Excalibur Chess Computer Comparisons

This forum is for general discussions and questions, including Collectors Corner and anything to do with Computer chess.

Moderators: Harvey Williamson, Steve B, Watchman

Forum rules
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Excalibur Chess Computer Comparisons

Post by spacious_mind »

I have started this Post as a place holder for comparing Excalibur Chess Computers.

As a kick off, I have completed my game 1 and game 2 Tests with Excalibur Einstein LCD Chess Wizard and Excalibur Touch Chess II,. The tests are available for download below if someone else wants to try the tests on their computers.

http://hiarcs.net/forums/viewtopic.php?t=6835

Excalibur Einstein LCD Chess Wizard - 2008

Image

http://www.spacious-mind.com/html/einst ... izard.html

Excalibur Einstein Touch Chess II - 2007

Image

http://www.spacious-mind.com/html/touch_chess_ii.html

Test Game 1 - 15th Century Masters

Below is the result of the first test, both computers played at Level F030 which corresponds to 30 seconds per move.

Image

Both computers scored really well in the first test, both beating the scores of Tasc Chess Machine 512K 30MHz The King 2.54 Normal Setting and Mephisto Berlin 68000!!

But unfortunately, both computers have exactly the same program and replayed every single move the same. Therefore Excalibur Einstein LCD Chess Wizard and Touch Chess II are 100% clones.

Test Game 2 - The Turk

Image

The performance in this game brought these computers back to earth. But same result 100% clones.

The average rating over two games for both computers is: RATING: 1806

Best regards
Nick
User avatar
Monsieur Plastique
Senior Member
Posts: 1014
Joined: Thu Jul 03, 2008 9:53 am
Location: On top of a hill in eastern Australia

Post by Monsieur Plastique »

Good testing there Nick, but I am surprised at the derived rating. I realise the rating is earned by the testing process rather than actual games, but in my own experience these programs are quite a lot weaker than that - as I say, maybe around the low 1500s Swedish ELO or mid 1600s USCF - and even then that would be really pushing it. I'm probably around 1700 Swedish ELO and my record against these programs is something like 80% wins. And those are against the "premium" models which are supposed to be twice as fast and therefore around 60 points stronger if we go by Kaufmann all those years ago. That is why I am bit dubious about the high rating it's been awarded via this test procedure.

I'm almost certain if you played a 40 in 2 match between either of those computers and a Mephisto Europa or original Constellation (both considered high 1600 machines), it would lose very badly - probably only getting around 2.5 out of 10 at tournament time controls. That would also be consistent with my own estimates of low 1500s at best.

Then again, your testing is that faster time controls so I guess that would increase the rating over 40 in 2 somewhat.

I am interested to know which Excaliburs are still available new and are reasonably strong. I am also curious to know whether the very last of the Excaliburs had the bug resolved that plagued the Alexandra, Phantom Force and DTTC - that is they often lose on time at the 40th move at the 40 in 2 level. It is as if the program thinks it only needs to make 39 moves for the time control instead of the full 40. These machines also have a "dotted segment" LCD display (my terminology) that slowly corrupts after the first time control in a 40 in 2 game - the more moves made after move 40 the more corrupted the display becomes - sometimes to the point where you can no longer read the display and have to restart. This effected every one of the above models that I have come across, but I wonder if these cheaper models have the problem given they don't have the "dotted segment" displays in them.

I am hoping that Sir Dave might chime in with this thread as he owns the Pavilion model (or something like that) which I believe would also be a clone of the above two models.
Chess is like painting the Mona Lisa whilst walking through a minefield.
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Monsieur Plastique wrote:Good testing there Nick, but I am surprised at the derived rating. I realise the rating is earned by the testing process rather than actual games, but in my own experience these programs are quite a lot weaker than that - as I say, maybe around the low 1500s Swedish ELO or mid 1600s USCF - and even then that would be really pushing it. I'm probably around 1700 Swedish ELO and my record against these programs is something like 80% wins. And those are against the "premium" models which are supposed to be twice as fast and therefore around 60 points stronger if we go by Kaufmann all those years ago. That is why I am bit dubious about the high rating it's been awarded via this test procedure.

