Oh, Atlanta

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.
Reinfeld
Member
Posts: 486
Joined: Thu Feb 17, 2011 3:54 am
Location: Tacoma, WA

Oh, Atlanta

Post by Reinfeld »

Just a sidebar to the rating-test thread.

Evidently, the Atlanta will run on batteries, per Wiki-Elo. (Whoops**)

If so, the Atlanta appears to answer two interesting questions:

1. What's the strongest dedicated machine that runs on batteries?**
2. What's the strongest Morsch machine available, *including* modules (excluding overclocking and other "bootleg" modifications)?

Once all of Nick's tests are complete, I'll be interested to see if Atlanta retains this position.

(I know not everyone will agree with this, but I tend to draw a distinction between modules and machines that play straight out of the box with one embedded engine. Just me.)

If point 1 and 2 above are true, the Atlanta is kick-ass strong, underappreciated and well worth owning. I'd be interested to hear reviews from those who are familiar with it. A match between Atlanta and Montreux or RISC 2500 would be very interesting; but I would suggest an RISC with the 2MB RAM upgrade.*

- R.

* Larry, he who possesses all three RISC 2500 models, has posted in the past about the different solving speeds associated with those upgrades. It's probably a hair-splitting distinction, but I'd be interested to know whether the 2MB version would outplay its brethren (and Montreux) in a series of games.

** Whoops - looks like I misread the Wiki-Elo translation. The Atlanta will run on a battery-powered adapter, but not exclusively on batteries, if I understand correctly. However, the bigger point (strongest Morsch machine) still appears to be solid.
"You have, let us say, a promising politician, a rising artist that you wish to destroy. Dagger or bomb are archaic and unreliable - but teach him, inoculate him with chess."
– H.G. Wells
User avatar
ricard60
Senior Member
Posts: 1285
Joined: Thu Aug 09, 2007 2:46 pm
Location: Puerto Ordaz

Post by ricard60 »

If Montreux runs with batteries it will be a great match between both of them to see wich is the strongest one.


Atlanta or Montreux regards
Ricardo
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Re: Oh, Atlanta

Post by spacious_mind »

Reinfeld wrote: ** Whoops - looks like I misread the Wiki-Elo translation. The Atlanta will run on a battery-powered adapter, but not exclusively on batteries, if I understand correctly. However, the bigger point (strongest Morsch machine) still appears to be solid.
No Atlanta does play on 6 AA batteries.

Regards,
Nick
Nick
Larry
Senior Member
Posts: 2269
Joined: Wed Aug 01, 2007 8:42 am
Location: Gosford, NSW Australia

Post by Larry »

Guys, the Montreux/risc2500 units do not run on batteries, only adaptor.
The Atlanta runs on either batteries or adaptor.
Berlin Pro runs only on adaptor.
I think we can say with confidence that the Atlanta is the strongest
battery powered dedicated chess computer ever made.
Larry
PS. Modules can run on batteries via a battery powered Exclusive board,
but I'm assuming we are excluding them simply because they are modules.
Larry
Senior Member
Posts: 2269
Joined: Wed Aug 01, 2007 8:42 am
Location: Gosford, NSW Australia

Re: Oh, Atlanta

Post by Larry »

Reinfeld wrote:


* Larry, he who possesses all three RISC 2500 models, has posted in the past about the different solving speeds associated with those upgrades. It's probably a hair-splitting distinction, but I'd be interested to know whether the 2MB version would outplay its brethren (and Montreux) in a series of games.
This of course assumes the Montreux to be the standard 128kb unit,
which as we know is the only RAM setting available in the released
Montreux. I personally have not had the curiosity to play risc2500 and
Montreux against each other, since they are virtually the same machine.
I have never played 2mb risc against 128kb risc. This is because I only
have the patience for fast games, where the ram setting probably would
not count.
The risc2500 came in at least three versions, which were no doubt
different only in a few very minor adjustments, resulting in a negligible
strength difference. We can look on the Montreux as the next version
of the risc2500 in a different color casing. The Englishman Ken Gorman
told me in an email once that the Montreux was released to rectify an
inbuilt problem in the risc2500 which caused it to freeze occasionally.
The Montreux New York Edition, released in honour of a tournament
in New York, should have included a much higher clock speed and 2mb
ram. Had this been the case, it would by now be a very highly sought
after unit selling for big bucks at auction. Hypothetical, I know...
L
Reinfeld
Member
Posts: 486
Joined: Thu Feb 17, 2011 3:54 am
Location: Tacoma, WA

