Portables Collector: Openings Diversity of 4 Novag Portables

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

Portables Collector: Openings Diversity of 4 Novag Portables

Post by SirDave »

OPENINGS DIVERSITY OF NOVAG JADE, JADE II, AMBER AND STAR OPAL

This is a study of the ‘openings diversity’ of four Novag portables in both Normal mode and Random mode (except the Star Opal): The original Jade (opening book 12,600 elo 1710) , the Jade II (opening book 13,000 elo 1900), the Amber (opening book 8,900 elo 1900) and the Star Opal (opening book 8,500 elo 1500). Novag particularly excelled at manufacturing portable dedicated chess computers and these four have been some of the most popular.

The Star Opal is out of its league in overall power in this group and it doesn’t have a Random key, but it is included because it is the only portable Novag unit still being sold at online stores and also, because it has (anecdotally) appeared to have a nice selection of openings. In addition, it was interesting to see how it would perform given its lower overall power.

The purpose of this study was to determine:
-The overall ‘openings diversity’ in response to E2E4 and D2D4.
-Whether the Random key affects the diversity of responses and if so, how much.
-If the overall number of openings that each unit has relates to the diversity of the openings. One might think that the more openings that were stored in the unit, the more possible responses to E2E4 and D2D4 would be offered.

This study was inspired by a thread that discussed the lack of ‘openings diversity’ of the Novag Obsidian:
http://www.hiarcs.net/forums/viewtopic. ... a1aa21d962

In that thread, the Obsidian was said to have a very restricted response to both E2E4 and D2D4. One misunderstanding in that thread:. It was said that the Novag manual states that the Random key does not affect the openings. In the manuals of the Jade, Jade II, Amber and Obsidian, the following (or a very similar) statements are made regarding the Random key:

Every time you switch the computer on, automatically the move with the highest rating is chosen, based on the depth of search which is determined by the skill level. This is always the case after the opening is finished.
However, this may not always provide the desired variety. If you press the Random key, you can choose various levels of randomness. The computer will select a move at random from a list of possible moves which are similar in their evaluation.


The statement, ‘However, this may not always provide the desired variety.’ refers to both statements that precede it with the apparent meaning that the Random key should affect both opening and the moves that follow. Of course, this could have been stated more clearly in the manual and it is easy to see why it was misunderstood.

Each unit was tested with 100 entries each of E2E4 and D2D4 in both Normal and maximum Random mode (ie. Random 3 out of a choice of 3) which gives a pure percent response. Any opening response that appears less than 1 in 100 would be of little practical use.

Note: It is difficult to format tables in forum posts hence the formatting used below. Nml means Normal mode, Rnd means Random mode. The figures, 02% and 05%, mean 2% and 5% and so on.
Last edited by SirDave on Tue Oct 23, 2012 1:28 am, edited 2 times in total.
SirDave
Full Member
Posts: 610
Joined: Thu Sep 09, 2010 4:59 am
Location: Southern California USA

Post by SirDave »

NOVAG OPENINGS DIVERSITY –Response to E2E4

NOVAG JADE
Response to E2E4:
Nml Rnd

54% 32% E7E5 (Open)
00% 00% E7E6 (French)
00% 00% A7A6
00% 00% B7B6
00% 02% B8C6 (Nimzo)
08% 23% C7C5 (Sicilian)
23% 21% C7C6 (Caro-Kann
10% 24% D7D5 (Scandinavian)
00% 00% D7D6 (Pirc)
00% 00% G7G5
05% 00% G7G6 (Modern)
00% 00% G8F6 (Alekhine)

Novag Jade E2E4 Openings Diversity Evaluation: The Random key obviously affects the openings, but not in a way that increases diversity. In fact, with Random on, the incidence of response of (only) four openings is simply reshuffled and the G7G6 (Modern Defense) is eliminated. The only discernible benefit of the Random key is a noticeable increase in the Sicilian Defense (C7C5). The total absence of the French (E7E6) and Pirc Defense (D7D6) is surprising. The Jade had the least overall diversity in response to E2E4 of all 4 units tested.