I'm almost certain if you played a 40 in 2 match between either of those computers and a Mephisto Europa or original Constellation (both considered high 1600 machines), it would lose very badly - probably only getting around 2.5 out of 10 at tournament time controls. That would also be consistent with my own estimates of low 1500s at best.

Then again, your testing is that faster time controls so I guess that would increase the rating over 40 in 2 somewhat.

I am interested to know which Excaliburs are still available new and are reasonably strong. I am also curious to know whether the very last of the Excaliburs had the bug resolved that plagued the Alexandra, Phantom Force and DTTC - that is they often lose on time at the 40th move at the 40 in 2 level. It is as if the program thinks it only needs to make 39 moves for the time control instead of the full 40. These machines also have a "dotted segment" LCD display (my terminology) that slowly corrupts after the first time control in a 40 in 2 game - the more moves made after move 40 the more corrupted the display becomes - sometimes to the point where you can no longer read the display and have to restart. This effected every one of the above models that I have come across, but I wonder if these cheaper models have the problem given they don't have the "dotted segment" displays in them.

I am hoping that Sir Dave might chime in with this thread as he owns the Pavilion model (or something like that) which I believe would also be a clone of the above two models.
A lot of good questions. If you read what I wrote for my tests, then I think you will have your first question answered. The goal is to have a total of 16 tests all deriving from human master games through the centuries. Since the players are human you will see some perfectly played games and some not so perfectly played when analyzed by a modern computer. After all they are humans.

The dedicated computers are not much different to this. There are certain game positions where a 2250 SSDF or 2400 Info Active rating such as the King or a Lang etc will play some games perfectly well above that rating and some games where it just does not have a clue. The same applies with the programmers of weaker computers. A lot of these weaker computers play on what you could almost call instinct because they do not reach anything beyond 2 or 3 ply at 30 seconds and maybe 4 ply if you are very lucky at 3 minutes per move.

Would it surprise you if I were to show you some test games were a Lang, King, etc performance in the game was as low as 1300 or even less? Hence the 16 test games that I plan to play. This will hopefully show all the good and the bad on all these tested computers and provide a final test rating across all these different scenarios.

BTW if it were possible tone down a Komodo, Hiarcs or Stockfish to play 3 or 4 or 5 ply only, you will soon realize how little improvement there is if any between what the programmers did in the early 80s and what the programs do today. The improvement lies on the higher speed and higher depth search improvement. At a minimalistic level these programs today have less knowledge than programs from the 80's.

The fun for me in individual test games such as for example Test Game 1 is to enjoy a weaker computer actually overall playing the game with more skill than a higher end computer, it is amusing and it teaches a lot about the different behaviors.

Regarding the rating, my instincts tell me that the rating ultimately most closely resembles the USCF ratings and surprisingly come a lot closer to what the manufacturers claimed for the computers than SSDF or Info.

But at the same time it is just a rating and best compared as results within the tests themselves and not against other rating systems, which after all are just guidelines as well.

Through the tests you will see the computers bounce around because of their knowledge and lack of knowledge in certain situations. Sometimes a program will do great and sometimes really bad. But this is exactly how it plays when you or I play them. Sometimes they surprise you and sometimes they disappoint you.

A further comment is that your performance against computers does not really acknowledge the strength of a computer because familiarity breeds contempt. Older computers once you have figured out how to beat them you will always continue to beat them because you learn how to respond to them and win. These computers will continue to faithfully play the same moves whereas you can adjust and learn from your losses or even make corrections if your mind tells you (oops I made a mistake here, not saying that you do that, but humans have this option) that you just made a mistake.

A more accurate test would be to place 10 1800 USCF players at a table and let them play in match conditions 10 1800 USCF Manufacturer rated computers and see what comes out of it. Now that would be a superb test!

A weakness with SSDF (which is why they have had to adjust their rating in the past) and other lists is that computers play against computers and invariably this pushes down weaker computers to a rating level which no longer makes sense, no longer resembles performances against humans or matches their true strength hence everyone who works with these tests a lot will tell you the same that the ratings should not be taken seriously and compared to other ratings lists. Only the performance of the contestants within the rating chart is what matters in the end.

