Champions Showdown: Improving the format

Advertising

Books, boards, sets: Chess Niggemann

Macauley Peterson Macauley served as the Editor in Chief of ChessBase News from July 2017 to March 2020. He is the producer of The Full English Breakfast chess podcast, and was an Associate Producer of the 2016 feature documentary, Magnus.

Discuss

Rules for reader comments

celeje 10/6/2018 02:16

@Petrarlsen:

I also hope lajosarpad will come back to check, read & comment.
I will keep reading & commenting.

Petrarlsen 10/5/2018 10:37

@ lajosarpad (2/2):

- About a castling in which the King goes toward the center.

My impression was that it was rather obvious that the King is less safe in the center that on the flank. After reading your last posts, I tried to convert this impression into objective arguments, which are :

1) As a preamble, my reasoning: If the King is in the center: a) Most of the time, it isn't possible to keep a "pawns' shield" in the center; at least some of the central (in the broader sense of the word - c to f) pawns must move in the opening phase to control the center in one way or another, the result being that a King in generally poorly protected in the center. b) When the King is in the center it can be attacked from every side; attacking possibilities are more limited when it is in a corner. (Obviously, there are many exceptions, but I think this holds most of the time.)

2) A strong GM's (Yasser Seirawan) opinion: "Castling allows a player to move his King out of the center (the main theater of action in the opening) to the flank, where the King can be protected by pawns." (In the glossary of Y. Seirawan's "Winning Chess" series.)

3) In quite a number of opening's lines featuring queenside castling, the player who castles queenside puts (just after castling or later in the opening) his King on b1 or b8: what I deduce from this is that, very probably, the safest square for a King is the b1/b8 square. And - generally - when kingside castling is used, the King nearly never moves until the end of the opening; I think this shows that there is quite a difference between b1/b8 and c1/c8: the safest square is b1/b8 - the square which is nearer to the flank.

4) When castling isn't possible anymore (generally because the King or the Rooks have moved), frequently, the players "castle by hand"; in my opinion, this tend also to show that the King is safer on the flank ; otherwise, the players wouldn't use quite a number of moves to get their Kings on the flank in these cases. (Secondarily, the fact that, in some games, the players "castle by hand" also tend to show that castling isn't such an "haphazard move" ; otherwise, the players wouldn't spend a whole series of move to obtain the same result when castling itself insn't possible.)

- As for the positions for which castling is possible from move one, my opinion isn't that this is a complete catastrophe; only that, compared with a position for which castling isn't immediately possible, it is inferior for the reasons I stated previously. And, in my opinion, if it is clearly inferior to other positions, it means that such starting positions would have to be discarded.

A supplementary element : When you say : "I do not think that having a very difficult castling possibility would have any effect on a position where a good castle can be done at the first move. People will simply not care about the other castling possibility.", I think that this also constitutes an argument against positions for which castling is possible from move one: in such positions, chess players will use only one of the two castling possibilities; considering once more this given starting position as if it was the only starting position of the game, it would mean that two different castlings would have been created, knowing perfectly well all the time that one of these two castlings will never be used by the players; it isn't coherent: there should either be ONE castling, or if there are TWO castlings, then it should be reasonably conceivable that the two castlings can be used in practice! Otherwise, this would give "one-legged games", theoretically featuring two different possible castlings, but necessarily ignoring completely, in practice, one of these two castlings!

Petrarlsen 10/5/2018 10:37

@ lajosarpad (1/2):

I am really sorry... I couldn't find time to answer you until now... I hope that, some time, you will come back to this page...

- About the question regarding the possible validation of the starting positions as if each was the only starting position, I am not completely sure to have understood perfectly what you meant.

As for myself, what I meant was in particular:

1) What a chess player creates are chess games - a writer creates books, a composer musical works, a painter paintings, a sculptor sculptures... and a chess player chess games. So, if a top-player's game (and more so for a particularly important game) is played with a inferior starting position, it is quite a pity, as the game will inevitably be affected by this. It is easy to think about a Chess960's starting position's defects: "This isn't so much of a problem; this position will only occur in one game out of 960." But if this position occurs in an important top-level game, this will affect this game, and I think that, if each starting position was taken into account individually, as if there wasn't any other starting position, this would avoid this problem.

2) I consider that to take each starting position as if it was the only starting position is some sort of a "condition": it is necessary to avoid distortions, when considering a starting position (the "one out of 960" effect...), but this doesn't mean that it is sufficient; it would indeed be necessary to define a minimal level of perfection for a starting position.

- About castling, I don't agree that castling can only be explained historically, and that it would appear as being as strange as petrarlsening, if we weren't accustomed to it.

In my opinion, castling has a conceptual explanation (and not only a historical explanation): if the concept is to implement a special move improving the King's safety, castling is, in my opinion, a quite good answer to this concept. Perhaps that, if we wouldn't know castling, it would appear to us as a strange move if it was presented to us without any further explanation, but I think that, if someone said : "I tried to devise a special move, having for an objective to improve the King's safety", castling wouldn't appear to us as something quite strange. This is the difference with petrarlsening, which doesn't seems to answer to ANY understandable concept.

Petrarlsen 10/2/2018 03:04

@ lajosarpad : Sorry I couldn't find time to answer you for the moment... I think it will be possible for me tomorrow...

@ celeje : Yes, it will be interesting to see your ideas about castling !

celeje 9/30/2018 08:47

@ Petrarlsen:

I think I should explain my so-far unwritten thoughts on castling "logical coherence" etc. in my next posts (instead of posting about other discussion points), because until you see them you'll assume they don't exist & I'm just being [put negative words here]. I was going to do this in the last article's comments. When this article appeared, I had to remember everything again.

You may not agree at all with what I'm gonna say, but that's okay. They can be food for thought even if you disagree with them.

@lajosarpad:

I agree the distance to a chess960 castled position in the absence of castling depends on the metric you use.