NOVAG JADE II
Response to E2E4:
Nml Rnd

59% 54% E7E5 (Open)
00% 02% E7E6 (French)
00% 00% A7A6
00% 02% B7B6
00% 00% B8C6 (Nimzo)
16% 16% C7C5 (Sicilian)
11% 15% C7C6 (Caro-Kann
08% 07% D7D5 (Scandinavian)
00% 03% D7D6 (Pirc)
00% 00% G7G5
00% 00% G7G6 (Modern)
06% 01% G8F6 (Alekhine)

Novag Jade II E2E4 Openings Diversity Evaluation: The Random key resulted in the addition of 4 ‘new’ responses (partly at the expense of the Alekhine (G8F6)), but the overall affect was not particularly significant. The absence of the French and Pirc Defense in non-Random/Normal mode and very limited presence in Random mode is surprising. Also interesting is that the Modern Defense (G7G6) was not present at all, while it showed up on a limited basis in both the original Jade and Amber. Still, the Random mode took the number of different responses from 5 to 8 and overall ‘openings diversity’ for E2E4 is second only to the Amber.


NOVAG AMBER
Response to E2E4:
Nml Rnd

50% 32% E7E5 (Open)
04% 06% E7E6 (French)
00% 07% A7A6
00% 01% B7B6
00% 02% B8C6 (Nimzo)
16% 19% C7C5 (Sicilian)
19% 09% C7C6 (Caro-Kann
05% 12% D7D5 (Scandinavian)
01% 05% D7D6 (Pirc)
00% 01% G7G5
01% 02% G7G6 (Modern)
04% 04% G8F6 (Alekhine)

Novag Amber E2E4 Openings Diversity Evaluation: The presence of 8 different responses in Normal mode and 12 in the Random mode with a reasonable presence in Random mode of the French, Pirc and Alekhine Defenses (not to mention the appearance of a couple of Nimzos) is impressive. The Novag Amber, by far, wins the award for most ‘openings diversity’ in response to E2E4.


NOVAG STAR OPAL
Response to E2E4:
Nml Rnd

50% E7E5 (Open)
00% E7E6 (French)
00% A7A6
00% B7B6
00% B8C6 (Nimzo)
22% C7C5 (Sicilian)
12% C7C6 (Caro-Kann
09% D7D5 (Scandinavian)
00% D7D6 (Pirc)
00% G7G5
00% G7G6 (Modern)
07% G8F6 (Alekhine)

Novag Star Opal E2E4 Openings Diversity Evaluation: The ‘openings diversity’ for E2E4 was less than expected, but at least the most important Defenses were represented since there is no Random mode. The presence of the Alekhine is appreciated given its surprising absence in the more powerful Jade.
Last edited by SirDave on Tue Oct 23, 2012 1:16 am, edited 1 time in total.
SirDave
Full Member
Posts: 610
Joined: Thu Sep 09, 2010 4:59 am
Location: Southern California USA

Post by SirDave »

NOVAG OPENINGS DIVERSITY –Response to D2D4:

NOVAG JADE
Response to D2D4:
Nml Rnd

30% 49% D7D5
51% 43% G8F6
16% 01% B7B5
00% 00% B7B6
00% 00% B8C6
03% 07% C7C5
00% 00% C7C6
00% 00% D7D6
00% 00% E7E5
00% 00% E7E6
00% 00% F7F5
00% 00% G7G6

Novag Jade D2D4 Openings Diversity Evaluation: Just as with responses to E2E4, diversity is very limited and the Random mode doesn’t seem to accomplish much. And why is the B7B5 response fairly frequent in Normal mode (when it isn’t even available at all in the other tested units in Normal mode) and then in Random mode it almost disappears? It doesn’t seem to make much sense. The original Jade wins the overall ‘least openings diversity award’ of the 4 units tested.


NOVAG JADE II
Response to D2D4:
Nml Rnd

46% 23% D7D5
44% 36% G8F6
00% 02% B7B5
00% 02% B7B6
00% 02% B8C6
03% 10% C7C5
00% 01% C7C6
02% 14% D7D6
00% 02% E7E5
02% 04% E7E6
00% 03% F7F5
03% 03% G7G6

Novag Jade II D2D4 Openings Diversity Evaluation: The ‘openings diversity’ in response to D2D4 was very good in Normal mode and excellent in Random mode providing 12 different responses. It was noticeably the best of the 4 units in response to D2D4.