Post by Reinfeld »

Larry said:
I think we can say with confidence that the Atlanta is the strongest
battery powered dedicated chess computer ever made.
Larry
So let me ask the other big question again: Is Atlanta the strongest of all Morsch machines?

Separately, regarding RISC 2500 and its multiple versions...

Larry said this:
I personally have not had the curiosity to play risc2500 and
Montreux against each other, since they are virtually the same machine.
I have never played 2mb risc against 128kb risc. This is because I only
have the patience for fast games, where the ram setting probably would
not count.
By "curiosity," I suspect you mean "patience." :D Totally understandable.

However, here's a couple of reasons why it might be worth it:

1. Montreux has a far larger opening book (380,000) than RISC (100,000).

Testing the 128K models against each other might show whether the book is any kind of factor. Testing the Montreux against the 2 MB RISC might show whether the bigger RAM chip neutralizes the opening-book factor. Then again, the differences might be so negligible that it's not worth measurement. Still, I wonder.

2. I suspect (but do not know) that Wiki-Elo's ratings are based on the RISC 128K model. Is the 2 MB version *stronger*? If so, by how much?

3. If it isn't obvious already, I'm aiming at the Plastic World Championship, a matter that should consume us all. Let's forget the wooden divas and their ceaseless demands for attention, dusting and coddling. Only four machines matter in this argument:

- Berlin Pro
- Montreux
- RISC 2500
- Atlanta

I say this with all due respect to the tragic Star Diamond, the Alexei Shirov of tabletops. Love that machine - but it ranks just below these monsters. Fidelity, sweet old Fidelity, brings up the rear

Same goes for the Berlin Pro London, it seems - strong, but not quite as strong as its father, if Wiki-Elo can be trusted.

And here's where the RISC versions come in. You'd have to start with a 2 MB RISC v Montreux and go from there. Which machine gets to participate?

The ELO gap for the four machines above is about 30-40 points. These are small distinctions.

Wiki-Elo ranking (Aktiv) looks like this:

1. Berlin Pro - 2284
2. Atlanta - 2267
3. Montreux - 2265
4. RISC 2500 (128 KB) - 2241

My personal ranking (an average of averages, including SSDF and Selective, as well as Wiki) looks like this:

1. Berlin Pro - 2242
2. RISC 2500 (128KB) - 2234
3. Montreux - 2231
4. Atlanta - 2214

- R.
"You have, let us say, a promising politician, a rising artist that you wish to destroy. Dagger or bomb are archaic and unreliable - but teach him, inoculate him with chess."
– H.G. Wells
Larry
Senior Member
Posts: 2269
Joined: Wed Aug 01, 2007 8:42 am
Location: Gosford, NSW Australia

Post by Larry »

Reinfeld wrote:
So let me ask the other big question again: Is Atlanta the strongest of all Morsch machines?

Separately, regarding RISC 2500 and its multiple versions...





1. Montreux has a far larger opening book (380,000) than RISC (100,000).

Testing the 128K models against each other might show whether the book is any kind of factor. Testing the Montreux against the 2 MB RISC might show whether the bigger RAM chip neutralizes the opening-book factor. Then again, the differences might be so negligible that it's not worth measurement. Still, I wonder.

2. I suspect (but do not know) that Wiki-Elo's ratings are based on the RISC 128K model. Is the 2 MB version *stronger*? If so, by how much?







And here's where the RISC versions come in. You'd have to start with a 2 MB RISC v Montreux and go from there. Which machine gets to participate?