Best regards
Nick
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

I am interested to know which Excaliburs are still available new and are reasonably strong. I am also curious to know whether the very last of the Excaliburs had the bug resolved that plagued the Alexandra, Phantom Force and DTTC - that is they often lose on time at the 40th move at the 40 in 2 level. It is as if the program thinks it only needs to make 39 moves for the time control instead of the full 40. These machines also have a "dotted segment" LCD display (my terminology) that slowly corrupts after the first time control in a 40 in 2 game - the more moves made after move 40 the more corrupted the display becomes - sometimes to the point where you can no longer read the display and have to restart. This effected every one of the above models that I have come across, but I wonder if these cheaper models have the problem given they don't have the "dotted segment" displays in them.
Sorry I didn't get a chance to respond to you this morning regarding the above.

The ones that I know that have the pixilation problem are all the LCD Chess models. Model 375 375-1 375-2 etc. as well as the original New York Times and Chess Station's Pocket Chess. I have not seen this problem repeated on other Excalibur computers yet.

Regarding Alexandra's problem with 40 in 2 hours I have never really noticed that. I can remember playing some 40 in 2 hours games with Alexandra a long time ago and it seems if I recall the computer allowed me to continue playing beyond 40 moves so I have never noticed this error. I will look for it next time.

Pavilion which is a Toys R Us brand has in fact sold Excalibur products under their Brand, but they were not really limited to Excalibur, Pavilion's Chess Man Pro II for example would have come from the same place as Lexibook.

Best regards,

Nick
Nick
SirDave
Full Member
Posts: 610
Joined: Thu Sep 09, 2010 4:59 am
Location: Southern California USA

Post by SirDave »

Jon: 'I am hoping that Sir Dave might chime in with this thread as he owns the Pavilion model (or something like that) which I believe would also be a clone of the above two models.'

I can't be of much help when it comes to tournament levels such as 40/2 because I don't really use them. I like to take my time and not be under any time constraints. I do know that the Pavilion absolutely is the Alexandra with a costume on though it may play a little weaker since it requires only 2 batteries. As to its Elo, I can't say for sure. I know I can reliably beat it at the average 30 sec. level and, fwiw, I can beat the GM at the average 15 sec. level. I have been picking up my game since 2009 after having not played for many years and I continue to progress so I'm not sure of my present Elo. At the moment, I think I'm still a little below Jon's 1700.
SirDave
Full Member
Posts: 610
Joined: Thu Sep 09, 2010 4:59 am
Location: Southern California USA

Post by SirDave »

spacious_mind wrote: Pavilion which is a Toys R Us brand has in fact sold Excalibur products under their Brand, but they were not really limited to Excalibur, Pavilion's Chess Man Pro II for example would have come from the same place as Lexibook.

Best regards,
Nick
Hi Nick,
The Pavilion unit is a bit of a crazy bird which I am very familiar with (having bought 3 of its variants). There was a pure Excalibur version which has 'Chess Champion' in big letters (with 'Talking Electronic Chess' under it in smaller letters) on the front of the box and 'Tournament Staunton' in big letters on the back. (What's with that? Was someone conflicted over what to call it?). The manufacturer/distributor is given as 'EbExcalibur, Inc. However, when you open up the box, the board has a plastic wrap with 'Pavilion' printed all over it and "Pavilion' and 'Talking Electronic Chess' printed on the board itself.