Petrarlsen 9/30/2018 04:43

@ lajosarpad : I haven't the time for the moment to answer your last posts, but I very much hope that you are right, when you say : "Anyway, I would be VERY surprised if after the first 15 tournaments of chess960 we would not have a very very long list of sensible suggestions of improvement." I find Chess960 to be extremely interesting, but much too much "unfinished" for my taste, and I would be extremely pleased if a really good version of it would appear at one point!

lajosarpad 9/30/2018 02:24

@Petrarlsen

With your reasoning about castling, as I described my disagreement is theoretical skepticism, that is, I do not actually believe that your statement is wrong, my position is that the king going towards the centre by itself does not make it less safe, or, to be more precise, I am not yet convinced about this pattern. On the other hand, I think randomity does not compensate for imbalance. If we accept randomity into the start of the game, then a match score will not necessarily determine who the better player is and chess would have the same defect as poker has. If one has bad cards, one has a defect caused by chance. And if one loses, one might argue about having bad cards, even if those cards were not so bad. The scientific nature of the game would disappear.

Interesting point about the starting positions where castling can instantly happen. My point of view is actually a provable fact. The value of instant castling will eventually be approached by machines, generating a consensus and if it is a much better move than anything else, then all games in the given position will start in the exact same way, making opening theory about the position redundant. I understand your point, however, the possibility that the games will always start with the same move would lead to a ridiculous situation. Hence I agree with your point:

"It would be as if, for example, castling was possible from move one in traditional chess ; I think it would be rather an impoverishment in comparison with traditional chess in its present form... "

In short, in the given position castling either improves the position or not. If it improves the position, then the first move is banal. If it does not improve the position, then it is self defeating. There is a third possibility, namely that it does not improve the position as the first move, but it needs to be prepared to be effective. While we cannot exclude this case, it does not seem to be too probable.

@Celeje

"here's a balance here that no one talks about. When castling is "unearned", the result of castling is smallest. The K & R move 1 square each."

I do not think we know the real result. If we look at the table as a grid where Manhattan distance is applied (https://xlinux.nist.gov/dads/HTML/manhattanDistance.html), then a king on f1 being swapped with a rook on g1 is a very small change indeed. But if we define the result in a term like

gain / cost

then cost is almost 0 and gain is a pretty large number if the king will be safer indeed. In the other hand, in classical chess you have to move at least a pawn, a knight and a bishop before you castle. That's four tempo of preparation and since the other side is fighting you, you might need to make other moves to prepare castling or prepare the preparation of castling. I do not think that having a very difficult castling possibility would have any effect on a position where a good castle can be done at the first move. People will simply not care about the other castling possibility.

I am more and more convinced that chess960 deserves a chance to be played as a separate sport and if it is successful, then it will take roots. If not, it will disappear eventually. Anyway, I would be VERY surprised if after the first 15 tournaments of chess960 we would not have a very very long list of sensible suggestions of improvement.

lajosarpad 9/30/2018 01:45

@Petrarlsen

Step 1: I understand your reasoning and I quite like the idea that all possible starting positions must be perfect. It is difficult to define what perfect is, but the introduction of a standard for what "perfect" is in terms of a starting position is certainly possible if there is a consensus that a standard needs to be reached. Also, I am not convinced that the validation of a starting position would be ideal if we validate the starting position as if it was the only possible starting position. Whether the validation is correct under such conditions highly depends on what "perfect" is in our case. If the hypothetically accepted perfect is possible to be accurately validated without taking into account other possible starting positions, then you are right, of course. I tend to agree with you, but I am not convinced. The thought experiment which makes me thinking that you are probably right is the case when A and B are playing a single game and after randomization they get a position where A has the advantage from the start. This case is unacceptable, but knowing these pitfalls a tournament system could avoid them even if imbalanced starting positions are kept, but the frustration of hobby players cannot be resolved by FIDE. In short, I tend to agree with you about step 1, at least I understand the rationale behind your arguments.

Step 2: Petrarlsening, that's a good one!

Before reacting to step 2, let me give you an intermezzo about castling, so my approach to the rule will be more understandable. Castling in classical chess is inconsistent with other rules and it is a remedy to fast attacking possibilities. En passant is another example of inconsistent rule. These have their historical reasons, but, if castling would not exist and somebody would bring up the idea of castling, I think it would be as strange for us as petrarlsening. I think one of the most important factors we are so fond of castling is that we are used to it and accepted its existence when we learned the rules of the game, before we developed a chess taste.

Now, since castling is a construct we accept due to historical and cultural reasons, we either abolish castling (which will not happen) or use it as a standard for castling in chess960. According to these terms, I think its goal is king safety. A possible addition is rook development, but I am not convinced in this, since, if the rule of castling would not exist, rooks would develop on another route, possibly via moving a pawn on the a or h file. The reason we are not moving these pawns to the fourth or fifth rank to develop the rooks quickly (besides bishop and queen diagonals) is that castling exists and we do not want to break the future pawn barrier of the king. If the rule of castling would not exist, then the route of the king to the position it lands at after castling would be very long, especially due to the own pieces. This is why I think castling in chess960 should keep the principle of king safety improvement.

Step 3

I understand the reasoning and it is quite possible that in all the possible starting positions where at the end of castling the king would be closer to the center than before, the king would be less safe by default. If that is the case, then your thesis is correct. If this thesis is true in most cases, but there are a few exceptions, then your thesis describes a pattern, but there are exceptions. My position is that the king should be safer after castling than before in most cases of each position and here we agree. I, however, do not see being closer to the centre after castling by itself would decrease the safety of the king. If this rule is correct, then you are right. The question is interesting, I think research is needed.

Petrarlsen 9/30/2018 12:33

@ celeje:

- I agree that, some people react just instinctively to Chess960, without any real arguments, but, in my opinion, you do rather the same sort of thing when you answer me by saying : "(...) it's just a familiarity thing.".

