The Fidelity Excel Mach II & The 1987 World Open

[Cyberchess]

Back in 1987, I had the summer off from college and decided to work full-time, enjoy some swimming and the great outdoors on my days off, and play in The 1987 World Open Chess Tournament in Philadelphia, PA.
Came late June, I found myself cramming opening preparation and playing thru some instructive games in the hopes of being inspired to higher realms of play. I had been rated about U.S.C.F. 1900 at the time, and had hoped to make a big splash in the A section.

Upon arrival at the tournament, I noticed that Fidelity Electronics had entered their new Excel Mach II 16 Bit 68000 unit in the U-2200 section. There was a notice pinned to the wall for players that didn’t wish to compete against the computer, as they don’t actually comply with U.S.C.F. rules and regulations pertaining to opening books, etc. Not surprisingly, a large crowd of entrants lined up to sign the exclusion list, so Fidelity offered additional cash incentives to players willing to take on the silicon dragon.

After scoring a quick win in round 1 due to my opponent’s inattentiveness, I found myself unfortunately saddled with drawn game after drawn game. When my 4th round game also resulted in a drawn K+R vs. K+B endgame, my opponent and I had both been knocked out of any shot at decent prize money, so we were forced to make a mutually beneficial compromise deal. One of us would have to lose the point. True, it was unethical, but handshake deals were common practice at all levels of play, and I really wanted to minimize my losses and salvage what was left of my long 4th of July weekend, so I agreed to sell my half point.

With that out of the way, I decided to see how things were progressing with the Fidelity hopeful. The game in progress looked promising for the human challenger, and several onlookers were gathered around. He had wisely played a closed system against the machine and had emerged with a slight positional advantage, though no clear-cut path to victory as far as I could tell. He chose to proceed with what I believe to have been a premature, hastily construed attack, and soon found himself on the brink of material loss. Much to his credit, he did manage to avert loss by way of perpetual check. He was one of the fortunate ones that collected the Fidelity cash bonus and walked away unscathed.

When all was said and done, the Excel Mach II 68000 had earned a U.S.C.F. 2265 rating and began selling in the fall of ‘87. I passed the unit by due to my huge disappointment with the Par Excellence, though I would have liked to own such a strong dedicated unit at reasonable cost.

The housing was reminiscent of the early ‘80s Sensory Challenger units. According to this website, this computer is said to be the first to use endgame hash tables:

http://chesscomputer.org/fidelity-excel ... hallenger/

Please note that the photos are actually from the 1988 Mach III which uses the same casing.

Happy Collecting!

John

[Steve B]

my opponent and I had both been knocked out of any shot at decent prize money, so we were forced to make a mutually beneficial compromise deal. One of us would have to lose the point. True, it was unethical

According to this website, this computer is said to be the first to use endgame hash tables:

Well the Fidelity 68000(Mach I) was the first to use Hash Tables

anyway...in an effort to return you to those joyful ..although ethically challenged ..days of yesteryear ...

you can grab this Designer Mach III
Same computer as the Excel Mach III but in the Designer Housing

http://www.ebay.com/itm/231159530209

Ethically Chess Challanger Regards
Steve

[Cyberchess]

Well the Fidelity 68000(Mach I) was the first to use Hash Tables

anyway...in an effort to return you to those joyful ..although ethically challenged ..days of yesteryear ...

you can grab this Designer Mach III
Same computer as the Excel Mach III but in the Designer Housing

http://www.ebay.com/itm/231159530209

Ethically Chess Challanger Regards
Steve

Yeah, it just goes to show how failure to properly prepare for a major tournament can lead one down the slippery slope of turpitude and depravity. Oh well, at least I fessed up after all these years.

Kudos to the designer of the “Designer Mach III” unit. The original Sensory Chess Challenger, Excel Mach I,II,III, etc. units should henceforth be dubbed “The Christmas tree models” with all those tacky protruding LEDs everywhere. I find these to be especially obtrusive when the squares are so small on these plastic units.

Regards from the land of fire and brimstone…

John

[Steve B]

Well the Fidelity 68000(Mach I) was the first to use Hash Tables

anyway...in an effort to return you to those joyful ..although ethically challenged ..days of yesteryear ...

you can grab this Designer Mach III
Same computer as the Excel Mach III but in the Designer Housing

http://www.ebay.com/itm/231159530209

Ethically Chess Challanger Regards
Steve

Yeah, it just goes to show how failure to properly prepare for a major tournament can lead one down the slippery slope of turpitude and depravity. Oh well, at least I fessed up after all these years.

