ChessConquest - official discussion thread

[yotta]

...and any thoughts are always most welcome

Our original thinking was that with face-to-face chess, you can only play out your moves in advance in your head (not physically on the board), and since playing out your moves on the board makes it much easier we removed the conditional moves link purposefully, in an attempt to keep the online games as close as possible to FTF games.

We are thinking about adding an option in version 2 of ChessConquest that give players the ability to specify before a game is started if the link is visible during a game or not. This way players who feels it to be an unfair advantage can disable it and others who wants the link can enable it.

However, if more players want the conditional moves link to be permanently available, we can look at making a quick change to the current live version of ChessConquest.

So anyone who feels the conditional moves link should OR should not be available can post their view in this forum.

 

[tallfred]

Yotta

Thank you for your invite to comment.

In my view conditional moves are for addressing a problem that is not met in FTF. In FTF when your opponent moves and you know what you are going to do you simply make your move almost instantly. There is no time loss.

In turn-based chess this is not the case. The opponent makes his move, and then goes off and does something else. Some time later you go back, see that the expected move has been made, and instantly make your move. You then go off and wait for the opponent to get back and notice that you have done something. This can take a considerable amount of time in turn-based games.

In other words, conditional moves are a convenience aimed at speeding up games. Until the opponent has made his move and seen the conditional move I do not see them as legally binding. I think that it is wrong to penalize a player for making use of the facility. Therefore a player should be able to see what conditional moves he has made, and even be able to change them, provided of course that the move has not already been played.

I therefore completely support the idea of having conditional moves permanently visible.

 

[yotta]

You make a very strong point TallFred, and you convinced me.

If no-one else objects or gives a counter argument against the link being permanently available, I'll update our task-list to make the change that MarkFace suggested to version 1.

 

[Cicero]

Sweet, hopefully this will help my Chess K/D!

 

[yotta]

Something I've always wondered about is, how big is white's advantage with having the first move?

Lets assume that there hypothetically exists two unbeatable chess players in the world. It is impossible for either one of them to loose a game.

Now you pair them up in a game against each other. Will white always win?

 

[Solitude]

I suppose that if both white and black play perfectly then white should win. However this is just what I'm thinking, I can't really back it up. Or will it always be a draw? Hmmmm. It feels like white should always win if both players play the perfect game.

 

[ViperGTI]

Back in my prime ... I used to prefer playing black and it was my stronger colour. Can't really remember why.

I should actually have a look at my stats on games played thus far, although with my head being all over the place, not sure how accurate the stats will be since I've been pretty good at blundering games in both colours recently .

At the most extremes though, I suppose white should win. They have an extra move and everything.

 

[tallfred]

Something I've always wondered about is, how big is white's advantage with having the first move?

Lets assume that there hypothetically exists two unbeatable chess players in the world. It is impossible for either one of them to loose a game.

Now you pair them up in a game against each other. Will white always win?

In order to answer your question about whether white will always win if both sides “perfectly” (whatever that may be) with complete certainty we need to examine every possible game of chess and choose the moves that give the best result to each player.

In trivial games like tac-tac-toe this is quite simple because the possible number of games is fairly small. Chess is far more complex than that. The mathematician Claude Shannon did some thinking on this (see http://en.wikipedia.org/wiki/Shannon_number) and came to the conclusion that if you were to try to solve chess in the same way as for tic tac toe you would have to look at 10 the power of 120 moves. That is a number too vast for mere humans to comprehend: the observable universe has only 10 to the 80 atoms in it. In other words, for each atom in the universe there are more than one quintillion quintillion chess moves. Oh yes, I’m glad you asked what a quintillion is. It is a billion billion.

Computing such a high number of positions is out of the question, at least in my lifetime. So your question has no proven answer today, but it is still interesting to speculate on.

(Did you think that computers studied the position by looking at every move to a certain depth and then choosing the best move? I am sorry to disappoint you: computers have sneaky ways of guessing which are the critical moves in a given position and they then concentrate on those. This is one of the key ways in which one chess program gets better results than another. This is different from a human who will really only look at a small number of moves, say three or four.)

Taking your question one step further, I wondered what is the practical importance of the fact that White moves first. I took a look at 365chess, a site with 2.78 million games and found White would be (statistically) expected to get 53.9 points and black 46.1 I correlated this to two much smaller game collections containing approximately 14,000 and 56,000 games and found that in those sets white took 54 points and 55.5 points out of hundred.

In other words in practice White has tangible advantage over Black, and will get 20% more points.