The people who criticize Chess960 instinctively don't have any real argument, so, more or less, they say that Chess960 is bad because they don't like it. But, I think it is more or less the same for you ; you answer to me: "(...) it's just a familiarity thing", because you haven't any other answer, and that, by saying this, you can evacuate any criticism regarding the incoherent and farcical character of certain traits of Chess960. As I said, you could also answer "(...) it's just a familiarity thing" about an elephant dancing a waltz, but it wouldn't be true; it wouldn't be a question of familiarity: an elephant dancing the waltz will NEVER be something serious, even if you see one doing it every day of your life. And, in my opinion, it is the same for many things in Chess960 : it isn't a question of familiarity; some things are just clearly suboptimal, for different reasons I (and others) have developed previously, without it being a question of being accustomed to them or not. So I consider that, when you answer me: "(...) it's just a familiarity thing.", it is not a good answer to what I said...

- "What info we saw indicated that there were different rules in different places & they had to change and adopt one set if they wanted to play with outsiders. No doubt the rules in places where there were more strong players had a bigger chance of adoption."

It seems also just simply natural that the chosen rules wouldn't have been the most "outlandish" ; it is much easier to agree on something "normal" that on something completely "outlandish".

It is in a way some sort of a "natural selection" ; what remains is the most "normal" ; the outlandish, bizarre, and incoherent rules are eliminated. The fact is, anyway, that it it precisely the "normal" sort of castling that remains, and that the other ones disappeared... You will say that we consider this castling as being "normal" because we are accustomed to it, but I will not agree, not more that for the elephant dancing the waltz!

And it is the way things work most of the time in history; things are perfected through years, decades, and centuries; for example, the Chinese porcelain didn't attain immediately the level of perfection it had at the time of the Ming dynasty; it took centuries and centuries of evolution to obtain such a result. And I think that very rare are the things that attain immediately (or nearly immediately) their definitive state.

- "If it seemed so bizarre, it would never have been a rule in the first place."

Why??? A person who devises a game isn't infallible! He (or she...) can have strange ideas (and even very strange ideas) as anyone else! And it isn't because this game has a certain success that everyone agrees about everything in this game! It is quite possible that many people play this game, and, at the same time, that they don't like some things in it, the result being that at the first occcasion for them to "push" for a change in the rules, they do it!

- "When castling is "unearned", the result of castling is smallest."

Yes, perhaps ; but this doesn't make for something well-balanced ; a well-balanced castling, in my opinion, is one for which you must work, and that gives a significant result. To have an castling in embryo form doesn't appear to me to be quite satisfying!

- And you didn't answer my last question of my last post to you : "On that subject, how would you explain that Chess960 - which hasn't even existed for a single quarter of a century - is quite generally considered by its proponents to be fixed for good in its smallest details, while traditional chess evolved through many centuries before reaching its modern form?"

celeje 9/29/2018 11:29

@ Petrarlsen: (continued)

Another thing about your "castling must be earned (... you must "work for it")":

There's a balance here that no one talks about. When castling is "unearned", the result of castling is smallest. The K & R move 1 square each. I think the "opposite" gets the most complaints. The "opposite" is where the K flies from b1 to g1. But that largest result requires the most work to "earn". So it sort of fits well.

celeje 9/29/2018 09:59

@ Petrarlsen:

I agree that if something is unfamiliar that does not mean it's automatically good. It doesn't mean it's automatically bad either. I just mean that IN THIS CASE (Chess960), most people have not thought out carefully whether (parts of) Chess960 are good or bad. They just have kneejerk reactions. We see this because they don't come up with arguments. They just make one-line claims, some of which are plain wrong. We saw that in the first article and comments.

We talked about history with jacob woge and others. What info we saw indicated that there were different rules in different places & they had to change and adopt one set if they wanted to play with outsiders. No doubt the rules in places where there were more strong players had a bigger chance of adoption. We saw they also changed rules because they wanted to speed up play. Nowhere did we see they wanted rule changes because they thought something was conceptually "bizarre". If it seemed so bizarre, it would never have been a rule in the first place.

I am not saying we have to follow history either.

And I don't want to go backwards in history and adopt 1.Kd3!! It seems bizarre TO ME because I live NOW, not then. That does not mean it MUST be okay either.

Petrarlsen 9/29/2018 05:53

@ celeje:

- "(...) it's just a familiarity thing."

The fact to know if something is serious, farcical, bizarre, etc., isn't only a familiarity thing: for example, in another domain, if a film is a comedy, no one will ever think it is a tragedy, or else there is a big problem somewhere... And if I would see an elephant dancing a waltz, I would not consider this as something serious, and I don't think that anyone would think differently. This, while if it was for example a purebred Arabian horse dancing the same waltz, even if it would still be an animal, it would seem much less of a farce than for an elephant.

Indeed, I think that sentences like this sentence ("It's just a familiarity thing.") is some sort of a refuge ; Chess960's proponent use them to avoid seeing the stark reality, which is that some Chess960's starting positions have real problems. If you consider, for example, that nothing can be absurd per se, it permits to consider that even the most glaring absurdity isn't really an absurdity. It is very convenient, but I think that it is much more a technique (unconscious, most of the time, in my opinion) to avoid the real problems than a real and serious argument.

And when you say : "Remember historically it came from the King leap e.g. 1. Kd3!!! People obviously had no problem with the King leap at the time.", I think that, precisely, as this was rather bizarre, progressively, other solutions, more normal, were found. And this is exactly what seems to be lacking in the "Chess960 universe": "when Chess was young", ameliorations were found for centuries and centuries, while, for Chess960, most of its proponents consider it as being set forever in the present state.

On that subject, how would you explain that Chess960 - which hasn't even existed for a single quarter of a century - is quite generally considered by its proponents to be fixed for good in its smallest details, while traditional chess evolved through many centuries before reaching its modern form?

celeje 9/29/2018 10:22

@ Petrarlsen

Petrarlsen: "I find that this "first move castling" business would be considered as being too farcical if this position was the only starting position..."

Possibly. It's hard to tell.

But what the chess public thinks changes and depends on what they're used to at the time. Again it's just a familiarity thing.
Remember historically it came from the King leap e.g. 1. Kd3!!! People obviously had no problem with the King leap at the time.