Yes, Atlanta is clearly the strongest of all Morsch machines.
Regarding the size of Montreux's opening book, we had this out on
the forum maybe a year ago. We could not find any official documentation
anywhere to confirm this. I remember someone producing a document
from the mid 1990's which mentioned an opening book way smaller
than 300,000 half moves. Here again we are only going on the Wiki
claim, which is someone's guesstimate that was passed around until
it was accepted as fact. The huge opening book assumption is linked
to Montreux's 256kb rom, compared to risc2500's 128kb rom. A very
knowledgeable guy told us that a good part of Montreux's
256kb rom is empty.
Referring to: http://www.schachcomputer.info/html/ssdf.html
we can see that the 128kb ram risc2500 is only 8elo below the upgraded
Montreux 512kb ram. Be assured that there is a sharply diminishing
return on ram upgrade size, ie, a 512kb ram upgrade is darned near
as beneficial as 2mg upgrade. I have heard from more than one source
that the estimated elo gain from 128kb to 2mb is about 40elo. It does
depend heavily on the time controls. At very fast games, there is
likely hardly any difference, at 'Aktiv' there will be some difference, and
at 40 moves in 90 minutes, eg, the difference will be the full 40elo.
We have had over the years a number of comments about the
risc2500/montreux saga. One guy called the integrity of SSDF ratings
into question, remembering that the initial Montreux rating was below
the earlier risc2500, so they (SSDF), resorted to giving Montreux weaker
opponents to jack it's rating up, and thereby help it's sales along.
It is an ongoing mystery to me as well, because almost nowhere do they
seem to want to specify the ram setting of the risc2500. Also, a LOT of
Montreux's are upgraded, and by some definition "modified".
Playing games between my 128kb risc and my Montreux presents me
with a little problem. My Montreux has 2mb ram. My other two risc2500's
are both with upgraded ram. This means that I would have to take the
backs off two machines, swap ram chips around, and reset jumpers,
play the games, then do everything in reverse. I'm not sure I'm willing
to do that at this time.
I'm sure we will get to the bottom of the matter one way or another.
L
User avatar
chessguru
Member
Posts: 47
Joined: Mon Feb 21, 2011 7:33 am
Contact:

Post by chessguru »

Hello,
Larry wrote:A very
knowledgeable guy told us that a good part of Montreux's
256kb rom is empty.
Yes, 40% are empty. The Montreux opening book is 33% larger.

But 380.000 halfmoves? Mephisto marketing? I don't know.

Following openings can be found in the programs:

Code: Select all

Mephisto Montreux Openings		Saitek Risc 2500 Openings (V1.03/1.04)
		
