Not logged inRybka Chess Community Forum
Up Topic The Rybka Lounge / Computer Chess / Mundial Chess official thread
1 2 3 4 5 Previous Next  
- - By Ozymandias (****) Date 2009-10-31 19:41
Hello chess friends,

My name is Juan and I've been participating in freestyle tournaments for several years under the nick "Ozymandias". As most of you, I've seen the last year and a half go by without any significant open to talk about. Well, all this is finally behind us, because as of January 8th, 2010, the mother of all competitions is coming to us at mundialchess.com.

I know some will be reluctant to believe this event will actually take place and, with so many delays in the past (including a couple of our own), who's to blame them? But the thing is, after seeing the amount of work these people have put into this project I not only believe it's going to happen, but I also will try to help shape it. From this moment on, I'm going to represent mudialchess on this forum. Any doubt, suggestion or criticism you'd like to express, you may do so on this very thread.

For all of you who have already registered or are considering that option, please don't hesitate to try the last working beta here. The more of us the merrier, as we all want the tournament to run smoothly. The server has only been running for a couple of weeks, so please come and test our application, strength in numbers!

The software is pretty bare at the moment, but it's minimalism also makes for an intuitive interface. Currently you only get to play at the tournament's time control of 10'+1'', but more options like rematch, algebraic notation, arrows to indicate last move, board flipping and much more are coming soon. The best news right now? It's solid, stable and FIDE-compliant. You can play a game of chess and rest assured nothing strange is going to happen. But, all this could change under considerable strain (as seen in past events) which is why I'm asking you to come and stress it. There will be more tournaments trough 2010, so let's start with the right foot on this one.

That would be all for now, but I have read about legitimate concerns that deserve some answer and I'm going to try and provide one:

1- How are they going to pay so much money? Prizes for this tournament are guaranteed, no matter what; evidently, in the long run, the companies behind this project expect subscriptions to their playing site to produce a healthy revenue. But that's the long run and it's up to them to worry about it.

2- Delayed twice, why not again? The first dates were rushed to make it coincide with the VIII International Chess Festival in Benidorm, now that it not longer needs to make that appointment, the established date of January 8 looks like a sure bet.

3- Is it 40 or 20€? Depends on what you want, if you're only interested in playing the tournament go ahead and save half the fee, but as I said, this won't be an isolated event, which is why we're still getting more full registration payments than anything else. It's more or less like it was being a member of the CSS or paying for each individual PAL. The difference here is that not only do you get a discount, but you also have a full year's access to the playing site.

I hope to have provided you with some useful information about this puzzling tourney and to witness a new dawn for Freestyle soon.

Regards,

Juan José Molina
Parent - - By Nelson Hernandez (Gold) Date 2009-11-01 03:32
Juan, it seems a bit strange that even now, after two abortive tries, no one has attempted to persuade the tournament organizers that 10+1 is not really a sustainable Freestyle time control.  It really gives no opportunity for skilled centaurs to show their ability, and it puts entirely too much premium on human OTB skill.  I fear is that all the sponsorship money will warp the whole nature of Freestyle and that this time control will become the new standard.

As you know, a typical Freestyle game runs in neighborhood of 40-50 moves.  Seldom are there protracted games, though they do happen from time to time.  But even assuming 60 moves you are compelling the players to kick out a move every 10 seconds on average.  Given that Freestylers often use multiple computers, they need several seconds to register their own and their opponent's moves.  Thus every move they make at 10+1 will involve about 50% administrative time (to make moves on their machines) and about 50% to make a move-choice. 

I don't know how you feel about this but I think that such a frantic pace favors a) people with strong hardware, b) people with video game reflexes, and c) strong OTB players who can, at some stage of the game, abandon the computer and play very rapidly.  Good centaurs have no niche in such a time control, really, and as a result all opportunity to see the full creativity that can emerge from a well-founded man-machine partnership is lost.

There is also the possibility that a player could game the system.  For instance you could adopt a Pablo Restrepo strategy of creating a pawn blockade with the intent of inducing the opponent to run out of time.  This strategy was attempted against us in Benidorm two years ago and it might well have worked at 10+1.

If there will only be one game played per day then I see no reason why the time control could not be 75+15 or at least 60+15.  Is it too late for the organizers to consider my arguments and make adjustments?
Parent - - By Ozymandias (****) Date 2009-11-01 20:00
The organizers are well aware of our dislike for this fast time control, but as I told Ciron in another post, we're a minority (5 - 10% of all registrations at best). Add to that, the ones who will absolutely refuse to play because of that are even a small fraction of that minority, which means we'll hardly be heard in that regard. Really, it's all a question of practice. When I started practising for this tournament I could hardly play more than 40 or 50 moves, right now I'm between 80 and 100 depending on what kind of middlegame I'm facing.