Petrarlsen 9/29/2018 02:19

@ celeje : I agree that castling wasn't necessarily the only first move possible. But, even taking this into account, I find that this "first move castling" business would be considered as being too farcical if this position was the only starting position of a "serious" chess variant; it is only accepted because such positions are only encountered from time to time, in Chess960. But it isn't at all because a position isn't the ONLY starting position that the level of expectations must be lower...

celeje 9/29/2018 02:10

@ Petrarlsen:

I think an important question is: Can we be certain that castling was the best first move in the Carlsen-Nakamura example? Right now, I don't think we can be.

Petrarlsen 9/29/2018 12:34

@ celeje :

"I don't see why fundamentally a player must "work for it"."

Besides what I've said on this subject in my last post, it feels as if the real starting position was after the castling, if the two players castle immediately (as did, if I remember well, Carlsen and Nakamura in one of the games of their Chess960 match). In this case, after move one, you have something that completely resembles a starting position; just better, because the King's position is safer.

I think that, if such a starting position was the only starting position of a chess variant, people could say : "What does it mean to implement a castling, if you can do it immediately? Either you arrange things in such a manner that it isn't immediately possible, or then, be logical with yourself, suppress the castling altogether, and take directly for a starting position the position after castling." In my opinion, it rather feels like a farce, this "first move castling"...

The problem is always that, as there are 960 starting positions in Chess960 and that they haven't been directly invented by someone who could be criticized for that (as they are randomly generated with some restrictions due to a series of given criteria), the level of exigence is MUCH lower than if some person had directly designed one of these positions as the one and only starting position of a chess variant. What, in my opinion, wouldn't be admitted for one moment in this last hypothesis is considered as quite normal when considering one of the Chess960's starting positions...

This is not an enormous defect, but, in my opinion, it is nonetheless one of these little thing that leave the feeling that Chess960 is somewhat "unfinished", is some sort of a "draft" ; a very interesting draft, but a draft nonetheless...

celeje 9/28/2018 11:45

@Petrarlsen:

Petrarlsen (to lajosarpad): "One little question : what do you think about the starting positions in which castling is possible from move one? ... My opinion is rather that castling must be earned (that you must "work for it"), and that for it to be possible from move one isn't logical."

I don't see why fundamentally a player must "work for it".

My first reaction is that the only significance of being able to castle so soon is whether it makes it harder to attack the K than usual. But a successful early attack on an uncastled K doesn't happen at high levels anyway. GMs defend too well for that.

And now I'm not sure it matters if there's a slightly different balance between attack & defence in the opening, anyway. Maybe it's good that players have to play slightly differently in different positions, just as they have to if their opponents choose different openings.

Petrarlsen 9/28/2018 08:54

@ lajosarpad :

One little question : what do you think about the starting positions in which castling is possible from move one? (i.e. positions in which the King is alongside one of the Rooks)

My opinion is rather that castling must be earned (that you must "work for it"), and that for it to be possible from move one isn't logical, this having for a consequence that I would rather suppress all the positions featuring a King alongside one of the Rooks.

To elaborate a little : if we consider that castling answers to the concept "King's safety", and if castling is possible from move one, it means that the players can simply put their Kings into safety from move one. It seems to me a little too simple ; to say it this way, in such positions, you really haven't to sweat much to put your King into safety! And, once more, if you consider each starting position as if it was the only possible starting position of the game, I wouldn't find fully satisfying to be able to implement the "safety measure" that castling represents so easily as that... It would be as if, for example, castling was possible from move one in traditional chess ; I think it would be rather an impoverishment in comparison with traditional chess in its present form...

Petrarlsen 9/28/2018 11:02

@ celeje :

"Theoretically, it is a plus to me, as we discussed before."

- "(...) as we discussed before."

I didn't remember we already discussed it... (there was so much discussion on Chess960 that I've lost track of the details!...)

- "Theoretically, it is a plus to me (...)"

Why? Have they prepared or not, theory will have a great importance in this game ; it will favor, either the player who plays regularly traditional chess besides Chess960, either the player with the best long-term memory, either the very thorough player who will have revised traditional chess' theory specifically for this tournament, etc.... and all this doesn't seem to me to be in the spirit of Chess960: the specific interest of Chess960 is precisely to allow to play the game with a minimal part of opening theory...

I would find it more logical to rule out Position 518, but then, it means the idea of suppressing some positions is accepted, and then, why wouldn't other positions be suppressed for other reasons? (But it would be possible to say that it could theoretically be possible to suppress other positions - simply, there isn't any problem with any other position ; I wouldn't agree with the fact that there wouldn't be any problem with any other position, but it would still be possible, and I would find this more logical...)

celeje 9/28/2018 09:08

@Petrarlsen:
About the small point...
Some tournaments have ruled out 518 and made it Chess959. Even if they don't, a 1/960 chance is rather small & the practical difference is that the players would not have prepared for that.

(Theoretically, it is a plus to me, as we discussed before.)

Petrarlsen 9/28/2018 02:42

One small point : I think that, contrary to what many Chess960's proponents think, the fact that Position 518 in Chess960 corresponds to traditional chess is something negative, and not at all something positive. This for the reason that Chess960's main interest is to allow to play a game for which opening theory is either suppressed, or reduced to an absolutely minimal level ; if a Chess960's game is played with Position 518, the main interest of Chess960 will thus be lost, and I think this should be avoided (you don't play a Chess960's tournament to play a traditional game of chess!).

Petrarlsen 9/28/2018 02:27

@ lajosarpad (2/2) :

* Step Three : Each detail of a special move must answer to the corresponding concept. (And this is at this moment that I disagree with you about a castling for which the King would go in the direction of the centre of the board.)

For example, castling has two elements : a King's displacement, and a Rook's displacement. For each starting position, it must be checked that these two elements really concur to the corresponding concept (for example : King's safety). And, in my opinion, in the Chess960's starting positions in which, for at least one of the two castlings, the King will move towards the center when castling (as, for example, in the starting position corresponding to the Wikipedia example that was discussed previously in this and the previous article about Chess960 - https://en.wikipedia.org/wiki/Chess960#Castling_rules), the element "King's displacement" doesn't concur to the concept "King's safety".