Sicilian Defence			       Spanish Opening
Najdorf Variation			      Zaitsev Variation
Polugayevski Variation			 Breyer Variation
Poisened pawn Variation			Marshall Gambit
Fischer Variation			      Open Defence
Byrne Variation				     Berlin Variation
Zagreb Variation			       Dilworth Variation
Scheveningen Variation			 Trifunovic Variation
Richter-Rauzer Attack			  Riga Variation
Sozin Attack				        Walbrodt Variation
Boleslavski Variation			  Centre Attack
Dragon Variation			       Anderssen Attack
Moscow Variation			       English Attack
Taimanov Variation			     Steinitz Defence Deferred
Accelerated Dragon			     Exchange Variation
Rossolimo Variation			    Steinitz Defence
Pin Variation				       Cordel Variation
Paulsen Variation			      Konikowski Variation
Kelly Variation				     Berlin Defence
Nimzowitsch Variation			  Schliemann Gambit
Wing Gambit				         Bird Defence
Morra Gambit				        Italian Game
Closed Variation			       Mueller Attack
Alapin Variation			       Giuoco Piano
Keres Variation				     Evans Gambit
Grand Prix Attacks			     Hungarian Defence
Two Knights Defence			    Two knights Defence
Canal Variation				     Four Knights Opening
Max Lange Attack			       Belgrade Gambit
Classical Variation			    Leipzig Gambit
Traxler Variation			      Scotch Game
Italian Game				        Ponziani Opening
Evans Gambit				        Petrov Defence
Hungarian Defence			      Steinitz Variation
Four Knights Opening			   Lasker Variation
Italian Variation			      Nimzowitsch Variation
€Rubinstein Variation			  Cochrane Gambit
Scotch Game				         Paulsen Variation
Classical Variation			    Philidor Defence
Mieses Variation			       Latvian Gambit
Scotch Four Knights Game		  Damiano Defence
Goering Gambit			         Queen's Pawn Counter-Gambit
Scotch Gambit			          King's Gambit
Spanish Opening			        Vienna Game
Schliemann Gambit			      Centre Game
Cordel Defence			         Bishop's Opening
Berlin Defence				      Sicilian Defence
Cozio Defence				       Alapin Variation
Marshall Gambit				     Rossolimo Variation
Open Defence				        Accelerated Dragon
Dilworth Variation			     Maroczy Variation
Riga Variation				      Dragon Variation
Centre Attack				       Scheveningen Variation
Exchange Variation			     Najdorf Variation
Bird Defence				        Nimzowitsch Variation
Ponziani Opening			       O'Kelly Variation
Latvian Gambit				      Closed Variation
Petrov Defence				      Grand Prix Attack
Steinitz Variation			     Wing Gambit
Four Knights Opening			   Keres Variation
Three Knights Opening		  	Morra Gambit
Philidor Defence			       Caro-Kann
King's Gambit				       Nimzowitsch Variation
Kieseritsky Gambit			     Capablanca Variation
Berlin Defence				      Knight Variation
Hungarian Defence			      Fianchetto System
Brentano Defence			       Panov Variation
Paulsen Defence				     Advance Variation
Long Whip				           Two Knights Variation
Allgaier Gambit				     Pirc Defence
Hanstein Gambit				     Austrian Attack
Hanstein Gambit				     Classical Variation
Philidor Gambit				     Kholmov Variation
Polerio (Muzio) McLean Attack	Byrne Variation
Lolli Gambit				        Czech System
Herzfeld Gambit				     Centre Counter Game
Ghulam Kassim Gambit			   French Defence
Rosentreter Gambit			     Winawer Variation
Euwe Defence				        Alekhine Gambit
Modern Defence				      Classical Defence
Fischer Defence				     Steinitz Variation
Becker Defence				      Burn Variation
Schallopp Defence			      MacCutcheon Variation
Bishop's Gambit				     Rubinstein Variation
Classical Defence			      Tarrasch Variation
Mason Gambit				        Alekhine Defence
Petrov Gambit				       King's Fianchetto
Breyer Gambit				       Queen's Fianchetto
Falkbeer Countergambit			 Nimzowitsch Defence
Nimzowitsch C.gambit Declined   St. George Defence
Vienna Game				         Slav Defence
Bishop's Opening			       Dutch Variation
Centre Game				         Carlsbad Variation
French Defence				      Wiesbaden Variation
Winawer Variation			      Steiner Variation
Petrosian Defence			      Geller gambit
Alekhine Gambit				     Meraner Variation
Classical Defence			      Schlechter Variation
Steinitz Variation			     Exchange Variation
MacCutcheon Variation		     Breyer Variation
Classical Variation			    Winawer Counter-Gambit
Burn Variation				      Exchange Variation
Rubinstein Variation			   Queen's Gambit Accepted
Advanced Variation�			     Vienna Variation
Tarrasch Variation			     Queen's Gambit
Exchange Variation			     Tartakower Variation
Caro-Kann				           Ragozin Variation
Panov Variation				     Exchange Variation
Alekhine Defence			       Tarrasch Defence
King's Fianchetto			      Chigorin Defence
Austrian Attack				     Albin Counter-Gambit
Centre Counter Game			    Marshall Variation
Czech System				        London System
Pirc Defence				        Blackmar-Diemer Gambit
Nimzowitsch Defence			    Nimzo-Indian Defence
Queen's Fianchetto			     Classical Variation
Nimzo-Indian Defence			   Rubinstein Variation
Classical Variation			    Leningrad Variation
Rubinstein Variation			   Modern Benon
Leningrad Variation			    Taimanov Variation
Queen's Indian Defence		    Queen's Indian Defence
Petrosian Variation			    Bogo-Indian Defence
Bogo-Indian Defence			    Catalan Opening
Catalan Opening				     Budapest Gambit
Budapest Gambit			        Grue�nfeld Defence
Wolga Gambit				        Exchange Variation
Benoni Wall				         Sevilla Variation
King's Indian Defence			  King's Indian Defence
Grue�nfeld Defence			      Averbakh Variation
London System				       Glicoric System
Torre Attack				        Exchange Variation
Trompowski Attack			      Old Indian Defence
Queen's Gambit				      Old Indian Defence
Tarrasch Defence			       Wolga Gambit
Schara-Hennig Gambit			   Benoni Wall
Semi-Slav				           The Vulture
Chigorin Defence			       Torre Attack
Slav Defence				        Trompowsky Attack
Winawer Countergambit			  Dutch Defence
Exchange Variation			     Staunton Gambit
Queen's Gambit Accepted			Polisch Defence
Alekhine Variation			     Barcza Opening
Albin Countergambit			    Reti Opening
Austrian Defence			       Lisitsin Gambit
Blackmar-Diemer Gambit			 English Opening
Dutch Defence				       Mikenas Variation
Staunton Gambit				     Nimzowitsch-Larsen Attack
Charlick Gambit				     Orang Utan Opening
English Opening				     Heinrichsen Opening
Opening					          Grob Opening
Barcza Opening				      Bird System
Lisitsin Gambit				     Saragossa Opening
Bird System		
From Gambit		
Nimzowitsch-Larsen Attack		
Grob Opening		
Orang Utan Opening		
Heinrichsen Opening		
King's Fianchetto Opening		
Van 't Kruijs Opening		