NOVAG AMBER
Response to D2D4:
Nml Rnd

51% 30% D7D5
41% 34% G8F6
00% 00% B7B5
00% 00% B7B6
00% 02% B8C6
03% 05% C7C5
00% 00% C7C6
02% 01% D7D6
00% 00% E7E5
01% 07% E7E6
00% 07% F7F5
02% 14% G7G6

Novag Amber D2D4 Openings Diversity Evaluation: Just as with E2E4, the Amber shows a reasonable diversity in Normal mode and a noticeable increase in diversity in Random mode, particularly with the increase in the Modern Defense (G7G6). The Amber, as with E2E4 has very good ‘openings diversity’ in Normal mode and has a very good increase in diversity in Random mode, though this time, bested by the Jade II in Random mode.


NOVAG STAR OPAL
Response to D2D4:
Nml

53% D7D5
24% G8F6
00% B7B5
00% B7B6
00% B8C6
09% C7C5
00% C7C6
07% D7D6
00% E7E5
00% E7E6
00% F7F5
07% G7G6

Novag Star Opal D2D4 Openings Diversity Evaluation: The ‘openings diversity’ in response to D2D4 is almost comparable to the Jade II and Amber and better than the original Jade when in Normal mode. The frequency of the Modern Defense (G7G6) is welcome considering that the only other unit that has it in useful numbers is the Amber in Random mode.

Overall Conclusions:
The Novag Amber and Jade II both had exceptional ‘openings diversity’ in Normal mode and particularly in Random mode with the Amber superior in response to E2E4 and the Jade II superior in response to D2D4. The Novag Jade was particularly disappointing in its relative lack of diversity in both Normal and Random modes in response to E2E4 and D2D4. The Star Opal not only held its own up against the Jade in Normal mode, but might be said to be slightly superior to it in diversity though the Star Opal didn’t come close to the Amber and Jade II.

I would give the overall ‘openings diversity’ award to the Amber, slightly edging the Jade II because of the Amber’s more diverse responses to E2E4 which is still (perhaps arguably depending on one’s skill level) the overall most popular opening.

The Random key had a definite affect on the ‘openings diversity’ of all 3 units that have a Random key, but when it comes the original Jade, the Random key seemed to serve little practical purpose. The number of openings stored in a unit apparently doesn't directly affect the diversity of the openings. The Amber with 8,900 openings was at least the equal of the Jade II with 13,000 openings. Likewise, the Star Opal with 8,500 openings was at least the equal in Normal mode with the original Jade with 12,600 openings.

Incidentally, a small test was also conducted using the Next Best key available with the Jade, Jade II and Amber. It appears that when it comes to openings, all the Next Best keys does is act as a toggle: If E2E4 is entered, the Next Best key responds with either E7E5 or C7C5. This was true even in the case of the Jade which is interesting since in Normal mode it by far responds with C7C6 (Caro-Kann) over C7C5 as the second most frequent response to E2E4. In the case of D2D4, the Next Best key toggles between D7D5 and G8F6. In short, the Next Best key allows only two different responses to E2E4 and D2D4 and does not choose from a diverse menu of openings.

To anticipate a possible question: The ‘openings diversity’ of these units beyond the initial response to E2E4 and D2D4 was not studied. It would be interesting to see what diversity exists when it comes to all the possible responses to openings such as the Open Game and the Sicilian Defense.
User avatar
Steve B
Site Admin
Posts: 10140
Joined: Sun Jul 29, 2007 10:02 am
Location: New York City USofA
Contact:

Post by Steve B »

Excellent Scholarly Research Work Dave
will take some time to digest

what was your testing methodology?
you played games or just reset the computer after it made a reply to the opening move?
if only one move...did you do a full reset after each response or only one reset at the beginning and then just started a NEW Game after each response?



