REBEL 11.0 review
by Keith Kitson
REBEL CENTURY 3
I have to say this is the most versatile and comprehensively specified DOS
based chess program on the market. Ed Schroeder and his team have solved
the assignment of Hash Ram to a DOS shell program running within the windows
environment. Two other notable DOS based programs MCP8 and Hiarcs 7.0/7.01
are not able to assign the same levels of Hash (i.e. upto 400Mb with Rebel,
depending on available memory on the system) running within a DOS shell.
The refinements built into successive versions have rendered Rebel as
windows like as possible. A comprehensive, intuitive, context sensitive
help facility, pull down menus, right click context sensitivity,
comprehensive layout facilities, this program has all the conveniences one
would expect of a mature chess program.
Position setup mode is a pleasure to use, it being so easy and intuitive.
Truly international in nature Rebel caters for 6 languages, although I stick
with English, more than a handful for me.
The EOC (Encyclopaedia Of Chess) facility has been enhanced and a utility is
provided to convert old EOCs to the new more efficient format. The EOC
supplied with REBEL has 20 million unique chess positions!
A slight change in approach has been made to the default layout to cater for
the presentation of more feedback information on the current position or
game in progress. The default board layout has been reduced to 75% to
accommodate this. Many other alternative layouts are available, and the
user has the opportunity to define their own layout by right-clicking and
dragging the individual windows as required, then saving the new layout for
CAT (Computer Analysis Tool) database, is a new feature provided with this
release. It is a database mechanism allowing the orderly storage of
computer v computer games and computer analysis for quick recall as
required. The big advantage of CAT is it can be switched into the program
and when a position is found in the CAT database it will use the information
enabling the engine to compute one ply deeper in most cases. Other features
of CAT include automatic learning, import of positions and analysis from
several chess programs to compare results to better determine which was the
best move generated. The CAT feature alone is worth the cost of Rebel!
Apart from the above Rebel is supplied with a database of games which has
been increased with this release by 300,000 games to total 800,000 high
The Openings book has been expanded with an extra 60,000 new book moves
concerning latest opening theory. Now with 46,000 variations the book
references some 2.6 million unique positions.
Ed has managed to find something in the region of +100 elo improvement in
Computer v Computer games. This is bourne out by the testing I have done
against the other top programs currently on the market. Elo of +100
equates to approx. +20 to +30 elo against human opposition.
Hash table persistency has now been incorporated. This helps where
calculated search results are stored in hash ram for future use on
successive moves. The search algorithm has also been speeded up.
Other interesting facilities, begging to be investigated are, Bluff Chess,
Tactical Engine and Club Player.
One of the most interesting facilities is the personality change. Various
parameters have been tapped into by the programmer and allow the user to
vary some of the critical criteria that the program uses to decide the way
in which the game will be played. A separate utility is built to set new
personalities, but they can also be set directly within Rebel, and saved
with a unique name for later recall.
I suspect we have not heard the last of the personalities facility. Rebel
was set to what was thought to be the strongest personality setting, but
extended parameter experimentation/testing has revealed small improvements
on the default engine. This suggests that there is room for the owner to
experiment and perhaps find a stronger version of Rebel within the
personality changes. Perhaps there is another +100 elo to be found by
experimentation. Who knows? The benefit is the fun in experimenting and
CHESS TIGER 13
This is the second major release of Chess Tiger under it's new Chess Partner
interface (at version 5.0 now). The GUI environment has been enhanced
considerably. The menu options are now Windows/Office 2000 styled.
Winboard engines are supported, and access to Chess Internet servers is
available directly from the Interface. Of particular note (though not a new
facility) is the ability to add an infinite number of new levels to the
existing mechanism. The owner has the ability to define their own tailored
levels and name them for seamless incorporation into the default levels
listbox, a very neat facility.
A version of Rebel is now available and can be switched into Tiger for
Hash Ram assignment for Tiger has been improved. Set the maximum available
for the system and let Tiger assign the necessary ram according to the level
of play set. Much more civilised.
Tiger now shows pondering during the opponents thinking time. This is a
good improvement for Tiger owners as the more information feedback that can
be presented the more interesting each game becomes.
This new release of Tiger makes it possibly the strongest single processor
chess program on the market at the time of release. All the top programs
are much closer in strength now than in years gone by. Therefore a
considerable number of games is required at a variety of time controls in
order to determine which programs have the advantage in various areas of the
game. It is becoming increasingly more difficult to say that one specific
program is overall strongest in all areas of play. My prediction is that
Tiger will be at the top, or at worst in the first three. With Gambit Tiger
(see below) perhaps not that far behind.
During the beta test phase Christophe Theron released a slight change to the
Tiger program for the testers to experiment with. It was supplied with the
note, not quite as strong but a different approach. Suddenly reports came
in from Beta testers of a revelation in chess playing ability. All games of
this changed engine were exciting, cavalier, edge of the seat stuff. As
more and more reports came in of its abilities Christophe decided to release
the engine in its own right, and due to the nature of the reports coming in
the new engine was given the title Gambit Tiger. It appears to be slightly
more careless in it's play sometimes in comparison with Chess Tiger but this
has not deterred it from winning two tournaments and creating a lot of
discussion about a new direction in Computer Chess.
In the main chess writers looked to build the strongest engines they could
produce regardless of how sedate the programs play was, suddenly Gambit was
showing a new approach (it had been thought about by others but never really
successfully implemented), a new exciting, attacking, cavalier style. Who
cared if it lost the odd game here and there, the games were never dull.
Christophe may have set a new trend or direction for the rest of the
marketplace to follow, or be left behind in the race.
Not only should a program play excellent chess, if the program can play in
an exciting and cavalier manner this has to be a crowd puller, and better
for chess in general. It should see more converts wanting to jump on the
bang wagon and experience this new style of play, for a chess computer.
My prediction is that Christophe has opened up a new direction for chess
programming in general. The chess marketplace must follow suit or fall by
Interestingly Rebel has this capacity built in to alter certain parameters
of the program to suit a more cavalier style of play.
The Rebel camp are ahead of the game.
Who will be the first to catch up?