To illustrate this, we can imagine the following scenario :

A is presenting a proposed starting position for a chess variant to B ; A explains that, in this variant, castling will correspond (more or less as in traditional chess) to the "King's safety" idea, and he further explains that, in this variant, the King will move towards the center when castling, but that this will nonetheless quite correspond to the "King's safety" idea, because of the displacement of the Rook.

B can answer that, yes, globally, this castling is in accordance with the "King's safety" concept, but that, when taking separately its two elements, the Rook's displacement element concurs to the "King's safety" concept, but not the King's displacement, so that this castling haven't the necessary quality level to be implemented in this form.

It is a question of coherence : it seems to me to stand to reason that, if you are wanting to devise a special move answering to a specific concept, logic demands that all the elements of this special move will concur to this concept. And if one element doesn't concur to the general concept, this special move, devised as it is, is suboptimal, and thus, cannot be kept in this form.

And I think that the sole reason for which, in Chess960, such castlings as the castling figuring in the Wikipedia example are accepted is only that no one takes each Chess960 starting position separately, to check if it is satisfying (according to precise criteria...) or not ; the general idea seems to be that, as these positions are randomly generated, they are necessarily good. (Or perhaps, that there isn't any need for them to be "good" ; as they are randomly generated and that this means that, globally, they are fair to the players - they can give an advantage to a player in one game, but not on the long run, on a great number of games, as there is no human choice involved -, there is no need to ask for more...)

Would you agree or disagree with all this, and if you disagree, with what would you disagree ?

Petrarlsen 9/28/2018 02:24

@ lajosarpad (1/2) :

About castling :

I already explained what I will now explain in previous posts, under the previous article about Chess960 (https://en.chessbase.com/post/the-problem-with-chess960), but I rather think that you haven't followed the comments under this article ; if you know already my "theories" about Chess960's castling, you can quite well skip all the following, as you shouldn't normally learn much there about my arguments about Chess960's castling !...

I will explain my reasoning step by step, so that you can point out more easily with which step you disagree (...or perhaps, if I am sufficiently convincing, you will not disagree anymore !!...).

* Step One : I consider that, for "validation purposes", each Chess960's starting position would have to be treated as if it was the only starting position of the given game. (And it would be the same for another chess variant with multiple possible starting positions.) To elaborate a little : as each starting position can be the starting position of an important game between great players, if a starting position has a significant degree of imperfection, it will potentially affect the quality (as a "chess work of art", so to say) of some important games, and I don't find this satisfying, so I consider that the requirements' level for a starting position to be usable must be exactly the same as if this starting position would be the only one for this variant.

* Step Two : For me, for a special move (as castling) to be implemented, this special move must answer to a concept ; in my opinion, an arbitrary and haphazard special move shouldn't be implemented, as this inclusion would not be coherent with the rest of the game. (I gave an imaginary example that I called "petrarlsening" : as I put it in an old post, petrarlsening "would consist of exchanging the positions of Rooks with the positions of Knights (you put the Rooks on the squares previously occupied by Knights, and the Knights on the squares previously occupied by the Rooks)." I supposed that if this petrarlsening would be implemented, it would be used from time to time - and perhaps frequently - by the players, but, as it is completely arbitrary and haphazard - and doesn't answer to any concept -, it shouldn't in fact be implemented.)

Until now, for castling, I followed the concept described by GM Seirawan (a double idea : King's safety + Rook's development). But, in fact, it can probably indeed be reduced to the King's safety aspect (at least for Chess960), in accordance with what you proposed yesterday. I think that another concept, completely new, could also be invented specially for Chess960, but it would be necessary for this concept to be coherent with castling in all Chess960's starting positions.

(I know that we very probably globally agree on this "Step Two", but I put it nonetheless because I think it is an important element in this reasoning.)

celeje 9/27/2018 11:29

@lajosarpad:

@lajosarpad: "You might be right about your guess that all 960 positions are theoretical draws. Or not. I cannot determine it at this point."

Yup, no one can determine at this point that even the traditional starting position is a draw, though everyone believes it. And though everyone believes it, that doesn't influence whether they play it.

I'd say the same will apply to the other 959.

lajosarpad 9/27/2018 08:28

@celeje

Those 1000 moves are in the case of perfect play, but not impossible to memorize at all. Some people are memorizing much larger samples. You might be right about your guess that all 960 positions are theoretical draws. Or not. I cannot determine it at this point.

celeje 9/27/2018 02:18

@lajosarpad:

@lajosarpad: "If a problem is difficult to prove, then in practice there is no problem, until the problem is found. If there is a forced win for white in a given starting position and it will be found in 2050, then there will be no problems playing that position before 2050, but the position will not really be playable for black after 2050. If there is a problem with a given position, then there are theoretical reasons for it, regardless of whether we are already aware of them or not."

Yes, agreed... except not necessarily unplayable in the example after 2050. (e.g. could be 1000-move TB win.) I say that means there's no problem playing any of the 960 positions...
And I'd guess all 960 positions are theoretical draws.

lajosarpad 9/27/2018 02:05

@Petrarlsen

"I am sorry, but I didn't understand what you meant by that (the beginning : "I think it is fair to put the fairness of the positions"). Could you elaborate a little ? "

I was unclear, sorry about that. I have meant that a large part of the proponents for chess960 will find all chess960 positions to be fair, regardless of their attributes. I think the real power behind the support of chess960 is the frustration of theoretical defeats. Of course, there are people like Celeje who have very reasonable arguments, at least, so far, but a large amount of people are relying more on their feelings and they see chess960 as a game without openings, where they can avoid the frustration of theoretical defeats.

I think castling has safety as its purpose. If the king is safer after castling, then castling makes sense, even if the king is closer to the center after castling than before. I therefore do not share your opinion that castling is illogical if the king is closer to the center than before as a result, but if from the starting position it is foreseeable that the safety of the king will drop if the player is castling (assuming a majority of normal scenarios), then the rule is illogical. To put it into different terms: if players will have reasons to not want to castle, while in chess a very frequent strategic goal is the castle, then the rules of castling have some problems.

@Celeje

Interesting thought about US point of view.