This belongs as a Wiki Entry Regards
Steve
Larry
Senior Member
Posts: 2269
Joined: Wed Aug 01, 2007 8:42 am
Location: Gosford, NSW Australia

Post by Larry »

Steve B wrote:Excellent Scholarly Research Work Dave
will take some time to digest


Steve
Yep, never seen such in depth testing on chess comp openings. I'm
tempted to get out the Novag Emerald Classic Plus and do a check to
see if it agrees with the Amber. They are supposed to be the same
software.
I don't own a Jade or Jade2, keeping a look out for them. A few weeks
back I found an Opal at a flea market for $5. Not much playing strength
but a neat little critter.
Good work Dave,
Larry
SirDave
Full Member
Posts: 610
Joined: Thu Sep 09, 2010 4:59 am
Location: Southern California USA

Post by SirDave »

Steve B wrote:Excellent Scholarly Research Work Dave
will take some time to digest

what was your testing methodology?
you played games or just reset the computer after it made a reply to the opening move?
if only one move...did you do a full reset after each response or only one reset at the beginning and then just started a NEW Game after each response?

This belongs as a Wiki Entry Regards
Steve
Thanks Steve and Larry- it turned out to be more of an undertaking than I had thought.

That's a good question Steve. Your last statement is what I did: Turned the unit on once for each separate study ie. E2E4 normal, E2E4 random etc and then simply entered a New Game after each response so the results reflects more what happens after the unit has been turned on once.

It's hard to say if the results would have been different if I had turned the units off & on for each New Game. I did a short test to see if that would be true and it didn't appear to be, but the short test may not have been significant enough.

One thing I know from my programming days which probably would have included when the firmware for these units was programmed- randomizing is more complicated than it would seem and has to be done correctly or else random is not really random. Part of that process usually meant setting a 'seed' before starting the randomizing process. What if the seed was not reset properly unless the unit was turned off?

In reality, I wasn't prepared to turn these things off & on 200 times each- too much wear & tear on the electronics (and myself, given how tedious the process was as it was :)). However, the results, particularly for the Amber and Jade II indicated a very useful randomizing process going on.

One errata in the results: The results for the Jade show 2 B8C6 (Nimzo) responses in Random mode. That is an error; there were zero B8C6 responses. Nimzo (sort of surprisingly) showed up only in the Amber in Random mode.
RadioSmall
Member
Posts: 273
Joined: Fri Jul 02, 2010 7:12 pm

Post by RadioSmall »

SirDave wrote:
Steve B wrote:Excellent Scholarly Research Work Dave
will take some time to digest

what was your testing methodology?
you played games or just reset the computer after it made a reply to the opening move?
if only one move...did you do a full reset after each response or only one reset at the beginning and then just started a NEW Game after each response?

This belongs as a Wiki Entry Regards
Steve
Thanks Steve and Larry- it turned out to be more of an undertaking than I had thought.

That's a good question Steve. Your last statement is what I did: Turned the unit on once for each separate study ie. E2E4 normal, E2E4 random etc and then simply entered a New Game after each response so the results reflects more what happens after the unit has been turned on once.

It's hard to say if the results would have been different if I had turned the units off & on for each New Game. I did a short test to see if that would be true and it didn't appear to be, but the short test may not have been significant enough.

One thing I know from my programming days which probably would have included when the firmware for these units was programmed- randomizing is more complicated than it would seem and has to be done correctly or else random is not really random. Part of that process usually meant setting a 'seed' before starting the randomizing process. What if the seed was not reset properly unless the unit was turned off?

In reality, I wasn't prepared to turn these things off & on 200 times each- too much wear & tear on the electronics (and myself, given how tedious the process was as it was :)). However, the results, particularly for the Amber and Jade II indicated a very useful randomizing process going on.

One errata in the results: The results for the Jade show 2 B8C6 (Nimzo) responses in Random mode. That is an error; there were zero B8C6 responses. Nimzo (sort of surprisingly) showed up only in the Amber in Random mode.
The Novag Sapphire is also worth mentioning. It has one of the most varied opening responses I have ever seen for a dedicated Computer or otherwise.Without the need the push the Random Key as White it plays 1.e4 , 1.d4 , 1.Nf3 , 1.c4 , 1.b4 , and 1.g4. As black in response to 1.e4 and 1.d4 it plays almost everything imaginable under the sun, too numerous to mention here. I know this from over one hundred games I have played throughout the years pitting it against both other dedicated machines and full PC programs. It would be interesting to know why this design Philosophy did not carry over to the Obsidian.
SirDave
Full Member
Posts: 610
Joined: Thu Sep 09, 2010 4:59 am
Location: Southern California USA

Post by SirDave »