I have heard from more than one source
that the estimated elo gain from 128kb to 2mb is about 40elo. It does
depend heavily on the time controls.


A look at the CSS:

Image

and

Image


Risc 2500 BT2630 Test / 2 MB and 512 KB

Image

at 'Aktiv' there will be some difference
No difference (128K <-> 2MB) at "aktiv".
at 40 moves in 90 minutes, eg, the difference will be the full 40elo.


Yes, I would agree.


If you want games (Risc, Montreux...), no problem.

Best regards,
Micha
User avatar
chessguru
Member
Posts: 47
Joined: Mon Feb 21, 2011 7:33 am
Contact:

Post by chessguru »

Sorry, the diagrams:

Image
  • 1...Kb8 - remis
[fen]2k5/8/8/8/1PK5/8/8/8 b - - 0 1[/fen]
  • 1.Nb6 - white wins
[fen]k1N5/2p5/8/1P6/8/8/7K/8 w - - 0 1[/fen]
  • 1.Kg3 - remis
[fen]8/1p6/1P6/8/7K/8/8/1k6 w - - 0 1[/fen]
  • 1.Kh8 - remis
[fen]6K1/8/6k1/8/8/p7/P7/8 w - - 0 1[/fen]
  • 1.Kb1 - white wins
[fen]8/8/5k2/8/p7/8/1PK5/8 w - - 0 1[/fen]
  • 1.c7 - white wins
[fen]2k5/8/1pP1K3/1P6/8/8/8/8 w - - 0 1[/fen]
  • 1.Ke6 - remis
[fen]8/8/8/5K2/7p/8/5P1k/8 w - - 0 1[/fen]
  • 1.Kg6 - remis
[fen]8/6p1/k1P2p1p/7K/8/8/8/8 w - - 0 1[/fen]
  • 1.Ba7 - white wins
[fen]1B6/8/7P/4p3/3b3k/8/8/1K6 w - - 0 1[/fen]

Best regards,
Micha
Larry
Senior Member
Posts: 2269
Joined: Wed Aug 01, 2007 8:42 am
Location: Gosford, NSW Australia

Post by Larry »

chessguru wrote: Image
Could you please tell me the link where you got this image from? I see
you have cut the full image to show only what is needed to make your
point, but I'm interested in the link to the full article.
thanks,
Larry
User avatar
chessguru
Member
Posts: 47
Joined: Mon Feb 21, 2011 7:33 am
Contact:

Post by chessguru »

Hi Larry,

no problem.

