Version History

•	Correction for incorrect detection of stalemate in positions with white pawn capture moves.

•	Support for Polyglot opening books has been added, including support for Cerebellum Polyglot books with multiple moves.

•	Houdini 6 Pro: A correction made to Nalimov code to prevent multi-threaded crashes.

•	A new "Hide Redundant Output" option makes the engine output cleaner by removing some output lines. The option hides information about fail high and fail low situations, and statistics at the end of each search iteration.

•	When analyzing positions with the "Analysis Contempt" option checked, the Contempt value will now apply to the White side (instead of to the side that is to move). To use contempt that is optimistic for the Black side you should enter a negative Contempt value.

Houdini 6 continues the impressive surge made by its predecessor by adding another 50 to 60 Elo in strength, to become once again the strongest engine that money can buy.

Every aspect of the engine has been thoroughly reviewed and refined: evaluation, search, time management and table base usage.

Houdini 6 makes better use of multiple threads and will be particularly strong on hardware with many cores. To take into account the recent availability of mainstream processors with 8 cores, Houdini 6 Standard version now supports up to 8 threads. On high-end hardware Houdini 6 Pro can make use of up to 128 threads.

The evaluations have again been calibrated to correlate directly with the win expectancy in the position. A +1.00 pawn advantage gives a 75% chance of winning the game against an equal opponent at blitz time control. At +1.50 the engine will win 90% of the time, and at +2.50 about 99% of the time. To win nearly 50% of the time, you need an advantage of about +0.60 pawn.

The enhanced Tactical Mode turns Houdini 6 into the most impressive position solver ever, improving on the records the previous Houdini versions set for solving tactical test suites. When running multiple threads the Tactical Mode can now be activated for only some of the threads.

•	The Chess960 castling codes are now correctly handled. As a bonus, the Chess960 support is now compatible with Arena which uses the non-standard O-O and O-O-O as Chess960 castling moves.

•	Houdini 5 Pro: Setting the thread affinities now also works on Windows 7 and Windows Server 2008.

•	Work-around for some GUI's not showing engine output at depth 1: Houdini now generates a line with depth 99 in table base positions.

•	The behavior of the MultiPV_cp parameter ("Delta") in Aquarium is reverted to what it was in Houdini 4, which seems more intuitive to Aquarium users.

•	For the "Contempt" option the minimum value has been changed from 0 to -10. This means that Houdini 5 can now play with negative contempt, which is not recommended but interesting to try.

•	A new "Move Overhead" option for users playing engine games in which the move transmission has a time lag; the option specifies the extra time needed for each move in milliseconds.

•	The new "UCI Log File" parameter, when set, will produce a log file of all the UCI communication of the engine.

Released nearly 3 years after the previous version, the eagerly anticipated Houdini 5 is a rejuvenated and vastly improved engine.

Incorporating techniques and ideas developed over the past three years, Houdini 5 is about 200 Elo stronger than its predecessor. An important contributor to this giant leap in strength is the completely rewritten positional evaluation function. More aggressive King safety and sophisticated piece mobility are the keys to Houdini 5 unique playing style, combining aggressive ideas with sound positional judgment and plans.

By being more selective in its move choices, Houdini 5 now searches significantly deeper than before. This pays huge dividends in every stage of the game, but particularly in end games. For the best end game play, the new Houdini continues to support Syzygy end base tables, which can now be combined with the Nalimov tables to find the shortest path to mate.

The new Houdini implements the so-called "lazy SMP" algorithm and performs better than before on hardware with multiple processor cores. Up to 128 threads are now supported by the Houdini 5 Pro engine.

All major features of previous Houdini versions are still available: multi-PV analysis with user-definable evaluation threshold, Chess960 support and Persistent Hash (saving analysis results to a disk file for later reuse). The unique Tactical Mode has been updated and is now stronger than ever to find tactical solutions in positions.

A development version of Houdini 5 achieved a resounding victory with a score of 56/62 (50 wins and 12 draws) in the TCEC Season 9 Rapid tournament that featured the best 32 engines in the world. The final release version of Houdini 5 is about 30 Elo stronger than the engine that played in the TCEC tournament.

Houdini 4 contains many evaluation and search improvements in all phases of the game and is about 50 Elo stronger than its predecessor. The improved strength means that Houdini 4 nearly doubles the chess performance of your PC for game analysis and match play.

A major feature is the support of the new 6-men Syzygy table bases. Houdini 4 has been extensively tuned for this state-of-the-art end game table base system and will provide optimum precision and strength in the late endgame at any time control and with any number of cores.

While we strongly recommend using the 6-men Syzygy tables, Houdini 4 continues to support Nalimov and Gaviota table bases.