RadioSmall wrote: The Novag Sapphire is also worth mentioning. It has one of the most varied opening responses I have ever seen for a dedicated Computer or otherwise.Without the need the push the Random Key as White it plays 1.e4 , 1.d4 , 1.Nf3 , 1.c4 , 1.b4 , and 1.g4. As black in response to 1.e4 and 1.d4 it plays almost everything imaginable under the sun, too numerous to mention here. I know this from over one hundred games I have played throughout the years pitting it against both other dedicated machines and full PC programs. It would be interesting to know why this design Philosophy did not carry over to the Obsidian.
It would be interesting to know how the Sapphire compares to the other portables above. I'm surprised that over all the years that chess boards have been manufactured and during which so much attention has been given to things such as processor speed and number of stored opening moves that there has been practically no mention of openings diversity.

It would seem to me that openings diversity would be one of the most important things to support one's learning the various openings and also to increase the interest in any particular game. Why would I want to keep seeing E5 in response to E4 when advanced or club players that I might play are responding with a lot of different 'surprise' openings? Seems to me that, as in the case of the Jade, it is a waste of a 12,600 book if only a few responses to E4 and D4 are offered.
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 »

Hi Dave,

Very thorough and interesting work! Thanks for posting this. I guess what the survey can't really show by its nature is what happens after the first few moves. This is where I think Novag excels in that they generally have very broad responses after the first few moves. The Star Opal in particular has always surprised me as regards the relative diversity of variations after the first few moves as opposed to the relatively restricted responses on the first and second move as your results show.

Novags have always had nice books and this has been the case ever since the orginal 2.0 Mhz Constellation - although it had a tiny book by the standards of later dedicated machines, there was a lot of interest and variety once the random button had been pressed.

I wish these machines, however would show more "consistent diversity" for want of a better expression. By that I mean that all responses to a given opening move that are considered "not unsound" (as opposed to strictly sound) are given equal probability. I'd like to see machines, for example (in random mode), open just as often with b3, g3, f4 and even b4 and Nc3 as often as d4 or e4. And as black I'd like to see responses to e4 such as b6, Nc6, d6 as well as the much more common c6, d5, e6 and Nf6 handled with equal consideration to the "big" major openings.

I know I keep saying it, but this is why I think the best openings book ever in a small hardware device is the one in the Nintendo Fritz - it really is superb in terms of not showing any particular biases and has huge variety even from the very first move.

Sorry if I missed you mentioning it, but doesn't the Amber have several levels of random? I seem to recall the button could be cylced several times and an extra "random bar" would appear in the display each time it was pressed.
Chess is like painting the Mona Lisa whilst walking through a minefield.
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 »

Larry wrote: I don't own a Jade or Jade2, keeping a look out for them.
If you can find one in good condition snap it up because these machine seem to be quite rare relatively speaking. I very rarely see them come up at auctions and they don't seem to often get mentioned either. I wonder whether these series of machines were less popular than other series and that may explain the relative rarity. When these machines were on the market, I suspect that many people were at the point of branching off into PC chess engines, since we are talking about the early / mid 90s here when PCs were strong enough to support a chess program playing at Master strength.

Whateve the case, I have always maintained that these Novags (Jade, Jade II) were the best looking and arguably the most practical portable chess computers ever invented.
Chess is like painting the Mona Lisa whilst walking through a minefield.
SirDave
Full Member
Posts: 610
Joined: Thu Sep 09, 2010 4:59 am
Location: Southern California USA

Post by SirDave »

Monsieur Plastique wrote:Hi Dave,

Very thorough and interesting work! Thanks for posting this. I guess what the survey can't really show by its nature is what happens after the first few moves. This is where I think Novag excels in that they generally have very broad responses after the first few moves. The Star Opal in particular has always surprised me as regards the relative diversity of variations after the first few moves as opposed to the relatively restricted responses on the first and second move as your results show.

Novags have always had nice books and this has been the case ever since the orginal 2.0 Mhz Constellation - although it had a tiny book by the standards of later dedicated machines, there was a lot of interest and variety once the random button had been pressed.
Hi Jon- glad you found this useful. When it comes to the Star Opal's presentation/diversity of variations, you would have a better sense of that than I. I agree that even a unit with a relatively small book can have a fair amount of diversity. And I don't understand why those with a relatively large book (eg. Jade) would have such a limited openings diversity.