If a problem is difficult to prove, then in practice there is no problem, until the problem is found. If there is a forced win for white in a given starting position and it will be found in 2050, then there will be no problems playing that position before 2050, but the position will not really be playable for black after 2050. If there is a problem with a given position, then there are theoretical reasons for it, regardless of whether we are already aware of them or not.

celeje 9/26/2018 06:16

@ lajosarpad:

lajosarpad: "I agree with the way you validate, but I doubt we have enough knowledge to determine whether there are many positions which largely contribute to the result or not. I understand it is hard for you to find a position which highly contributes to a given result, but I believe you would face considerably higher difficulties in proving about all chess960 positions that they are all playable, balanced and not drawish. "

ccrl computer Chess960 results suggest no problem of semi-forced results. When the Carlsen-Nakamura match was on, someone showed supercomputer analysis to high depth, which suggested the same.

AND... if it's difficult to prove a problem, then in practice there will be no problem. And if there is a problem with some positions, theoretically it's interesting to know why, because there's no reason there should be.

Petrarlsen 9/26/2018 06:24

@ lajosarpad (2/2) :

- "There were always powerful belief systems, Katolocisism in medieval Europe, Islam in the Islamic world, Liberalism in the Jacobine era, Communism in the Soviet Union and China, National Socialism in the III. Reich and, for a lot of chess players, the unquestionability of chess960."

I think this evolved from religious beliefs. For religion, it is logical to follow completely a given religion, if you consider that this religion has been revealed by a god - the only real question is to know if REALLY a god is the source of this given religion, which isn't necessarily an easy question to answer, but, clearly, if the answer is "yes", it is logical to follow this religion, and if it is "no", then this given religion is simply of no value ; either an imposture or a series of erroneous beliefs, and it would be illogical to follow it. And I think that, later (from, more or less, the 18th and 19th centuries), the thought mechanisms linked with religious beliefs were transfered to non-religious beliefs (as, for example, as you said, Communism or National Socialism, for quite a number of people). And this gives very strange results, because people will take as an "absolute truth" elements which are not at all necessarily demonstrated as being true, and, furthermore, they aren't even really interested in any demonstration about if these elements are true or not : they "believe" in them in the same way that one can "believe" in something linked to a religion ; something "revealed" that you must adhere to... But the result is very paradoxical when the given "belief" is related to something purely created by humans.

And I think that you are right : the "unquestionability of Chess960", for some of its proponents, is a "belief system", that works in this way.

By the way, I wonder how this occured ; it stands to reason that, if I would invent today a new chess variant, not one would think that it is necessarily perfect and that nothing can be changed in it. Is it because it is Fischer who invented Chess960 ? But Fischer was a genius, but not a god ! A creation by Fischer isn't necessarily perfect and impossible to improve... Is it something in Chess960 itself that tends to generate such reactions ? Globally, I wonder, and don't quite understand these (excessive) reactions...

- My criticism of the present Chess960's castling is linked to the fact that the difference between kingside and queenside castling in traditional chess is due to the positioning of the King, which generate an asymmetrical castling. But this asymetrical castling has been kept in Chess960, in which the King can be positioned on 6 different squares, without any difference between the kingside and the queenside. So I don't see why this asymmetrical castling should be kept in Chess960. Furthermore, this difference between a-side - = queenside - castling and h-side - = kingside - castling tend to generate more frequently positions for which, after castling, the King is nearer to the center than before castling, which I find illogical.

Petrarlsen 9/26/2018 06:24

@ lajosarpad (1/2) :

- About chess variants with pieces' placement, my impression is that celeje is quite spot on when he says : "I think Bronstein chess is very interesting, but that is a single new starting position -- the blank board. Opening theory is then placing the pieces. If you convince lots of chessplayers to play Bronstein chess seriously, they'll be forced to develop the opening theory of placing pieces on the initially blank chessboard."

My impression is that we would in the end have the same situation that in traditional chess, the difference being that most of the opening theory would consist of the pieces' placement part. Some theory would been also developed for the "normal play" part, but, obviously, as the number of initial positions would certainly be rather great, the limitations of the human memory would necessarily imply that the "normal play" part's theory wouldn't be as much developed as the pieces' placement theory. Would you agree or disagree with this ?

- "I think it is fair to put the fairness of the positions as all chess960 positions are fair for the proponents of the game."

I am sorry, but I didn't understand what you meant by that (the beginning : "I think it is fair to put the fairness of the positions"). Could you elaborate a little ?

Petrarlsen 9/26/2018 06:15

@ celeje :

(About draws, tied matches, etc.) "I believe this is just the US view. Americans are just not familiar with draws, ties, etc. American sports are like that. Non-American sports often have draws. The solution may be just to persuade the Americans that a sport with ties is more nuanced, even if it leaves people without instant gratification."

Very interesting. And this could obviously also explain why the sole World Champion from the United States since the beginning of the 20th century, Fischer, was so intent on obtaining a "draws not counting" match...

I rather wonder what Caruana (who could perfectly well become the first World Champion from the United States since Fischer) would think of that ?

Petrarlsen 9/26/2018 12:49

@ celeje :

""Completely satisfying" is so vague and subjective." (about one of my posts in which I said : "In fact, my impression (...) is that most Chess960's proponents consider as some sort of a general principle (...) that there isn't even one of these 960 starting positions which isn't completely satisfying.")

"Completely satisfying" isn't per se subjective ; it all depends on what you mean by this. "Vague", yes ; and it was deliberately so ; each person could have a different reason (good or bad) to find a position "not satisfying", so I didn't want to enter into too much details. As for me, I have written a sufficient number of posts under this article or the last one for everyone to know for exactly which reasons I consider some starting positions as being unsatisfying (in particular for reasons linked with Chess960's castling...).

Petrarlsen 9/26/2018 12:24

@ fgkdjlkag :

- "I am wondering what your background in classical music is?"

Near-professional level in piano and history of music (I quite hesitated between law, on the one hand, and to teach piano or history of music, on the other hand), + other disciplines that goes with one or the other (musical theory, musical analysis, etc.) + a rather good level in violin (I learned violin playing for more or less 20 years, but I've stopped many years ago, because it is too time-consuming to do everything at the same time - my maximum was more or less at the difficulty level of Bach's Sonatas and Partitas) + the basis of organ playing + some experience in harpsichord and clavichord playing.