Houdini 4 maintains and improves on all the goodies that made previous Houdini versions the preferred analysis tool for many chess players. The Tactical Mode of Houdini 4 will find more solutions and provide faster solution times in tactical test suites than any other engine.

Persistent Hash and Learning are essential tools for storing your valuable analysis results for future use.

Houdini's very efficient multi-PV analysis and its support for Chess960 a.k.a. Fischer Random Chess remain the reference in the industry.

Houdini 4 uses calibrated evaluations in which engine scores correlate directly with the win expectancy in the position. A +1.00 pawn advantage gives a 80% chance of winning the game against an equal opponent at blitz time control. At +2.00 the engine will win 95% of the time, and at +3.00 about 99% of the time. If the advantage is +0.50, expect to win nearly 50% of the time.

Houdini 3 contains many evaluation and search improvements in all phases of the game and is about 50 Elo stronger than its predecessor Houdini 2 (which at the time of writing was still unmatched by any other engine). The opening improvements relate mostly to piece mobility and space management and are most convincingly demonstrated by the progress in Fischer Random Chess for which Houdini 3 has become about 75 Elo stronger. In the middle game Houdini 3 has significant enhancements for recognizing pieces with limited mobility and in king-side safety. Houdini 3 will seek deeper in end games and solve more positions than before.

In the new Tactical Mode Houdini 3 will prefer tactical instead of positional solutions. In tactical test suites the Tactical Mode will find more solutions and provide significantly faster solution times, often by a factor of 3 to 10.

The Accelerated Principal Variation Search or “Smart Fail-High” is especially useful in very deep analysis when a different move becomes best at very high search depth. Houdini 3 will apply an automatic depth reduction that often speeds up finding the Principal Variation by a factor of 5 to 10.

Besides Nalimov and Gaviota End Game Table Bases, Houdini 3 now also supports Scorpio bitbases. These bitbases are loaded in memory when the program starts (requiring about 300 MB of memory) and are then readily available to the engine.

Hash usage has been optimized, improving back-tracking analysis. Houdini 3 Pro will now support hash tables up to 256 GB.

The engine evaluations have been carefully recalibrated so that +1.00 pawn advantage gives a 80% chance of winning the game against an equal opponent at blitz time control. At +2.00 the engine will win 95% of the time, and at +3.00 about 99% of the time. If the advantage is +0.50, expect to win nearly 50% of the time.

Fun fact: Over 10 million chess games were played for the development and tuning of Houdini 3!

•	Nalimov EGTB probing would not recognize KvK end game.

•	Nalimov EGTB: Houdini would exit when a corrupt table base file was encountered.

•	On some Windows systems the internal timer would overflow after approx. 30 minutes.

•	New MultiPV_cp option to limit multi-PV analysis to moves within a range of the best move.

•	New FiftyMoveDistance option to make the 50-move rule kick in earlier.

•	New UCI_Elo and UCI_LimitStrength options as UCI standard-compliant alternative to Strength option.

•	Houdini now exits when it detects that the communication with the GUI is broken.

•	Repetition in games would have Houdini evaluate the repeating position as 0.00.

•	When specifying a fixed depth search, Houdini would wait for a "stop" command when analysis finished early.

•	Omitting the "uci" command at start-up would produce erratic results in the first analysis.

Improved Analysis support: Save Hash to File, Load Hash from File, Never Clear Hash

The complete hash table can be saved to a disk file and reloaded later for continuing the analysis.

Houdini 2 makes better use of the data in the hash table to restart instantly the analysis at the point where it was previously interrupted.

By saving the hash table in a disk file you can interrupt the analysis anytime. At a later date you can reload the hash table file in memory to continue the analysis as if no interruption had happened.

The new "Never Clear Hash" option will keep the hash data in memory even when the position changes or when starting a new game.

The "Position Learning" mode will automatically save analysis results in a learning database that will be reused in future analysis. This is a more convenient, but less powerful method of using previous analysis results than the manual Save Hash/Load Hash operations of the previous point.

The strength is limited through a combination of techniques: limiting the number of positions searched, purposely picking a non-optimal move, and ignoring table bases.

To speed up solving deep mates, you can fix a limit on the search depth during the mate search.

For advanced servers with many cores on multiple sockets (usually with so-called NUMA architecture), a special Houdini Pro version is available. It significantly improves the scaling of Houdini up to 32 cores and makes use of up to 32 GB of hash memory.

Improved evaluation and search make Houdini 2 Standard about 25 Elo stronger than its predecessor Houdini 1.5a. On high-end hardware Houdini 2 Pro with its additional NUMA-related speed improvements is about 40 Elo stronger than Houdini 1.5ab-16.