Digressing a bit- it's amazing how squirrely the decisions are when it comes to the size of the opening book in some machines. I'm doing a little study of the Saitek/Mephisto clamshell portables. The Saitek Express 16K and Saitek Cavalier (much the same firmware) have an ELO of 1450 and an opening book of 5000, not too bad for these small, earlier portables and with an apparent fairly good openings diversity. Then a few years later, in almost exactly the same physical unit, came the more advanced (ELO 1870, H8 processor) Saitek Champion Advanced Trainer (Mephisto Champion is a sibling) but with an opening book of only 2000. What was the crazy-think behind that?
I wish these machines, however would show more "consistent diversity" for want of a better expression. By that I mean that all responses to a given opening move that are considered "not unsound" (as opposed to strictly sound) are given equal probability. I'd like to see machines, for example (in random mode), open just as often with b3, g3, f4 and even b4 and Nc3 as often as d4 or e4. And as black I'd like to see responses to e4 such as b6, Nc6, d6 as well as the much more common c6, d5, e6 and Nf6 handled with equal consideration to the "big" major openings.
I totally agree. Wouldn't one of the reasons for buying these things in the first place be the learning a variety of openings including the less common ones? Saitek seemed to do a better job in the units where you can select a Passive book, Active book and Entire book (ie. not necessarily the best opening, but a selection from the entire book).
Sorry if I missed you mentioning it, but doesn't the Amber have several levels of random? I seem to recall the button could be cylced several times and an extra "random bar" would appear in the display each time it was pressed.
Yes, at the beginning I mention that the test occurred 'in both Normal and maximum Random mode (ie. Random 3 out of a choice of 3)' The Jade, Jade II and Amber all had a choice of Random 1 to 3. I chose to use the maximum setting to see what was the 'best' diversity that a unit could come up with. Even at maximum, the Jade didn't come up with much. I did some minor testing of Random 1 and 2. Random 1 in particular didn't seem to affect the openings very much. Maybe the Random modes 1 to 3 have more useful differences when it comes to moves after the opening.
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 »

SirDave wrote:Then a few years later, in almost exactly the same physical unit, came the more advanced (ELO 1870, H8 processor) Saitek Champion Advanced Trainer (Mephisto Champion is a sibling) but with an opening book of only 2000. What was the crazy-think behind that?
I've never been quite certain what this 2000 move opening book really means. I have just pulled my mid-90s Saitek Chess Computer catalogue off the shelf (yes, from when you could still get a chess computer catalogue in physical form) and it has numerous 16K programs, all similar but with vastly differing opening book sizes - at least in terms of what is quoted.

This is likely the last catalogue before the Miami and GK2000 came on the market, since it only lists Turbo Advanced Trainer, Executive and Virtuoso as the 16K programs. It shows Turbo Advanced Trainer as having 6,000 moves, Executive as 2,000 moves and Virtuoso 6,000 moves. All have the 64 levels and the 1900 rating. The later Miami of course is also 1900 and the back of the box actually states that it has a 6,000 ply opening book.

But the question is whether that book really is 6,000 ply (and whether the Turbo Advanced and Virtuoso books are 6,000 ply), or whether they are actually similar 2,000 move books to that Champion Advanced Trainer, but the 6,000 came from a bit of creativity on the part of the marketing people, since there were different modes of opening book so I understand (well I know that is the case for Miami, but I don't own the other three). I have always considered that Miami is a 2,000 book machine but the book comes in three flavours. And somehow someone multiplied 2000 by three flavours and decided to quote 6,000.

That is my belief anyway - pretty hard to know without reverse engineering the code. But I find it really hard to believe that a 16K program with 2,000 moves and a 16K program with 6,000 moves can have pretty much the same rating by the same manufacturer at the same point in time unless the books are both actually the same size - there is just too much to lose in terms of chess knowledge and playing strength when you change the size of the library in the small 16K models.

