ZCT Gaunlet - 40/30 40/20 40/10 Repeating

Discussion of computer chess matches and engine tournaments.

Moderator: Ras

swami
Posts: 6664
Joined: Thu Mar 09, 2006 4:21 am

ZCT Gaunlet - 40/30 40/20 40/10 Repeating

Post by swami »

I don't know if Zach currently has even improved beta, but I seem to recall having seen newer version playing in one of the tournaments, must be ChessWar, If that version is almost similar to .2447, then this tournament and games analysis would be useful. But I'd organise this tourney anyway despite the rather abrupt power shut down and lack of UPS functioning (for this month).

Code: Select all

  Engine             Score ZC
01: ZCT 0.3.2447       0.0/0 · 
01: Natwarlal 0.14     0.0/0   
01: NOW Jan 08         0.0/0   
01: Rodin              0.0/0   
01: ChessAlexUCI       0.0/0   
01: Sorgenkind 0.4     0.0/0   
01: RomiChessP3k       0.0/0   
01: LittleThought-1.03 0.0/0   
01: Danasah397b_WB     0.0/0   
01: Scidlet 3.6.1      0.0/0  
01: Joker1114w         0.0/0   

0 of 40 games played
Name of the tournament: ZCT Gaunlet
Site/ Country: WELCOME-D2B6707, United States
Level: Tournament 40/30 40/20 40/10
Hardware: Intel(R) Core(TM)2 Quad CPU    Q6600  @ 2.40GHz4x  2400 MHz with 2,032 MB Memory
Operating system: Microsoft Windows XP Professional Service Pack 2 (Build 2600)
PGN-File: H:\Arena 1.99 beta 5\Tournaments\ZCT Gaunlet .pgn
Website: 
E-Mail Address: 
User avatar
Zach Wegner
Posts: 1922
Joined: Thu Mar 09, 2006 12:51 am
Location: Earth

Re: ZCT Gaunlet - 40/30 40/20 40/10 Repeating

Post by Zach Wegner »

Hey Swami,

Go ahead and use .2447. I have a few other later versions that are just about equal strength, but nothing higher. The version in ChessWar, .2479, seems weaker by my tests. I could get you a later version later, but I have to compile it. I think the latest version with a similar strength is .2463.

The newer versions have been going through some experimental stuff, so they're not as good. Soon I will go through a bugfix stage where I'll only fix things. I hope this will be reflected in the results!
User avatar
Werner
Posts: 3030
Joined: Wed Mar 08, 2006 10:09 pm
Location: Germany
Full name: Werner Schüle

Re: ZCT Gaunlet - 40/30 40/20 40/10 Repeating

Post by Werner »

Hi Swami, Hi Zach,
I have had so bad results with ZCT Version 0.3.2447 JA
that I decided to wait for an update and not to publish it.

Please Zach tell me, when you are ready with a new version.

Best
Werner
User avatar
Zach Wegner
Posts: 1922
Joined: Thu Mar 09, 2006 12:51 am
Location: Earth

Re: ZCT Gaunlet - 40/30 40/20 40/10 Repeating

Post by Zach Wegner »

What do you mean by bad results? If you mean bugs or time losses, then please let me know. If you mean general weakness, that's another issue, but a quick look shows that there are plenty of engines with much lower ratings on your list.

I am slowly working on better versions. I'm waiting to release until I'm really satisfied, so it might be a while.
swami
Posts: 6664
Joined: Thu Mar 09, 2006 4:21 am

Re: ZCT Gaunlet - 40/30 40/20 40/10 Repeating

Post by swami »

Zach Wegner wrote:Hey Swami,

Go ahead and use .2447. I have a few other later versions that are just about equal strength, but nothing higher. The version in ChessWar, .2479, seems weaker by my tests. I could get you a later version later, but I have to compile it. I think the latest version with a similar strength is .2463.

The newer versions have been going through some experimental stuff, so they're not as good. Soon I will go through a bugfix stage where I'll only fix things. I hope this will be reflected in the results!
Sure Zach, 7 games over and running, I may add more engines into this list later: Rooster, Lime (Engines that ZCT could win against, so I could determine the advantages)
swami
Posts: 6664
Joined: Thu Mar 09, 2006 4:21 am

Re: ZCT Gaunlet - 40/30 40/20 40/10 Repeating

Post by swami »

Werner wrote:Hi Swami, Hi Zach,
I have had so bad results with ZCT Version 0.3.2447 JA
that I decided to wait for an update and not to publish it.

Please Zach tell me, when you are ready with a new version.