However, I expect that this number is skewed by the fact that the games bases contain games by much stronger players than you and me. I suspect that weaker players are not skilled enough to make the first move count, and that White and Black will both come out at close to exactly 50 points out of 100. It might be interesting to see wehether I am right by analysing the results of the games on ChessConquest. Of course, if we were to use those stats we would eliminate all the games that were decided by timeouts.

Do you want to troll your database and have a look at what advantage (if any) White has over Black in ChessConquest?

 

[yotta]

Back in my prime ... I used to prefer playing black and it was my stronger colour. Can't really remember why.

I should actually have a look at my stats on games played thus far, although with my head being all over the place, not sure how accurate the stats will be since I've been pretty good at blundering games in both colours recently .

At the most extremes though, I suppose white should win. They have an extra move and everything.

Had a look at your stats and looks like you are very slightly stronger with black (62 wins) than white (60 wins).

By the way, I see on your ratings chart graph that you are at an all time high... well done ViperGTI!

 

[yotta]

In order to answer your question about whether white will always win if both sides “perfectly” (whatever that may be)...

I suppose a "perfect" game could be defined as, each time making the move that has statistically the highest probability of resulting in a win. Now if we could just find a way to calculate that valuable number

Calculating such a probability, given the number of possible outcomes you mentioned, is a bit above my level of mathematics though. Perhaps a good exercise would be to first try such a calculation on tic-tac-toe OR perhaps on a game with only kings and pawns.

...I’m glad you asked what a quintillion is. It is a billion billion.

I want a quintillionbyte harddrive ...

Do you want to troll your database and have a look at what advantage (if any) White has over Black in ChessConquest?

Better yet, I think I'll add the stats of games (excluding timeout victories) to the stats fields shown at the online page... that way we've all got access to the information.

http://www.chessconquest.com/index.php?p=online

PS. For those who posted requests on the forum, I plan to get my teeth into new feature development this long weekend, so keep a look out for the new updates.

 

[yotta]

I suppose that if both white and black play perfectly then white should win. However this is just what I'm thinking, I can't really back it up. Or will it always be a draw?

Hmmm, if both players are really perfect (unable to loose), then a draw should hypothetically be the most likely outcome.

Once again, when it comes to the numbers TallFred mentioned, I cannot really back this up with a probability equation.

 

[tallfred]

I suppose a "perfect" game could be defined as, each time making the move that has statistically the highest probability of resulting in a win. Now if we could just find a way to calculate that valuable number (Comment 1)

Calculating such a probability, given the number of possible outcomes you mentioned, is a bit above my level of mathematics though. Perhaps a good exercise would be to first try such a calculation on tic-tac-toe OR perhaps on a game with only kings and pawns. (Comment 2)



I want a quintillionbyte harddrive ... (Comment 3)



Better yet, I think I'll add the stats of games (excluding timeout victories) to the stats fields shown at the online page... that way we've all got access to the information. (Comment 4)

http://www.chessconquest.com/index.php?p=online

PS. For those who posted requests on the forum, I plan to get my teeth into new feature development this long weekend, so keep a look out for the new updates. (Comment 5)

Hmmm, there is much to comment here. I have listed my comments, so here goes:

Comment 1) In games theory terms, a perfect zero-sum game (such as chess) might be defined as one in which each player maximized his expected payoff at each move. There are subtle issues that come up in this regard. For example, in a recent game with one of the top players on the site I was in a terrible situation. I played a move to try to force a draw. Other moves would lose slowly and certainly. My opponent made the expected reply and we took the draw. What both of us missed was that instead my opponent could have played a brilliant Grandmaster move that would have have caused me to lose immediately and outright. (Chessmaster was delighted to point this out to me in my post-mortem of the game.) So did I play "perfectly"? Some would say no, because my move did not take into account my opponent's strongest move, and should have lost the game instantly against best play. Some would say yes because in practice it gave my opponent problems that he could not solve. I just don't know. This philosophy thing is not my strong point.

Comment 2) I have programmed a tic-tac-toe program. It is an amazing experience seeing your creation showing expert behaviour, albeit in a simple domain.

Comment 3) A quintillionbyte is technically called an exabyte. For reference, an exabyte is a billion gigabyte. It is actually technically possible to own such a drive: on Wikipedia (http://en.wikipedia.org/wiki/Exabyte - a very interesting article) it says that "Oracle Corporation claimed the first Exabyte tape library with the SL8500 and the T10000C tape drive in January 2011." Does this mean that we are close to solving the problem of storing the information for chess? Hardly, but we have made a (pathetically modest) start though. We are exactly one quintillionth of the way, assuming we store one byte for every chess move. That is the equivalent of the first four millimetres of a journey to Proxima Centauri, the closest star to earth apart from the sun.

