Classical chess or Fischer Random? Or both?

by Arno Nickel
11/10/2022 – Which positions arose in Fischer Random games, which from classical chess games? This is what Arno Nickel asks, and in his essay he takes an in-depth look at Fischer Random Chess. His proposal: give players at tournaments the option of playing either classical or Fischer Random!

ChessBase 18 - Mega package ChessBase 18 - Mega package

Winning starts with what you know
The new version 18 offers completely new possibilities for chess training and analysis: playing style analysis, search for strategic themes, access to 6 billion Lichess games, player preparation by matching Lichess games, download Chess.com games with built-in API, built-in cloud engine and much more.

More...

Classical chess or Fischer Random?

Reflections on a (still) unequal pair of twins

Although the second FIDE World Championship in Fischer Random recently took place with Nakamura, Nepomniachtchi, Carlsen and Abdusattorov, to name just four of the eight finalists, the followers of this chess discipline with its 960 possible basic positions still feel like something exotic in chess clubs and chess communities. Not everywhere, but this sense of the matter is very widespread. More than 20 years of intensive efforts by private organizers and individual clubs, in our country above all by Hans-Walter Schmitt and his Chess Tigers, have not changed this decisively.

It is well known that many grandmasters who are tired of the sprawling opening theory and its consequences for competitive chess prefer to play Fischer Random rather than so-called ‘classical chess’ and have far fewer problems with the changeover than the average amateur who brings his opening variations and systems to the board like a life insurance policy. It is not the unknown that attracts him, but the solid ground under his feet, where he is sure not to fall into some hole. That is better left to the opponent. Fischer Random is not an issue for him, although he hardly knows it. Typical objections: “I would have learned everything I know about openings for nothing”, or “I hardly understand normal chess, why should I do this to myself? You can’t beat the classical starting position with its symmetry and inner harmony”.

The point here is not to convince critics and sceptics of the merits of Fischer Random, especially as this is most likely to succeed if someone has their own experience with it. And, of course, there is no good arguing about taste. But a basic insight, which also reached FIDE in 2009 at the latest, should by now be indisputable: Fischer Random or Chess960 is in no way inferior to classical chess and is at least as substantial, whether you like it or not. The fact that many, if not most, games that originate from a Chess960 starting position take on more and more moves of the positional patterns we are familiar with is an unmistakable indication that this type of chess basically hardly differs from classical chess. The games only have a different history of origin.

In the appendix, four positions are given as examples, each after White’s 20th move, two of which come from classical chess and two from Fischer Random games. Try to match the positions correctly without using tools such as databases. It might not be that easy, and most of the time you will hardly succeed. (You can find the corresponding games in our dynamic replayer at the end of the article).

Occasionally, observers attribute the tendency towards traditionalism to the fact that the players strive for familiar positional patterns. This is probably true to some extent, but should not be overestimated. What is more decisive is that the inner logic of chess leaves them little choice but to follow the laws of space, force and time. These apply in Chess960 just as they do in classical chess. What other strategic and tactical motives, what other techniques should work here?

What really does not promise success would be the attempt to transfer the opening principles from classical chess one-to-one to Fischer Random. It is true that all these opening principles generally apply in the random variant, but the priorities often have to be set or observed quite differently if, for example, a pawn is uncovered in the starting position and can already be threatened by a bishop or queen on the first move.

A typical case: an uncovered pawn on a7 is attacked by a bishop on g1 after 1.f3 or 1.f4, and if there is still a trapped black rook on b8, then hallelujah! Occupying the centre with 1...d5 or 1...e5 wouldn’t be such a good move, but that doesn’t mean that the centre no longer plays a role. After the second move, things can already look significantly different.

It is precisely this relativity of the theorems that makes the Fischer Random so appealing to its followers and all those who value creativity as the highest good. My conclusion: classical chess and Fischer Random are basically twins. According to my definition, Fischer Random is nothing other than neoclassical chess and should no longer be thought of and treated only as an antithesis to classical chess. As far as I am concerned, it is also neo-romantic, but that is another matter. We should not narrow the definition of classical chess to the starting position, but consider the totality of chess rules as constitutive. Changes at one point do not necessarily mean that the whole edifice collapses. On the contrary, one must even make small changes from time to time, measured in decades and centuries, to keep the whole thing alive.