I guess whatever the truth is, there is no doubting that as far as 16K program opening bookss are concerned, the Novag Star Opal leaves all rivals in it's wake.
Chess is like painting the Mona Lisa whilst walking through a minefield.
SirDave
Full Member
Posts: 610
Joined: Thu Sep 09, 2010 4:59 am
Location: Southern California USA

Post by SirDave »

Monsieur Plastique wrote:
SirDave wrote:Then a few years later, in almost exactly the same physical unit, came the more advanced (ELO 1870, H8 processor) Saitek Champion Advanced Trainer (Mephisto Champion is a sibling) but with an opening book of only 2000. What was the crazy-think behind that?
I've never been quite certain what this 2000 move opening book really means. I have just pulled my mid-90s Saitek Chess Computer catalogue off the shelf (yes, from when you could still get a chess computer catalogue in physical form) and it has numerous 16K programs, all similar but with vastly differing opening book sizes - at least in terms of what is quoted.

This is likely the last catalogue before the Miami and GK2000 came on the market, since it only lists Turbo Advanced Trainer, Executive and Virtuoso as the 16K programs. It shows Turbo Advanced Trainer as having 6,000 moves, Executive as 2,000 moves and Virtuoso 6,000 moves. All have the 64 levels and the 1900 rating. The later Miami of course is also 1900 and the back of the box actually states that it has a 6,000 ply opening book.

But the question is whether that book really is 6,000 ply (and whether the Turbo Advanced and Virtuoso books are 6,000 ply), or whether they are actually similar 2,000 move books to that Champion Advanced Trainer, but the 6,000 came from a bit of creativity on the part of the marketing people, since there were different modes of opening book so I understand (well I know that is the case for Miami, but I don't own the other three). I have always considered that Miami is a 2,000 book machine but the book comes in three flavours. And somehow someone multiplied 2000 by three flavours and decided to quote 6,000.

That is my belief anyway - pretty hard to know without reverse engineering the code. But I find it really hard to believe that a 16K program with 2,000 moves and a 16K program with 6,000 moves can have pretty much the same rating by the same manufacturer at the same point in time unless the books are both actually the same size - there is just too much to lose in terms of chess knowledge and playing strength when you change the size of the library in the small 16K models.

I guess whatever the truth is, there is no doubting that as far as 16K program opening bookss are concerned, the Novag Star Opal leaves all rivals in it's wake.
This is an area where my guess is that your chess programming experience would give you a more accurate perspective. Like you, I have sometimes wondered whether the Book figures are anywhere near reality. How could the average player know for sure? I've never seen any evidence of anyone trying to determine whether the numbers given are anywhere near accurate.

For instance, if the Jade with a Book of 12,600 only offers 4 or 5 alteratives in the normal or maximum random mode in response to E4 or D4, where are all those thousands of opening moves showing up? Are there likely to be many Variations offered after those few openings? Likewise, are we to believe that the limited 4 or 5 'main' openings leads to dozens of Transitions? I rather doubt it. On the other hand, the Amber with a listed Book of 8900 offers 12 different responses to E4 in maximum random mode which could lead in all sorts of directions. Makes you wonder whether the Jade has a grossly overly-stated Book.
RadioSmall
Member
Posts: 273
Joined: Fri Jul 02, 2010 7:12 pm

Post by RadioSmall »

SirDave wrote:
RadioSmall wrote: The Novag Sapphire is also worth mentioning. It has one of the most varied opening responses I have ever seen for a dedicated Computer or otherwise.Without the need the push the Random Key as White it plays 1.e4 , 1.d4 , 1.Nf3 , 1.c4 , 1.b4 , and 1.g4. As black in response to 1.e4 and 1.d4 it plays almost everything imaginable under the sun, too numerous to mention here. I know this from over one hundred games I have played throughout the years pitting it against both other dedicated machines and full PC programs. It would be interesting to know why this design Philosophy did not carry over to the Obsidian.
It would be interesting to know how the Sapphire compares to the other portables above. I'm surprised that over all the years that chess boards have been manufactured and during which so much attention has been given to things such as processor speed and number of stored opening moves that there has been practically no mention of openings diversity.