Comment 4) Thank you very much. I have wondered about this for a long time.

Comment 5) Go yotta! You know where to find beta testers...

 

[Vixe]

How do you guys feel about somebody exploiting the rating system?
http://www.chessconquest.com/index.php?p=stats&playerid=4956
http://www.chessconquest.com/index.php?p=stats&playerid=5491
(Direct your attention to the game history.)

Edit:

Comment 3) A quintillionbyte is technically called an exabyte. For reference, an exabyte is a billion gigabyte. It is actually technically possible to own such a drive: on Wikipedia (http://en.wikipedia.org/wiki/Exabyte - a very interesting article) it says that "Oracle Corporation claimed the first Exabyte tape library with the SL8500 and the T10000C tape drive in January 2011." Does this mean that we are close to solving the problem of storing the information for chess? Hardly, but we have made a (pathetically modest) start though. We are exactly one quintillionth of the way, assuming we store one byte for every chess move. That is the equivalent of the first four millimetres of a journey to Proxima Centauri, the closest star to earth apart from the sun.

On the mention of exabytes, I'd like to add that IBM is planning on building an exascale supercomputer by 2024 for the Square Kilometer Array (SKA) project, in order to handle the astronomical quantities of data that needs to be processed. It will be 100 times faster than today's most powerful supercomputer.

 

[tallfred]

How do you guys feel about somebody exploiting the rating system?
http://www.chessconquest.com/index.php?p=stats&playerid=4956
http://www.chessconquest.com/index.php?p=stats&playerid=5491
(Direct your attention to the game history.)

Edit:


On the mention of exabytes, I'd like to add that IBM is planning on building an exascale supercomputer by 2024 for the Square Kilometer Array (SKA) project, in order to handle the astronomical quantities of data that needs to be processed. It will be 100 times faster than today's most powerful supercomputer.

I agree, there is very real evidence of abuse of the rating system. I suspect that there is nothing we can do about it though because AFAIK we have no policy on what constitutes acceptable and unacceptable behaviour. There can therefore be no way to sanction individuals who cross the line. Yotta, can we open the discussion? Is there an acceptable behaviour policy we can copy from somewhere else?

With regard to IBM, I notice that they love to blow their own trumpet about what they will do one day, but are less than transparent about current weaknesses. For example, did you know that they are currently only number 10 on the list of fastest computers, less than 10.4% of the speed of the fastest computer. You will have to go a long way to find mention of the world's fastest computer at the moment. It is not only faster than number 2, it is faster than numbers 2,3,4,5,6,7 and 8 combined. It's upgrade alone was faster than number 2. For more info, see http://www.top500.org/lists/2011/11

 

[Vuade]

conditional move always avail

Sorry - have not been active on the forum for a while. Just came across this

You make a very strong point TallFred, and you convinced me.

If no-one else objects or gives a counter argument against the link being permanently available, I'll update our task-list to make the change that MarkFace suggested to version 1.

I object.

<my 2c>

Even though I would love this feature, I believe it would give me (and my opponent should he/she wish to make use of it) an unfair advantage.

I agree with Tallfred that this is solving a problem that does not occur in Face to Face games. But all you need to solve the time problem, is the conditional move feature available AFTER you have made your move.

In my mind, one of the things that makes one a 'good' chess player, is the ability to look ahead in your mind. If you allow people to look ahead physically on the board (i.e allow auto move BEFORE you have made your move) then you give people the assistance/tools they would not have in real life chess. Imagine you are allowed to take an additional chess board into a live competition, to mirror the game you are currently playing. Then, before you make a move on the real board, you first 'test' it out on the extra board to see what would happen. In my case, I believe I would be playing a better game of chess this way - because as a human, I normally only think about 2 or 3 moves in advance. Having a 'test' board would allow me to look further ahead, enabling me to possibly see mistakes/situations I would not normally have seen in my mind's eye, which in turn would enable me to play a better game.

So I guess you could argue people can do this in any case. Playing in the private of their homes - who says they are not doing this already? (btw this does not apply to people (like me) who make most of their moves on the phone while 'on the move')

In my mind the decision boils down to this - do you want chessconquest to be a platform where you can play traditional chess? Or to be something new that gives people aid and tools they can not have in real life, so they can play better chess?