So I quite know what I am talking about, and I rather suppose from what you said that you also know what you are talking about, but that our opinions differ about many things in classical music. I will not continue this discussion about classical music, as it would very probably turn into a new debate, and this is not at all the place for a debate on classical music...

I will just state what my opinion is : that in classical music in particular, and, more generally, in many artistic fields, you can find many quite knowledgeable persons who can approve VERY strange things, and that this is the same with Chess960 : some Chess960's proponents would approve in the most absolute fashion the strangest things about Chess960.

celeje 9/25/2018 08:23

@lajosarpad:

@lajosarpad: "@celeje

Completely satisfying is looking like something subjective, but I think Petrarlsen referred to the criteria shown in one of my previous comments, which can be quantified and objectively measured, possibly Petrarlsen has some other criteria, but judging from his arguments I am sure he agrees with the criteria I have presented. "

Okay, thanks. I'll have to go back & look at your previous comments again. I already gave what I think is all that really matters --- starting position not practically pre-determining win/loss/draw.

celeje 9/25/2018 08:19

@lajosarpad:

@lajosarpad (to Petrarlsen): "They seem to be not interested to keep the game a zero sum game or close to it, but to have more decisive results. See fgkdjlkag's comments, from where we can see that some people have a problem with matches resulting in a tie. I fail to understand what the problem with tied matches is and they did not explain it in a way that I would understand it, they handled this position more like an axiom."

I believe this is just the US view. Americans are just not familiar with draws, ties, etc. American sports are like that. Non-American sports often have draws. The solution may be just to persuade the Americans that a sport with ties is more nuanced, even if it leaves people without instant gratification.

lajosarpad 9/25/2018 04:25

@Petrarlsen

when watching the Olympiad, the World Cup and being interested in a lot of games, the level of confusion will increase, people would want to be home at the front of their computer, rather than there personally. This would cut a large proportion of the profit, unless radical changes are done in the organization of the tournaments.

Anyway, back to theory. I think theory would switch to pattern recognition for chess960 rather than knowing variants by heart. Some years will be needed to even start the change of approach. As a result, if chess960 would replace chess, then I think the whole thing would die fairly quickly, because if everyone considers chess to have a problem and to have to be replaced, then convincing people to join chess would be increasingly difficult, chess would be defeated ideologically then. And chess960 as I have shown has a high chance to defeat itself. I do not follow chess960 either. The fact that I do not understand the position at the start and would need to look at it 1 hour before the game to have some basic understanding makes it unattractive for me and if they blitz out the first moves, then the whole thing will be random for me. As a spectator, I would rather read a book than to watch chess960 as it is today. If chess960 would replace chess, they would lose my humble self as an enthusiast. I know it is not a big loss, but I might be not alone.

@fgkdjlkag

"But you will find easily thousands of persons who like every chess960 position."

Really? Let's conduct a survey, asking chess players what they like exactly about a random chess960 position, giving them two minutes to think before they answer. I think the majority of answers will be "I don't know", "I did not study this position", "Huh?" and so on and I doubt we will get many meaningful answers to that.

@celeje

Completely satisfying is looking like something subjective, but I think Petrarlsen referred to the criteria shown in one of my previous comments, which can be quantified and objectively measured, possibly Petrarlsen has some other criteria, but judging from his arguments I am sure he agrees with the criteria I have presented.

"I'm gonna say it's one where the starting position largely determines the outcome --- win, draw or loss. It's hard to find one chess960 position where that seems to be true. "

I agree with the way you validate, but I doubt we have enough knowledge to determine whether there are many positions which largely contribute to the result or not. I understand it is hard for you to find a position which highly contributes to a given result, but I believe you would face considerably higher difficulties in proving about all chess960 positions that they are all playable, balanced and not drawish.

lajosarpad 9/25/2018 04:07

@Petrarlsen

I think it is fair to put the fairness of the positions as all chess960 positions are fair for the proponents of the game. They seem to be not interested to keep the game a zero sum game or close to it, but to have more decisive results. See fgkdjlkag's comments, from where we can see that some people have a problem with matches resulting in a tie. I fail to understand what the problem with tied matches is and they did not explain it in a way that I would understand it, they handled this position more like an axiom. Now, while I completely disagree with this position, I approach to it with respect and tolerance and show why I consider it to be incorrect. I have no problem with draws or tied matches, in fact if I have to choose between analyzing a correct draw or a decisive game with some mistakes, I would probably choose to analyze the correct game, as it is more instructive. A tied match shows that the players are around the same strength, while matches decided in a sudden death game would provide the false impression that one of the players is stronger than the other. A coin toss instead would at least be honest.

Chess960 has no guarantee to provide the highest quality according to your definition to quality, which probably is not far from my definition, it might be equivalent. However, the problem is with different ontologies. Proponents of chess960 seem to define quality quite differently. They seem to consider a decisive game full of drama to be of higher quality than a very deep draw with very intuitive ideas. The definition of quality is an important factor to form tastes. Avant guard art has a lot of supporters, who regard some artworks to be of high quality, while I find difficult to find value in the very same artworks at all, even though, occasionally I like some avant guard artworks. There are many, who like the drama of a decisive game, for them chess is mostly a sport. For me it is mostly an art, a science and an abstract philosophy, so there are people who will never agree with me about this, but I am okay with that. If they want to see chess960 tournaments, I have nothing against it, but let's keep classical chess for those, who find it more valuable than 960. There were always powerful belief systems, Katolocisism in medieval Europe, Islam in the Islamic world, Liberalism in the Jacobine era, Communism in the Soviet Union and China, National Socialism in the III. Reich and, for a lot of chess players, the unquestionability of chess960. Without doubt this is unscientific, that is why its proponents have a high chance to be shocked when we even show that we are not convinced that they are right. I was never shoked to see people who want to replace chess with chess960, I just disagree and explain my opinion. The other side tends to be more temperamental, but that's okay.