Best

Elo is not everything :) I like all kinds of engines, rated 1800 and above. Not only are they competitive in their own league,but they all have unique playing styles, That's more than enough.

ZCT's estimated rating is fairly decent: 2150-2250 elo. Besides, there are more engines out there that play weaker than ZCT.
I'm one of those people who would like to see the lively gain in elo, that's why I'm giving Pros and Cons advice, In a way, I wanted to know more or less how my advice goes strengthwise in future versions, even if it's of little help or anything at all :wink:
User avatar
Werner
Posts: 3030
Joined: Wed Mar 08, 2006 10:09 pm
Location: Germany
Full name: Werner Schüle

Re: ZCT Gaunlet - 40/30 40/20 40/10 Repeating

Post by Werner »

Zach Wegner wrote:What do you mean by bad results? If you mean bugs or time losses, then please let me know. If you mean general weakness, that's another issue, but a quick look shows that there are plenty of engines with much lower ratings on your list.

I am slowly working on better versions. I'm waiting to release until I'm really satisfied, so it might be a while.
Hi Zach,
I had a look in my mails: I was already in contact with you. I am sorry about the confusion - I tried to test 2CPU version. Here is your answer:

"Hello Werner,

I seem to have messed something up pretty bad after the WCRCC. I think
the main culprit for illegal moves is the stuff_pv() function, which
I've removed now. But I'm still getting some problems in SMP mode...

Thanks for your detailed reports. I would suggest keeping ZCT 0.3.2447
until I can get this sorted out."


Now my question: Did you have success and find the reason? Which version can I use now?
Werner
swami
Posts: 6664
Joined: Thu Mar 09, 2006 4:21 am

Re: ZCT Gaunlet - 40/30 40/20 40/10 Repeating

Post by swami »

Code: Select all

Engine             Score     ZC
01: ZCT 0.3.2447       14.0/40 ···· 
02: Natwarlal 0.14     4.0/4   1111 
03: Danasah397b_WB     3.5/4   11=1 
04: RomiChessP3k       3.0/4   ==11 
04: Sorgenkind 0.4     3.0/4   1011 
04: Scidlet 3.6.1      3.0/4   11== 
07: LittleThought-1.03 2.5/4   0=11 
07: NOW Jan 08         2.5/4   1=10 
07: ChessAlexUCI       2.5/4   =101 
10: Rodin              2.0/4   0110 
11: Joker1114w         0.0/4   0000 
11: Rooster            0.0/0        

40 of 44 games played
Name of the tournament: ZCT Gaunlet
Site/ Country: WELCOME-D2B6707, United States
Level: Tournament 40/30 40/20 40/10
Hardware: Intel(R) Core(TM)2 Quad CPU    Q6600  @ 2.40GHz4x  2400 MHz with 2,032 MB Memory
Operating system: Microsoft Windows XP Professional Service Pack 2 (Build 2600)
PGN-File: H:\Arena 1.99 beta 5\Tournaments\ZCT Gaunlet .pgn
Website: 
E-Mail Address: 
Added Rooster, I will add Lime to the tournament with ZCT next.
BubbaTough
Posts: 1154
Joined: Fri Jun 23, 2006 5:18 am

Re: ZCT Gaunlet - 40/30 40/20 40/10 Repeating

Post by BubbaTough »

In a way, I wanted to know more or less how my advice goes strengthwise in future versions, even if it's of little help or anything at all
I am not sure exactly what you mean here...you want estimates on how much changes in future versions were based on your advice? I would be happy to give you feedback on your feedback in whatever form you would like given the effort you have put in on behalf of LL. For example: version 0.31 main improvements over 0.30 is reduced sacrificing in certain types of positions and redone rook usage with a very rough ELO improvement of 10 points.
swami
Posts: 6664
Joined: Thu Mar 09, 2006 4:21 am

Re: ZCT Gaunlet - 40/30 40/20 40/10 Repeating

Post by swami »

BubbaTough wrote:
In a way, I wanted to know more or less how my advice goes strengthwise in future versions, even if it's of little help or anything at all
I am not sure exactly what you mean here...you want estimates on how much changes in future versions were based on your advice? I would be happy to give you feedback on your feedback in whatever form you would like given the effort you have put in on behalf of LL. For example: version 0.31 main improvements over 0.30 is reduced sacrificing in certain types of positions and redone rook usage with a very rough ELO improvement of 10 points.
Very good! 10 elo in higher level in just 2 fixes is pretty much Ok Imo. I will surely try to finish the rest of the games and build the report. I'm sure you can fix the remaining bugs from the report and gain more elos 8-)