Kudos to the designer of the “Designer Mach III” unit. The original Sensory Chess Challenger, Excel Mach I,II,III, etc. units should henceforth be dubbed “The Christmas tree models” with all those tacky protruding LEDs everywhere.

Actually the Excel Mach's housing design does have a name ascribed to it already
collectors worldwide affectionately refer to it as the "Hump Back" Design

Quasimodo sends his Regards
Steve

[Monsieur Plastique]

My term for that design of housing was Gelato.

[ricard60]

When all was said and done, the Excel Mach II 68000 had earned a U.S.C.F. 2265 rating and began selling in the fall of ‘87. I passed the unit by due to my huge disappointment with the Par Excellence, though I would have liked to own such a strong dedicated unit at reasonable cost.



John

But was not Mach III that was rated 2265 and not Mach II?

Designer 2265 regards
Ricardo

[Cyberchess]

Yes, I believe it was actually the third version that received the official master rating printed boldly along the carton.

http://electronicchess.free.fr/images/excelmach3.jpg

The Mach II is often cited as having the Los Angeles program. Now here’s where it gets confusing: there’s a II, a IIb and a IIc. Fidelity is like a fickle date I once had.

If anyone has a Mach II box, please let us know if states the machine’s rating anywhere.

Confusing versions regards…

John

[Larry]

My term for that design of housing was Gelato.

In Europe they call that housing "bathtub"
For ten years I've been looking for an Excel Mach4,
L

[Steve B]

Yes, I believe it was actually the third version that received the official master rating printed boldly along the carton.

http://electronicchess.free.fr/images/excelmach3.jpg

The Mach II is often cited as having the Los Angeles program. Now here’s where it gets confusing: there’s a II, a IIb and a IIc. Fidelity is like a fickle date I once had.

If anyone has a Mach II box, please let us know if states the machine’s rating anywhere.

Confusing versions regards…

John

Mach III was first to be rated at 2265 USCF
Mach IV got the 2325 USCF Rating

Excel Mach I(68000) and Excel Mach II boxes do not sport the CRA Rating banner nor do the computers have the "Official" CRA Rating Certificates

Excel/Designer Mach III's and Mach IV's all have the certificates and the
Labels on the boxes
the Designer Mach IV was not sold in a box (only a leather carry case) so no label there

Of course the fact that the computers play no wheres near this level is...

A different Matter Entirely Regards
Steve

[Cyberchess]

Mach III was first to be rated at 2265 USCF
Mach IV got the 2325 USCF Rating

Excel Mach I(68000) and Excel Mach II boxes do not sport the CRA Rating banner nor do the computers have the "Official" CRA Rating Certificates

Excel/Designer Mach III's and Mark IV's all have the certificates and the
Labels on the boxes
the Designer Mach IV was not sold in a box (only a leather carry case) so no label there

Of course the fact that the computers play no wheres near this level is...

A different Matter Entirely Regards
Steve

Ah yes, the CRA. Sure glad I’m not the only one roasting down here.

My Novag Diamond was also the subject of much controversy back in the mid ‘90s.

http://www.stmintz.com/ccc/index.php?id=303623

While I’m not going to publically lambaste one of my all-time favorite schach toys, I would encourage potential buyers/bidders of all dedicated units to first consult the appropriate SSDF rating list. All ratings are published in ELO format, though I will post the conversion formulae for converting this rating to the equivalent USCF and BCF ratings.

Now what did I do with that article

Franticly searching through archives…

John

[Monsieur Plastique]

Although it needs to be remembered the SSDF dedicated ratings were literally reduced by 100 points overnight over a decade ago. Whilst it is possible the machines did not play at the previous (100 point higher) ratings, I've never been entirely comfortable with chess computer ratings suddenly dropping like that unless it is because the said machines played more actual rated games and the rating dropped because of poor actual results. I suspect the real playing strength sits somewhere in between old and new, albeit probably closer to the reduced (new) ratings.