It's clear that Lukas will be very pleased with this tournament as it is, but I would fear Eros or Jóse (if he recovers his sight) more than a GM with a comp. The organizers also think that GMs will have the advantage (even after Tiviakov was demolished by Jóse) but we'll soon see who's right. It's really a pity the fact that Anson can't play, because I think that the ones who have more options are the fastest centaurs with the best books. Summarizing, they must be fast, but they must be top Freestylers too.

And you shouldn't worry about blitz becoming the new standard, this tournament is so only because we're all mixed (human and centaur players). This situation doesn't have to be repeated.

As for the Father techniques, I fear them as much as GMs, little. Anson's problem was in the age of Rybka 2 and "slow" quads. Besides, what you have to worry about is wining the game, if the position is clearly won, you will always have the option to ask for the GM team there to check it, so you won't lose on time.
Parent - - By Nelson Hernandez (Gold) Date 2009-11-01 20:26
Anson could not play in October.  I am not sure this holds true for January.  I am awaiting his reply.  My sense is that even if we don't finish first this is almost a slam-dunk for him to make some amount of money.  He can certainly acquire credible hardware and he does have video-game reflexes.
Parent - By Ozymandias (****) Date 2009-11-01 20:48
And with him not having ELO he can play all of the three tournaments ;-)
Parent - - By Lukas Cimiotti (Bronze) Date 2009-11-01 20:53
I won't participate - Jiri will. And no - I'm not pleased at all. I'd prefer a reasonable time control like at least 60+15.
Parent - - By Ozymandias (****) Date 2009-11-01 21:01
Don't you think that a sorter time control exacerbates hardware differences? Have you tested that?
Parent - - By Lukas Cimiotti (Bronze) Date 2009-11-01 21:04
Yes, I think at shorter time controls our/Jiri's chances rise. But still it's ridiculous to play at 10 + 1. I prefer a fair competition in which human thought might help.
Parent - - By Ozymandias (****) Date 2009-11-01 21:12
You're being far too noble, from an egoistical point of view, you should be pleased.
So the cluster will play under Jiri's name? Cool.
Parent - - By Lukas Cimiotti (Bronze) Date 2009-11-01 21:23
Jiri will use the cluster - if this event ever takes place - which I think is still unclear. Jiri is a very good Freestyle player - it would be a shame if he only had to transmit moves the cluster tells him.
Parent - - By Ozymandias (****) Date 2009-11-01 21:29
I can tell you the same I said to Nelson, you can play a centaur game at 10 +1 and not just be the computer operator. I've done it, it's just a lot more hectic at this speed.
Parent - - By Nelson Hernandez (Gold) Date 2009-11-01 21:54
It certainly does help a lot if you're deeply booked.  It's hard to tell what's better, following a book line to play quickly or exiting book quickly and betting on the cat-like video-game reflexes.
Parent - - By Ozymandias (****) Date 2009-11-01 22:00
Depends on the what gives you more confidence, your book or your skills. With a cluster I would exit sooooon, though (no matter what).
Parent - - By Nelson Hernandez (Gold) Date 2009-11-02 00:32
Yep, that's a no-brainer.  The only hope at this time control is that the cluster operator will mouse-slip.
Parent - - By Auryn (***) Date 2009-11-02 02:08
What discourages me the most about this kind of tournament, is that for sure some skilled computer players will find a way to play with an auto engine, or at least with a program opened together with the mundial chess interface chessboard. Those guys with a fast computer playing auto will have the biggest winning chances in my opinion, at that fast time control. So it would be no surprise for me if one or more unknown players reach the first places ...
Parent - - By Banned for Life (Gold) Date 2009-11-02 03:08
Automating the interface is certainly practical, but since this is illegal, it would have to look like there was a person involved. The short time controls also increases the value of a deep book though.
Parent - - By Auryn (***) Date 2009-11-02 03:31
I didn't know it was illegal, anyway my point is that the players who can manage better to automatize an engine at the limits of legality have a big advantage, so in this tournament not the strong players will win, but the most skilled ones in finding half-cheating ways to play.
Parent - - By Banned for Life (Gold) Date 2009-11-02 03:50
Sure, automation in bringing in and sending out moves, a fast computer, and a good book, maybe one with offbeat lines, will be key. Centaur skill probably won't be that important at 10-1. That's fine if it is what the sponsors want, but I get the feeling that they are doing this with the intention of increasing the importance of human interaction, which strikes me as rather bizarre.
Parent - - By Ozymandias (****) Date 2009-11-02 19:59
Yes, they want the human factor to have some weight. After all, that's more than 90% of the players, what's so bizarre about it?
Parent - - By Banned for Life (Gold) Date 2009-11-02 20:26
The human influence is increased at long time controls. Its not at all clear that a top player with computer assistance has any advantage against a fast computer running a strong engine, coupled to a good book and 6-man TBs, with instant communication with the server at 10-1 time controls. Of course at longer time controls, the situation would be reversed. The bizarre aspect is that the organizers seem to believe that going to shorter time controls will work against pure engine entries. It just doesn't work that way...
Parent - - By Ozymandias (****) Date 2009-11-02 20:35
"with instant communication with the server" is something that should be proven possible and you also would have to do it without nobody detecting it when it reaches the server, not a real worry.
As for the time not being favourable to humans, just imagine a master playing an easy ending (for him) against somebody who needs to look at the engine's score all the time to survive. I would bet on the master there.
Parent - - By Banned for Life (Gold) Date 2009-11-02 20:48
"with instant communication with the server" is something that should be proven possible and you also would have to do it without nobody detecting it when it reaches the server