This is the BT-2630-Test (Bednorz-Tönissen-Test / Calculation on 2630 Elo)

You can find the link here: The test results can be found at the bottom of the page. Two resolutions are available: The tables should be in the wiki, but I'm too lazy. Perhaps in the coming days...


Five entries you can find for the Saitek Risc 2500.

BT 2630 - Seite 1:
  • 2316 Test-Elo | 2 MB offensiv
    2310 Test-Elo | 512 KB offensiv
BT 2630 - Seite 2:
  • 2280 Test-Elo | 2 MB activ
    2271 Test-Elo | 2 MB normal
    2240 Test-Elo | 128 KB normal
Best regards,
Micha
Reinfeld
Member
Posts: 486
Joined: Thu Feb 17, 2011 3:54 am
Location: Tacoma, WA

Post by Reinfeld »

Yes, blast it, I got sucked in: picked up a (reportedly) new-in-the-box Atlanta from the dreaded auction god.

I couldn't stop myself. I've been casually searching for a year, and I've never even seen one used. The damage: roughly 600 USD.

I blame Steve, because he once advised me that new is always better - and Larry, because he answered two questions I asked a year ago in this thread:

Question A: Is Atlanta the strongest of all Morsch machines?

Larry:
Yes, Atlanta is clearly the strongest of all Morsch machines.


Question B: What is the strongest dedicated that will run on batteries?

Larry:
I think we can say with confidence that the Atlanta is the strongest
battery powered dedicated chess computer ever made.
OK. That's why it was worth it to me. I love plastic and I love strength. I can't store wood, and I can't justify the expense. I have the Mephisto/Saitek Master, but it clearly ranks second on the Morsch meter.

The Atlanta gets me four fifths of the way to the Plastic World Cup, a category I created for hobby rationalization purposes:

Kittinger
- Star Diamond (aka Star Sapphire)
OWNED

de Koning
- RISC 2500 2 MB
OWNED

Morsch
- Atlanta
OWNED!

Spracklen
- Designer 2325 (Mach IV)
OWNED

Lang
- Berlin Pro
??

One of these days, I'll snatch a Berlin Pro, and my life will be completely wasted.

- R.
"You have, let us say, a promising politician, a rising artist that you wish to destroy. Dagger or bomb are archaic and unreliable - but teach him, inoculate him with chess."
– H.G. Wells
Larry
Senior Member
Posts: 2269
Joined: Wed Aug 01, 2007 8:42 am
Location: Gosford, NSW Australia

Post by Larry »

Reinfeld wrote:Yes, blast it, I got sucked in: picked up a (reportedly) new-in-the-box Atlanta from the dreaded auction god.


- R.
Congrats on the Atlanta. I think you paid 450 euro, just 10 euro more
than I paid for my Atlanta as a buy-it-now on German ebay about 5
years ago. Mine was also complete and minty.I thought you got a good buy.
The Atlanta will kick the Berlin Pro's butt on blitz time control, but on
longer times the Berlin Pro will show itself as marginally stronger.(Berlin
Pro does not run on batteries).
have many hours of fun...
L
IvenGO
Member
Posts: 298
Joined: Tue Oct 18, 2011 5:37 am
Location: Moscow, Russia

Post by IvenGO »

Hello folks!

I got Montreux (128kb) several days ago + bought 4 x 512kb RAM planks for upgrade but would like to ask is there any practical reason (for increasing strength) to make it if I play only 30-60 min / game? Or this computer values more with original RAM size?

P.S. Past year got Star Diamond for 120euro and Atlanta for 150euro both in close to "as new" conditions 8)
IvenGO
Member
Posts: 298
Joined: Tue Oct 18, 2011 5:37 am
Location: Moscow, Russia

Post by IvenGO »

P.P.S Will the computer start-up and work correctly if I'll install only one 512kb RAM plank instead of original 4 x 32 ?

Thanks in advance!

Atlanta beats Star Diamond 7.5 : 2.5 in 30min/game match :wink:
Post Reply