I do not have an opinion about changing the way castle should happen with chess960. I have allocated five minutes to understand the rules of castling and stopped reading about it when the time elapsed. As far as I understand the castling results in the king and the rook being at the very same square as in chess, but I am not sure how it works if the king is on, say, b1 and the rook is on a1. Is castling possible if the rook of the opponent controls d1, for example? Intuitively I think it is possible, because the king does not pass d1. Also, if you start to watch a game and you did not see it from the start, you might have a harder time understanding the position. If the king is, for example on c1 and the two rooks are on b1 and a1, respectively, then it is unclear whether the king can castle with the b1 rook now or can castle with the a1 rook if the b1 rook moves. This kind of unclarity exists in chess as well, but to a much lesser amount, as the king has to be on e1/e8 in order to even consider castle as a possibility and the rook in a corner on the same row.

lajosarpad 9/25/2018 03:47

@Petrarlsen

Since thie was never studied by anyone deeply, we are on uncharted territory, so I can only give my opinion. Let's say there are only 96 decent starting positions in chess960, this is actually a small number, as it is onnly 10% of the possible positions. This means that there is a high probability that the complexity of the game was raised 96x. Even though computers can handle this (or more precisely, will be able to handle this if software catches up with chess960 expectations, which is not very difficult, actually), having a ready-made answer for roughly all the positions which could arise against oneself and to be up to date in the opening theory, this is so difficult a task, that it is analogous to writing a doctorate. It is humanly possible, but difficult, no wonder there are not hundreds of thousands of super GMs. Now, if we continue the analogy, doing this for 10 positions would be analogous with writing 10 doctorates at the same time. This is still humanly possible, but only for a very few people. Now, if you and I play and we are both super GMs (which is quite a pleasing thought experience), then, since 96 positions are humanly impossible to know by heart even close to sufficient number of opening variations and critical lines, both of you and me will have a few favorite starting positions, let's assume 4 for each of us and maybe 10 - 15 positions we will be okay to play. Now, if we are to place the pieces, then the idealistic goal is to achieve some of the positions we have top knowledge about. I want to get my positions and avoid your positions at the same time. I would rather want to play a position not known to any of us, than to play your positions. Another important factor here is that I would not object to get a position outside the 96 decent positions if the position gives me an advantage. Now, if there are a lot of positions which favor white or black, maybe 750 positions favoring white and 114 positions favoring black, then for you, as black, the start of the game would be to avoid those 750 undecent positions for you and my positions as well, which is of tremendous difficulty and in fact Black is under pressure before move one.

Now, back from our thought experiment with the assumption of having 96 decent positions and back to stable ground, where we do not assume anything, I see two possibilities with high probability: the first is that there are many decent positions and the second is that there are many undecent positions. The two cases can be true at the same time, our example of 96 decent positions has too many decent positions for a human to remember theory for all of them and a huge majority of undecent positions. If there is a large amount of undecent positions, then the game will not be fair and the side with less probability to achieve a decent position will be under pressure before the game actually starts. Also, if the number of decent positions is known, then theory will no longer be objective, it will be about knowledge about the other player's preferences. This exists even today, but to a much lesser amount.

celeje 9/25/2018 11:32

@ Petrarlsen:

Petrarlsen (to lajosarpad): "In fact, my impression (and I always find this very strange - I wonder why they reason like this) is that most Chess960's proponents consider as some sort of a general principle (almost a "belief") that there isn't even one of these 960 starting positions which isn't completely satisfying."

"Completely satisfying" is so vague and subjective. If you asked "most Chess960 proponents" whether they consider every one of those 960 positions is completely satisfying, I'm gonna guess they pause, look stunned & ask you what you mean by "completely satisfying".

I'll guess what may be "unsatisfying". If I'm asked what an unsatisfying chess960 starting position is, I'm gonna say it's one where the starting position largely determines the outcome --- win, draw or loss. It's hard to find one chess960 position where that seems to be true.

I do not find it "unsatisfying" if I or someone else thinks it's "ugly" or "inharmonious" or whatever. That's just an unthinking instant reaction to unfamiliarity. We all want to take comfort in the familiar.

fgkdjlkag 9/25/2018 07:27

I am wondering what your background in classical music is?

News

ChessBase '26 - Mega Package

Expand your Chess Horizon The perfect equipment for 2026 with the latest ChessBase program '26, Mega Database, ChessBase Magazine and Premium-Account!

€349.90

Reversed Sicilian Power - Win with 1.c4 e5!

In this course, IM Nico Zwirs presents the Reversed Sicilian as a powerful and practical weapon for White

€39.90

ChessBase Magazine Subscription (Digital)

The magazine for professional chess

€19.90

Master Class Tactics - Train your combination skills! Vol.4

In almost every chess game there comes a moment when you just can’t go on without tactics. You must strike to not giving away the advantage you have worked for the whole game.

€39.90

ChessBase Magazine Extra 229

Opening videos: Daniel King presents new ideas against Caro-Kann with 3.Nc3 dxe4 4.Nxe4 Nf6 5.Nxf6+. ‘Mikhalchishin's Miniatures’: Najdorf, Petroff and Scotch. ‘Move by Move’ with Robert Ris. ‘Lucky bag’ with 37 analyses by Ganguly, Illingworth et al.

€14.90

SHOP

SHOP

Champions Showdown: Improving the format

ONLINE SHOP

Mastering Chess Strategy Vol.3 - Exploit your opponent’s weaknesses!

Chess960 is here to stay

Sinquefield is sold

Room for improvement

Picking the position

Kasparov's final thoughts on the Champions Showdown

Links

Discuss

ChessBase '26 - Mega Package

Reversed Sicilian Power - Win with 1.c4 e5!

ChessBase Magazine Subscription (Digital)

Master Class Tactics - Train your combination skills! Vol.4

ChessBase Magazine Extra 229

Understanding Middlegame Strategy Vol 13: Reversed Colour Systems – Grünfeld & Dutch

A powerful 1.e4 Repertoire

ChessBase Magazine 229

Pop-up for detailed settings