Its not really clear from the current rules whether a direct connection is legal or not. Earlier versions of the rules made it illegal to reverse engineer the client.

just imagine a master playing an easy ending (for him)

I'm willing to stipulate that a person with good endgame skills and 10 seconds per move would be able to outplay a fast engine in many types of endgames, but I'm pretty sure he wouldn't reach these endgames often enough to prevent getting slaughtered.

against somebody who needs to look at the engine's score all the time to survive.

I'm postulating that some people will automate the interface and will not have to look at the engine's score (i.e. the engine will play the games by itself). I suspect that any feelings that this will not be possible will be overtaken by events at the tournament. :-)
Parent - - By Ozymandias (****) Date 2009-11-02 21:02
"Its not really clear from the current rules whether a direct connection is legal or not"

Let me clarify it for you, it's illegal. I'll try to update the rules.

"I'm postulating that some people will automate the interface and will not have to look at the engine's score (i.e. the engine will play the games by itself). I suspect that any feelings that this will not be possible will be overtaken by events at the tournament."

We'll have to play the waiting game on that one.
Parent - By Banned for Life (Gold) Date 2009-11-02 22:38
Let me clarify it for you, it's illegal. I'll try to update the rules.

Thanks for the clarification. I think updating the rules would be a good idea because right now this is open to interpretation.
Parent - By Nelson Hernandez (Gold) Date 2009-11-02 12:35

> not the strong players will win, but the most skilled ones in finding half-cheating ways to play.


Yes, and I can see your gears working all the way across the ocean, you devious and twisted devil!!  :)
Parent - By Ozymandias (****) Date 2009-11-02 19:57
What exactly do you refer to as "half-cheating"?
Parent - - By Vempele (Silver) Date 2009-11-02 12:43

> Automating the interface is certainly practical, but since this is illegal,


How so?

It is not permitted to interfere with the technical processes of the server or the game. Above all, it is not allowed to slow the game or to manipulate the system itself. Any attempt to hack the server entails the elimination of the account without prejudice to any criminal proceedings that might ensue.

As long as the automating program doesn't directly touch the client, I don't see how it could possibly interfere with the "technical processes" of anything.
Parent - By Banned for Life (Gold) Date 2009-11-02 16:47
It is not permitted to interfere with the technical processes of the server or the game. Above all, it is not allowed to slow the game or to manipulate the system itself. Any attempt to hack the server entails the elimination of the account without prejudice to any criminal proceedings that might ensue.

The original language precluded the automation of the interface with the client. This language is much less clear, but I'm guessing the organizers will not look kindly on any team that receives or sends moves in any manner other than via interaction with the output of the client GUI. At these time controls, the automation of I/O could save a team up to half of their time, and an unaided engine could certainly win, so I'm guessing we may see this tested at some point in the not too distant future.

I'm guessing that this would be disallowed based on the catch-all "not allowed to [ ] manipulate the system itself." clause. I also suspect that the organizers meant to say hack the client in the next sentence, but that is sheer conjecture...
Parent - - By Ozymandias (****) Date 2009-11-02 19:56
"a person involved"? like in, an insider? Sure you would, good luck with that too.
Parent - - By Vempele (Silver) Date 2009-11-02 20:02

> "a person involved"? like in, an insider?