Also, we can't point the finger at Fidelity for inflated rating claims, since every Novag box for example has also carried a similar inflated rating. Infact I've pretty much noticed that for the most part, machines tend to be advertised at around 300 - 400 points over their real strength, with the main exceptions being Saitek from the mid to late 90s who actually seemed to advertise their machine strengths according to the (then current) SSDF scale.

[Cyberchess]

The SSDF generated ELO ratings serve only as my relativistic guide to computer/program strength. Even prior to the SSDF rating deflation, a computer that I had personally tested (circa 1984) caused me to question the viability of equating these computer ratings with human ratings.

My chess computer retailer allowed potential buyers a trial period of about 10 days in which to take the prospective machine home and try it out. During this period, the customer was allowed to return the machine with all original packing for a full, no questions asked, refund. In those days I received mailings pertaining to current and past chess computer offerings, so I decided to try out the then new Novag Supper Connie (A.K.A. “The Blitz Monster”). I had been a long established (roughly 5 years) USCF A player rated approximately USCF 1900 at the time. During the trial period, I figured that my “good chess days” and “bad chess days” would cancel out, yielding a decent estimation of sparring partner potential. The “Blitz Monster” maintained a small but certain plus record against me throughout the testing period. This machine was nothing like the Fidelity Sensory Chess Challenger that I had thrashed so easily only about 2 years prior. I later learned of a game the Super Connie played in which it outplayed Diane Savereide, a top female player and US Chess Master:

http://www.schach-computer.info/wiki/in ... stellation

Curiously, the SSDF 1992 Rating List has the Super Connie rated at only 1800 ELO, and this before the 100 point rating deflation of about 10 years ago. Depending on which formula one uses to convert ELO ratings to USCF ratings, this new scale would put this machine somewhere around the upper part of the USCF B section, which makes no sense when compared to the play of humans with similar ratings.

On another note, I’m trying to locate an article printed by the USCF on this very topic.

Rating rant regards,
John

[Monsieur Plastique]

So are you saying you think they are under-rated or over-rated? I did read your post twice but am still not sure

I was the operator of the Super Constellation in an Australian tournament back in the 80s when the machine was state of the art here (i.e. there were still no strong Mephistos being imported and the Fidelity machines were still a bit behind). I remember it did not acquit itself particularly well in the under 1800 tournament. A mediocre result which included some sad thrashings by opponents in the low 1700s (Australian rating).

But back then (and even now), the Australian ratings were significantly lower for a given playing strength than FIDE ratings, let alone USCF ratings. An Australian in the low 1800s back then would have been in the mid to high 1900s FIDE and well over 2100 USCF. So I guess the result in that human tournament actually correlated to the official USCF rating the Super Constellation actually had back then.

One thing that has changed is that humans through databases have much better opening preparation than they would have three decades ago. I really wonder how well these machines really would go today against humans if, hypothetically, they had opening books as good as the commercial ones included with the best engines.

Either way, I think it is fairly safe to say that the current SSDF rating of a dedicated machine is a guaranteed "minimum" playing strength, regardless of one's personal playing style or ability to exploit the machine's weaknesses.

One thing I feel is that humans might actually have become a little "lazy" against dedicated machines. By that I mean that the average player these days needs to significantly handicap an engine to get a fair game. For me it is that ELO slider. But as soon as you weaken these engine, they will still sometimes play obvious blunders - even at expert level and beyond - which none of the dedicated machines would have played (since they are always playing at maximum possible strength).

So I actually think the dedicated machines would hold up pretty well.

[Steve B]

Curiously, the SSDF 1992 Rating List has the Super Connie rated at only 1800 ELO, and this before the 100 point rating deflation of about 10 years ago. Depending on which formula one uses to convert ELO ratings to USCF ratings, this new scale would put this machine somewhere around the upper part of the USCF B section, which makes no sense when compared to the play of humans with similar ratings.

the Current SSDF list shows the SuperC at 1633

Selective Search Magazine shows a rating of 1728 for the Super Connie and the Wiki shows it at 1799
a good rule of thumb when converting from ELO to USCF is to add about 125-150...bringing the SuperC to high 1800's -low 1900's USCF

Chess computers playing against other chess computers always result in somewhat different rating then when playing against humans

Really cant equate one to the other regards
Steve

[Monsieur Plastique]

And who here seriously believes that a 1633 ELO player could draw a match against a Super Conny?

No Chance Regards