evaluation of the chess engine, the analysis depth, the time, and the main variation (PV). Clicking the “Insert” link inserts the variation along with its evaluation into the notation. This can be convenient if you see an interesting variation that you wish to explore further.

Convert IDeA task to infinite analysis. There are two reasons why you may want to convert an IDeA task into an infinite analysis task, which runs until you stop it manually. Sometimes you may simply want to decide yourself for how long a position is analyzed. In other cases, you may be waiting for the analysis of the last task in the queue to finish. If you turn it into an infinite analysis task, the stage will finish earlier and the task will continue to be analyzed in the next stage.



If you want to convert an IDeA task into infinite analysis, start by selecting a task that is being analyzed (a “green” task in the queue) and then click the “Infinite” link below the queue, as shown in the image above. The link will disappear and instead the following text is displayed: “Infinite analysis (press ‘Finish task’ to end).” This means that the engine will continue analyzing the position until you click the “Finish task” link, which you can see in the image.

If an IDeA task is converted to infinite analysis task, then Aquarium will not wait for it to finish at the end of the stage. If the minimum analysis time is reached for the task (task gets a white “f” status), then IDeA will close the current stage and proceed to the next one. The infinite analysis task will continue to be analyzed.

Deactivate all IDeA projects. A new command has been added for deactivating all IDeA projects. It can be useful if you are analyzing more than one project at the same time.



It is implemented as a split button. If you click the menu part of the Active button, the menu displays this new option that you can select to “Deactivate all projects.”

Insert/copy line from Sampled Search window. As with previous versions, you can browse the variations in the Sampled Search window when analyzing with Rybka 3 and later. In Aquarium 4.0.6 you can also insert its contents into the notation or copy the variations in PGN format.



If you see an interesting variation, right-click over the Sampled Search window and then select “Insert analysis” or “Copy to clipboard” from the menu.

Use of IDeA tree configurations simplified. The IDeA tree configuration is changed in Aquarium 4.0.6. First of all, you no longer need to specify a tree configuration when creating a new IDeA project. The predefined IDeA configuration will always be used. It also displays more information than it did in older versions. Here is an example showing a screen-shot of the default IDeA tree window as it looks in version 4.0.6.



The new columns are as follows:

Master: Shows the evaluation stored in the master tree (if you use one). For information about master trees, see IDeA Master Trees and IDeA Game Links.

IA: The engine evaluation of the position. The evaluation is taken from the infinite analysis history. This is accessible from the status bar, as described above.

In the screen-shot above, you can see that Rybka 4 evaluated the position as – 0.85 (see the IA column). After running IDeA for some time, the analysis tree was exported to the master tree and at that time IDeA’s score for 38…b6 was – 1.91 (see the Master column). Since then IDeA has continued analyzing and the current score is -2.13 (see the Eval column). Comparing these three columns gives you useful information; for instance, about the progress you are making.

If you want to know more about the evaluations in the IA column, you can hover with the mouse pointer over the IA history field in the status bar to display additional information.

If you look at the IDeA tree configuration, you will see that it contains some additional columns, which are hidden by default.



Besides the default columns (described above), you can display three more data columns, by selecting the highlighted check boxes in the screen-shot (in the Visible column). These columns are “Total,” which shows the number of games where the move has been played (based on HugeBase); “Success,” which shows how well the move scored in those games; and “CAP,” which is the move score from the CAP project.

The purpose of these changes was to simplify IDeA and give users easy access to automatically collected information that they otherwise might need to look up elsewhere in Aquarium or store manually (on a piece of paper, in another system, etc.).

This concludes my description of the many improvements in Aquarium 4.0.6. Next month I’ll talk about something completely different!