I prefer the former - but that is my opinion.

And just to clarify - the 'unfair advantage' I mentioned earlier is not pertaining to the two opponents (because both players have the option to make use of the facility), but rather to the game of chess. In my mind, if both players make use of the feature, you'll end up with a better chess game than you would have otherwise.

The golden middle way to implement it on the website most probably is the way you have proposed already - allow the challenger to dictate whether this functionality should be available or not.

</my 2c>

 

[Cicero]

How do you guys feel about somebody exploiting the rating system?
http://www.chessconquest.com/index.php?p=stats&playerid=4956
http://www.chessconquest.com/index.php?p=stats&playerid=5491
(Direct your attention to the game history.)

Jeez! Go look at the games! Terrible boosting. I mean, if you're gonna boost, at least try make it look realistic. one of the resignation games was 1 pawn move each....wtf!

http://www.chessconquest.com/index.php?p=viewgame&gameid=19451

 

[Vixe]

I stand by Vuade's objection. There is no need for conditional moves while it is your turn.

 

[Cicero]

I stand by Vuade's objection. There is no need for conditional moves while it is your turn.

But if you play 10 games plus, which most of us do. And you dont make a move for 2 days or so, you quickly forget what your thought process was. I just dont see the harm.

 

[yotta]

How do you guys feel about somebody exploiting the rating system?

I agree, there is very real evidence of abuse of the rating system. I suspect that there is nothing we can do about it though because AFAIK we have no policy on what constitutes acceptable and unacceptable behaviour. There can therefore be no way to sanction individuals who cross the line. Yotta, can we open the discussion? Is there an acceptable behaviour policy we can copy from somewhere else?

Jeez! Go look at the games! Terrible boosting. I mean, if you're gonna boost, at least try make it look realistic. one of the resignation games was 1 pawn move each....wtf!

The problem with players exploiting the rating system has been pointed out numerous times, and a policy discussion would definitely be a great help.

I think Vaude also posted his "rating-experiment" results in this forum, illustrating how easy it is to exploit the system.

This type of "behaviour" is definitely one of the more challenging aspects of creating an enjoyable and "fair" chess community website.

We were thinking of perhaps adding a limit to the number of games a player is allowed to play against a single player. For example, for each 3 games you complete against one player you need to complete at least one game against another player. Also perhaps a rule that a timeout can only be claimed after at least 2 moves (instead of just one move on either side).

Onto the topic about having the conditional move feature permanently available...

I object.

<my 2c>

Even though I would love this feature, I believe it would give me (and my opponent should he/she wish to make use of it) an unfair advantage.

I stand by Vuade's objection. There is no need for conditional moves while it is your turn.

But if you play 10 games plus, which most of us do. And you dont make a move for 2 days or so, you quickly forget what your thought process was. I just dont see the harm.

It seems like everyone is divided on the conditional move feature.

Perhaps I'll create a short online survey for the players at ChessConquest , and put this one to a vote?

1. Conditional moves only available during opponents move.
2. Conditional moves permanently available.
3. Conditional moves availability decided before game commences.

 

[ViperGTI]

The problem with players exploiting the rating system has been pointed out numerous times, and a policy discussion would definitely be a great help.

...

We were thinking of perhaps adding a limit to the number of games a player is allowed to play against a single player. For example, for each 3 games you complete against one player you need to complete at least one game against another player. Also perhaps a rule that a timeout can only be claimed after at least 2 moves (instead of just one move on either side).

What if maybe you implement a rule that you can only challenge players with a higher rating than yourself? I suppose there will still be room for the obvious exploitation (like the sample shown above), but it should go a long way to prevent some of these Romanian guys picking on the lower players just for the purpose of a high rating (as in Vuade's experiment).
The 3 games against a single player rule as mentioned above (if all three games are all consecutive wins / losses), could perhaps work for trying to prevent the situation as pointed out by markface.

Onto the topic about having the conditional move feature permanently available...

It seems like everyone is divided on the conditional move feature.

Perhaps I'll create a short online survey for the players at ChessConquest , and put this one to a vote?

1. Conditional moves only available during opponents move.
2. Conditional moves permanently available.
3. Conditional moves availability decided before game commences.

I have no issue with it being permanently available or not as long as it is at minimum available as it currently is (1). It is a great time-saver especially in situations where the other player is cornered and only certain logical moves make sense... it can take things pretty far.