No. Like there was a human making the moves.
Parent - By Ozymandias (****) Date 2009-11-02 20:14
Ok, I get you now.
Parent - - By Ozymandias (****) Date 2009-11-02 19:54
I also exposed this possibility to the organizers, but the person responsible for the decisions there assures me that's next to impossible. That player would need the code to know how to send the command to the application, since we are talking about some binary, encrypted, closed code that only one programmer has (no .exe to install or asp or anything)... well, good luck with that.
Parent - - By Vempele (Silver) Date 2009-11-02 19:57
Or they'd have to know how to control the mouse programmatically (and recognize the opponent's moves by looking at the window)...
Parent - - By Ozymandias (****) Date 2009-11-02 20:21
With ActionScript 3 they won't know how to.
Parent - - By Banned for Life (Gold) Date 2009-11-02 20:33
Its certainly possible with the right skills and motivation. I'm guessing this would be much simpler than reverse engineering R3, and the people who did that did it just for fun...
Parent - - By Ozymandias (****) Date 2009-11-02 20:41
With R3 you have the .exe to "work" with.
Parent - - By Banned for Life (Gold) Date 2009-11-02 20:57
A quick review of Adobe's literature on Actionscript 3 doesn't even list security as a goal. I know nothing about AS3, but this alone makes me very skeptical about your claim that the client cannot be reverse engineered.
Parent - - By Vempele (Silver) Date 2009-11-02 22:10 Edited 2009-11-02 22:13
Most of the "security" seems to come from using a newer version than the #1 Google result for "swf decompiler" can fully understand.

Downloading the #1 Google result for "Actionscript 3 decompiler"...
Parent - - By Vempele (Silver) Date 2009-11-02 23:10

> Downloading the #1 Google result for "Actionscript 3 decompiler"...


Which won't do anything useful until you pay them $80. This, however, will. It doesn't let you copy the decompiled code to the clipboard, though. Here's a sample:

//class GameData
package com.snc.chess.model.playZone
{
  import com.snc.base.utils.*;
  import com.snc.chess.model.application.*;
  import com.snc.chess.model.remoteClass.*
  import mx.events.*;

  public dynamic class GameData extends com.snc.base.utils.BaseData
  {
    // snip
   
    public function getOtherUser(arg1:com.snc.chess.model.application.UserData):com.snc.chess.model.application.UserData
    {
      if (whiteUser.id == arg1.id)
      {
        return blackUser;
      }
      if (blackUser.id == arg1.id)
      {
        return whiteUser;
      }
      return null;
    }
   
    // snip
  }
}
Parent - By Banned for Life (Gold) Date 2009-11-02 23:26
Doesn't look like RE will be that tough and with a 33 822 Euro top prize, this could be a very interesting tournament.
Parent - - By Vempele (Silver) Date 2009-11-02 22:08
http://www.mundialchess.net/chess.swf

Right-click the above link and "Save linked content as..."
Parent - - By Uly (Gold) Date 2009-11-02 22:13
A flash? Doesn't that mean that reverse engineering or a controlling mouse application aren't required?
Parent - - By Vempele (Silver) Date 2009-11-02 23:18 Edited 2009-11-02 23:22
What do you mean?

In any case, decompiling the client was utterly trivial.
Parent - By Ozymandias (****) Date 2009-11-03 21:21
Decompiling doesn't give you all the code you need, some part is blocked, which means... good luck with that. ;-)
Parent - - By Vempele (Silver) Date 2009-11-02 20:33
Why on Earth would they use Actionscript 3? I'm thinking of a simple UCI command-line interface that also happens to control the mouse and analyze screenshots of a chess board.
Parent - - By Ozymandias (****) Date 2009-11-02 20:40
And how do you get it to pass the moves to the playing app without having the code? Really, if you find it that simple and that the people working there don't know what they have in their hands then nothing we say is going to change your mind.
Parent - - By Vempele (Silver) Date 2009-11-02 20:54 Edited 2009-11-02 21:07

> And how do you get it to pass the moves to the playing app without having the code?


The same way a human would:

Move mouse to somewhere on the from-square.
Press the left mouse button.
Move mouse to somewhere on the to-square.
Release the left mouse button.
Parent - - By Ozymandias (****) Date 2009-11-03 21:19
I'm awaiting an answer on this one.
Parent - - By Vempele (Silver) Date 2009-11-05 21:37
Still no answer?
Parent - - By Ozymandias (****) Date 2009-11-05 21:52
The only one I could give you right now is my take on it, but since I'm no programmer I don't think you'll find it of much interest. The people responsible for the technical side of the event are quite busy at the moment, so this isn't in their priority list, but rest assured they'll give it a look before the tournament. Besides, even if it could be implemented, it would be detectable and sanctionable.
Up Topic The Rybka Lounge / Computer Chess / Mundial Chess official thread
1 2 3 4 5 Previous Next  

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill