Hello all,

from my computer chess collection, I most usually run swiss tournaments including a mix of chess computers and chess programs; selected in a range of relatively close ELOs.

On another hand, it is a best practice to power up and use each and every chess computer now and then; so I decided few years ago to run a yearly cup competition dedicated to my chess computers.

My first cup (2019) included 32 devices, later I used 48 chess computers, last year I entered 57 of them, and now I reached the nice number of 64 chess computers - so I am about to run this 64-cup!

It should be a three months or so effort; but considering I am now retired I plan to store and share the games (which I did not, previously).

Strong chess computers can face weak ones in such a cup, this can be unusually fun to watch

How do I proceed:
- 8 groups are created with 8 competitors each; of which one is selected amongst the eight strongest chess computers ("top seed"); and seven from the rest of my collection are randomly assigned to this group.
This ensures the top chess computers will not prematurely eliminate each other during the selection phase.

- Then I run short swiss tournaments, one for each group. Three rounds are usually enough to rank 8 competitors (eventually using Sonneborn/Berger score). Each competitor plays once with the light pieces, once with the dark ones, thus two games per round. All the cup competition is played using 15secs/move level (or closest level).

- This selection phase provides 2+1 chess computers that will keep on fighting, and 5 that are eliminated. The first and second ranks will play for the cup, while the third one will play for a consolation prize.

- 1/8th final for the cup (featuring 16 chess computers) will cross-face for elimination 1st from a group and 2nd from another group (1st of group 1 plays 2nd of group 8; 2nd of group 2 plays 1st of group 7, 1st of group 3 plays 2nd of group 6, and so on...).

- next phases for the cup are classical quarter-final, semifinal, and final plus a secondary/small final for third place.

- the consolation prize for the third ranked computers from the selection phase is played in cup mode, starting at quarter-final step (featuring 8 chess computers).

Obviously the DGT Centaur (expert mode) is expected to win the cup; nevertheless this should not prevent some interesting games to happen. Which chess computer will challenge it as a finalist, and what about the third place? Forecasts for the first three horses are welcome...

MfG,
Tibono

Thank you Tibono.

I will follow this with great interest.

Good luck

Murat

Thank you for this interesting tournament - I'm looking forward to it.
Best
Roberto

Groups are now designed. Let's start with group 1, contenders are:

- DGT Centaur
- Mephisto Mirage
- Conchess Escorter
- Novag Beluga
- Mephisto Mein erster
- Systema Challenge
- Fidelity Sensory Chess Challenger 8
- Novag Citrine

Round 1 matches will follow the above randomly created order; thus first two games will match DGT Centaur with Mephisto Mirage. The Centaur lets no chance to the Mirage; each weak move is merciless punished, and both below games are shortly completed (21 then 24 moves). Funny enough, both are French games. As the Mirage as black chose this opening and lost the first game, I suspect the Centaur wanted to instruct him how to win using the very same opening


I have no clue when the DGT Centaur gets out of its book, as all moves (including the very first ones) are played after around 10 secs.

MfG,
Tibono

I guess the Centaur in Expert Mode is too strong in Expert Mode. Even the King would not been able to beat a Stockfish 8.

__________________
Mein Profil beim ICCF (International Correspondence Chess Federation)
https://www.iccf.com/player?id=89948&tab=3

Hello !

DGT Centaur in Expert Mode 2919 ?, not even the world Champion could beat him, an opponent with 1544 hm, for me a senseless match sorry.

Best Regards

Yes, you are right. Anyway this tournament is not meant to be fair: the fate of the Novag Chess Champion MK1 is not to be envied... By the way, as Scally stated, the Centaur engine is the last development version of Stockfish 8 (310118), prerelease of Stockfish 9 on the very same day...

Therefore, it is Stockfish 9.
MfG,
Tibono

Miniaturansicht angehängter Grafiken

 

Second match for group 1, round 1, is Novag Beluga vs Conchess Escorter (basic module).

According to my Elo scale, the odds were 2/3 to 1/3 in favor of the Beluga. Well, the actual score being 1.5 to 0.5 can be considered well done from the Conchess; but zooming into both games is a bit more deceptive.

The Conchess played rather well as white until late middlegame, reaching a rather balanced position after exchanges, where both computers had equal material: 1 Queen, 1 Rook, 1 minor piece and 6 pawns . This was already an achievement with regards to its weaker Elo; but later the Conchess was 2 Pawns down and failed to provide support to its own center Pawns that could have put some pressure on the Beluga. The Beluga did not shine then, using around 30 moves to slowly convert the game into a win .

On the second game as black, the Conchess manages to get a piece ahead thanks to a blunder move from the Beluga (on move 15), but lets a position repetition later happen. The Beluga can be thankful for that!



MfG,
Tibono

The Vonset L6 v2 had the potential to give Centaur expert a good challenge. But its software designers are very inexperienced.

All levels form 1-22 play in less then a second. They claim it is
working as intended. I say they have no clue.

Too bad.

Murat

Third match in group 1, round 1, is Mephisto Mein Erster against Systema Challenge (using my "Tibono setting" intended to provide a slightly more dynamic play, whilst staying close to the original style: F1=9, F3=12, F4=F5=8, F6=11, F7=9).

The humble "Zen" 4K program from Eric van Riet Paap has of course no chance against Horvath's 32K running on a much faster H8.

The first game is sort of a miniature with only 16 moves, as the Mein Erster blunders very early. On another hand, the Challenge provides evidence it can be very passive: without errors from its opponent , despite playing as white (and despite my setting ), it lets the slow, closed game, reach 30 moves before daring to open the position on the Queen side. Under pressure, the Mein Erster fails to provide sufficient King safety and quickly loses to a corridor mate.



Mfg,
Tibono