The Toys-R-Us version has a similar looking box, but with just 'Chess Champion' & 'Talking Electronic Chess' on it, but no 'Tournament Staunton' and 'Pavilion' is now on the front of the box. The manufacturer/distributor is given as 'Geoffrey LLC, a subsidiary of ToysRUs, Inc. Inside, everything is exactly the same as the Excalibur unit above EXCEPT, and this is a big EXCEPT, they have removed the magnetic weights from the pieces so the pieces have a tendency to slide all over the place. (I know they did this because of the hazard of children swallowing magnets, but this is just ridiculous!) I learned this the hard way when I bought the Toys-R-Us version to have as a backup to the Excalibur version. I ended up buying another Excalibur version off of eBay just to get the 'normal' pieces.
SirDave
Full Member
Posts: 610
Joined: Thu Sep 09, 2010 4:59 am
Location: Southern California USA

Post by SirDave »

Since some of this thread has been about Excalibur LCD units, I thought I'd just throw out a reminder of the following thread that may be of some interest to anyone who missed it and it does have a couple of interesting posts by Jon.

http://www.hiarcs.net/forums/viewtopic.php?t=4526
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

SirDave wrote:
spacious_mind wrote: Pavilion which is a Toys R Us brand has in fact sold Excalibur products under their Brand, but they were not really limited to Excalibur, Pavilion's Chess Man Pro II for example would have come from the same place as Lexibook.

Best regards,
Nick
Hi Nick,
The Pavilion unit is a bit of a crazy bird which I am very familiar with (having bought 3 of its variants). There was a pure Excalibur version which has 'Chess Champion' in big letters (with 'Talking Electronic Chess' under it in smaller letters) on the front of the box and 'Tournament Staunton' in big letters on the back. (What's with that? Was someone conflicted over what to call it?). The manufacturer/distributor is given as 'EbExcalibur, Inc. However, when you open up the box, the board has a plastic wrap with 'Pavilion' printed all over it and "Pavilion' and 'Talking Electronic Chess' printed on the board itself.

The Toys-R-Us version has a similar looking box, but with just 'Chess Champion' & 'Talking Electronic Chess' on it, but no 'Tournament Staunton' and 'Pavilion' is now on the front of the box. The manufacturer/distributor is given as 'Geoffrey LLC, a subsidiary of ToysRUs, Inc. Inside, everything is exactly the same as the Excalibur unit above EXCEPT, and this is a big EXCEPT, they have removed the magnetic weights from the pieces so the pieces have a tendency to slide all over the place. (I know they did this because of the hazard of children swallowing magnets, but this is just ridiculous!) I learned this the hard way when I bought the Toys-R-Us version to have as a backup to the Excalibur version. I ended up buying another Excalibur version off of eBay just to get the 'normal' pieces.
Hi Dave,

Yes I know the computer, I played a couple of games with it against Novag Constellation 3.6 MHz. It won a game and lost a game. It is related to Alexandra the Great.

Image

http://www.spacious-mind.com/html/talki ... chess.html

EB is the name that Excalibur changed to. All the Einstein's are EB's.

Best regards,
Nick
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

This comparison involves Excalibur E-Chess Model # 410 and RadioShack E-Chess Model # 60-2845.

EXCALIBUR E-CHESS MODEL # 410 - 2001

Image

http://www.spacious-mind.com/html/e-chess.html

RADIOSHACK E-CHESS MODEL # 60-2845 - 2001

Image

http://www.spacious-mind.com/html/e-chess1.html

E-Chess has 4 built-in play styles which I wanted to test. They are

1) FAST OFF + RANDOM OFF
2) FAST OFF + RANDOM ON
3) FAST ON + RANDOM OFF
4) FAST ON + RANDOM ON

TEST GAME 1 - 15TH CENTURY MASTER

Image

Both Excalibur and RadioShack E-Chess were 100% identical, therefore after playing them both through Test Game 1 I decided to lump them together.

The play level used is Level 23 taken from the RadioShack manual, which has a much better description of the levels than the Excalibur E-Chess manuals. Level 23 corresponds to 31 Seconds per move.

In Test Game 1 FAST ON seemed to perform best.

I have also included Excalibur Einstein LCD Chess Wizard and Excalibur Touch Chess II for comparison purposes. As you can see from the above chart, there is a lot of similarity between, but they are not exactly the same. This may be a result of different Hardware used, it is hard to tell.

TEST GAME 2 - THE TURK

Image

In Test Game 2, FAST OFF + RANDOM ON performed slightly better than the other settings.

The average rating after 2 games looks as follows:

Image

After two test games it looks as if the setting FAST ON + RANDOM OFF might be the best setting for E-Chess.