It would seem to me that openings diversity would be one of the most important things to support one's learning the various openings and also to increase the interest in any particular game. Why would I want to keep seeing E5 in response to E4 when advanced or club players that I might play are responding with a lot of different 'surprise' openings? Seems to me that, as in the case of the Jade, it is a waste of a 12,600 book if only a few responses to E4 and D4 are offered.
I am a big fan of the Novag Sapphire .I would say that the Sapphire compares very favorably to the above portables.It is the most diverse in terms of Openings and completely bug free in every respect .In the case of the Novag Obsidian as white it has a diverse opening , playing lots of stuff other than 1.e4 and 1.d4.The main problem with it is the with black , in response to 1.d4 it always plays 1...d5 and after 2.c4 it always plays 2....dxc4. One could get rid of this tendency by pushing the random key
, however there is a big catch.Although the random key does diversify the opening it also has a negative effect on middle game play, weakening it.With the Obsidian the Random option being on during middle game play creates the bizzare tendency of minor pieces going back to their original squares during game play.for example moves like Nc6-b8 or Bg4-c8 or Nd7-b8 will be quite common. It also plays 1....e5 95% of the time in response to 1.e4 playing 1...c5 the other 5%. With the RadioShack brands of dedicated chess Computers you cannot go wrong in terms of opening diversity.Thus I consider RadioShack to be superior to Novag in general.For example even the RadioShack 1650 which has a very small opening library is very diverse with both black and white.Also it seems that all RadioShack Chess Computers have a large random factor built into them by default since during middle game play and also during analysis one does not always get the same moves after equal thinking time, in the same positions. This is very important since sometimes a weak player can find an opening or formula to beat a stronger dedicated machine that does not have this randomness.
SirDave
Full Member
Posts: 610
Joined: Thu Sep 09, 2010 4:59 am
Location: Southern California USA

Post by SirDave »

RadioSmall wrote:
SirDave wrote:
RadioSmall wrote: The Novag Sapphire is also worth mentioning. It has one of the most varied opening responses I have ever seen for a dedicated Computer or otherwise.Without the need the push the Random Key as White it plays 1.e4 , 1.d4 , 1.Nf3 , 1.c4 , 1.b4 , and 1.g4. As black in response to 1.e4 and 1.d4 it plays almost everything imaginable under the sun, too numerous to mention here. I know this from over one hundred games I have played throughout the years pitting it against both other dedicated machines and full PC programs. It would be interesting to know why this design Philosophy did not carry over to the Obsidian.
It would be interesting to know how the Sapphire compares to the other portables above. I'm surprised that over all the years that chess boards have been manufactured and during which so much attention has been given to things such as processor speed and number of stored opening moves that there has been practically no mention of openings diversity.

It would seem to me that openings diversity would be one of the most important things to support one's learning the various openings and also to increase the interest in any particular game. Why would I want to keep seeing E5 in response to E4 when advanced or club players that I might play are responding with a lot of different 'surprise' openings? Seems to me that, as in the case of the Jade, it is a waste of a 12,600 book if only a few responses to E4 and D4 are offered.
I am a big fan of the Novag Sapphire .I would say that the Sapphire compares very favorably to the above portables.It is the most diverse in terms of Openings and completely bug free in every respect .In the case of the Novag Obsidian as white it has a diverse opening , playing lots of stuff other than 1.e4 and 1.d4.The main problem with it is the with black , in response to 1.d4 it always plays 1...d5 and after 2.c4 it always plays 2....dxc4. One could get rid of this tendency by pushing the random key
, however there is a big catch.Although the random key does diversify the opening it also has a negative effect on middle game play, weakening it.With the Obsidian the Random option being on during middle game play creates the bizzare tendency of minor pieces going back to their original squares during game play.for example moves like Nc6-b8 or Bg4-c8 or Nd7-b8 will be quite common. It also plays 1....e5 95% of the time in response to 1.e4 playing 1...c5 the other 5%. With the RadioShack brands of dedicated chess Computers you cannot go wrong in terms of opening diversity.Thus I consider RadioShack to be superior to Novag in general.For example even the RadioShack 1650 which has a very small opening library is very diverse with both black and white.Also it seems that all RadioShack Chess Computers have a large random factor built into them by default since during middle game play and also during analysis one does not always get the same moves after equal thinking time, in the same positions. This is very important since sometimes a weak player can find an opening or formula to beat a stronger dedicated machine that does not have this randomness.
The Sapphire (or Sapphire II) firmware in the Obsidian and the Amber firmware in the Star Opal would have resulted in two excellent dedicated boards- one a 'laptop' and the other a 'mini'.

Of course, almost everything attributed to RadioShack units would have originated with Saitek.
Post Reply