Best regards
Nick
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Here are some more comparisons. This time Excalibur Mirage, Igor, Ivan and Alexandra are compared.

Since Igor and Ivan have 12 MHz and Mirage 10 MHz, these tests were made were Igor and Ivan played Level 22 (25 seconds per move) and Mirage played level 23 which is 30 seconds per move. Alexandra was thrown in just to see if there is a relationship with Ivan, Igor or Mirage. Also to compare the difference between Level 22 (25 seconds per move) and Level 23 (30 seconds per move) I have also played Test game 1 with Level 23 just to compare.

TEST GAME 1 - 15TH CENTURY MASTER

Image

Both Level 22 and Level 23 are fixed time levels therefore different to the tests and ratings that are being done at:

http://hiarcs.net/forums/viewtopic.php?t=6835&start=15

If you look at the first test you will quickly see that Ivan is different to Mirage and Igor. Mirage (10 MHz) at level 23 and Igor (12 MHz) at level 23 play exactly the same moves! Therefore this seems to confirm that Igor is a Mirage clone with regards to software but is faster at 12 MHz.

Alexandra does not seem to have much in common with Mirage, Igor or Ivan.

TEST GAME 2 - THE TURK

Image

Test game reconfirms everything seen in Test game 1. Mirage and Igor varied moves only 1 time on White's move #8.

Alexandra played terribly in this game. So bad in fact that I repeated the game a second time to make sure that I did not have a bug or something in the first attempt. The second attempt ended up being even worse than the first attempt with Alexandra playing a different move on two occasions. This does seem to indicate that Alexandra has some built in randomness.

Based on these tests I would say that Ivan/Mirage/Igor are a different program to Alexandra.

Excalibur Avenger still remains a mystery, does it play at 12 MHz or 10 MHz? Doing this test at Level 23 (10 MHz) and Level 22 (12 MHz) might bring some more light.

Best regards,
Nick
User avatar
mclane
Senior Member
Posts: 1600
Joined: Sun Jul 29, 2007 9:04 am
Location: Luenen, germany, US of europe
Contact:

Post by mclane »

Avenger is the horvath program.
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

mclane wrote:Avenger is the horvath program.
Read CCS Apr/May 1996 and 1997 Nuernbuerger Messe.
Nick
User avatar
mclane
Senior Member
Posts: 1600
Joined: Sun Jul 29, 2007 9:04 am
Location: Luenen, germany, US of europe
Contact:

Post by mclane »

spacious_mind wrote:
mclane wrote:Avenger is the horvath program.
Read CCS Apr/May 1996 and 1997 Nuernbuerger Messe.
It plays like Horvath. Not different then my other Horvath machines such as legend, Regency, ...

Only the Avenger IS only Battery. While the others can be used by wire.
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....
masterclass
Member
Posts: 9
Joined: Wed May 11, 2016 11:01 pm

Post by masterclass »

have just acquired an Excalibur Igor model 711E with no manual. Any idea where i can get access to one - need it for the 100 playing levels
IanO
Member
Posts: 162
Joined: Sat Aug 15, 2009 11:43 pm
Location: Portland, OR

Post by IanO »

masterclass wrote:have just acquired an Excalibur Igor model 711E with no manual. Any idea where i can get access to one - need it for the 100 playing levels
Here are the levels

1: adaptive time

2-6: Fixed depth, 5 ply down to 1 ply (so level 6 is weakest available)

7-36: fixed time
37-66: average time (both are seconds per move)
1,2,3,4,5,6,7,8,9,10,
12,14,16,18,20,
25,30,35,40,45,50,
60, 75, 90, 105, 120, 135, 150, 165, 180

67-89: countdown (G/n minutes)
2,3,4,5,6,7,8,9,10,
15,20,25,30,35,40,45,50,55,60,
75,90,105,120

tournament levels:
90: 60/60
91: 40/60
92: 40/120, then 20/60
93: 20/60

94: infinite

95-100: problem solver: mate in 1 to 6

Note: Igor always thinks on the opponents time. There is no "easy" option to turn it off.
Post Reply