But enough of the long preface. The question that is on my mind, and I’m sure on the minds of others, is: what should happen next for those chess players who would like to see and play more Fischer Random, but find too few opportunities to do so?

Above all, there are too few opportunities for long games under tournament conditions. Online play is not a satisfactory solution in principle, apart from the fact that 90% of the games there are only blitzed. Quick tournaments, which are offered here and there by clubs and private organizers are much more attractive, but there are still far too few of them in each area, and they are (understandably under cost aspects) only rapid tournaments and not comparable to classical tournament chess.

Should we wait until the enthusiasm for Fischer Random to trickle down from the top to the bottom? Until the number of Fischer Random fans grow so large that large open tournaments are profitable for organizers and regional and local championships become a matter of course? That may take time because not only God’s but also Fischer’s mills grind slowly.

If classical chess and Fischer Random are basically twins, why can’t they appear together in a tournament? To put it quite concretely and simply: in any chess tournament I would like to have the freedom of choice to play a Fischer Random game with my opponent, if he ‘ticks’ the same way as I do, instead of a game with the classical starting position. In my opinion, there is nothing against this in principle, even if some technical peculiarities occur that mean extra work for the tournament directors, apart from the fact that they would have to familiarize themselves with the rules for Chess960.

The publisher Arno Nicke, here at a tournament in Berlin in 2006

The matter is not that complicated and would in no way endanger the existing hegemony of classical chess, since no one would be forced against their will to play Fischer Random in even a single game. But for everyone it could nevertheless be interesting to follow a few Fischer Random games here and there in a tournament in which mainly classical chess is played, and later possibly also to analyse them with the players concerned. If, during the years, one or the other of the initially few Chess960 disciples were to join, i.e. be baptized, that would be fine

Basically, I imagine it like this in an Open: Those players who would like to play Fischer Random in a tournament announce this to the tournament management before the start of the tournament, and are then provided the opportunity to do so, which is binding for both sides, if they come across a like-minded person in the draw according to the Swiss system.

The Swiss system rules should be strictly applied to the draw, i.e. no pairings for or against Chess960 should be arbitrarily created. The starting rankings will be determined by Elo (or possibly DWZ); Fischer Random rating numbers do not play a role for the moment, as long as there are no officially binding numbers for them. (Note: They could later, if recognized by the federations, play a secondary role in determining ranking within the Chess960 player list). The starting position should be drawn just before the start of a round and should be binding on all Fischer Random players in that round.

At the FIDE World Championship and also in Saint Louis (there under the new brand name Chess 9LX), the players were informed of the starting position a certain time before the start of the game to familiarize themselves with it (without electronic help) so that they did not have to invest a lot of thinking time on the first move, especially as they were playing with a rapid-chess time control. This could be handled in the same way in the open tournaments envisaged here, whereby the FIDE grace period of 15 minutes seems quite appropriate.

Another aspect concerns the avoidance of advantages and disadvantages in the choice of colours. The simple possibility that each player plays a starting position with both White and Black (as is the case in many Ches960 competitions) obviously does not work in an open. Here everyone has either White or Black, so the question arises whether the starting position drawn favours one of the players. This special topic, for which there are already computer evaluations of all positions, cannot be dealt with in depth here, but it is not of fundamental importance. To date — according to the current Wikipedia entry — no game-deciding White advantage has been proven for any of the 960 possible positions. The entry explains: “The chess program Stockfish, for example, evaluates the 960 opening positions with a search depth of 39 half-moves with values between 0.0 and 0.57 pawn units advantage for White (mean value 0.18), whereby the opening position of traditional chess is evaluated with 0.22 pawn units advantage for White”. In my opinion, however, this aspect should not be overrated.

In principle, the following experience applies to Fischer Random in particular: the stronger player or the one who has better understood and fulfilled the positional requirements under the specific tournament and time conditions in a specific case wins, regardless of whether White or Black is playing. And by the way, as a rule, no one can rely on computer preparation here. It would nevertheless be conceivable that positions proven to be ‘suspicious’ are taken out of the huge pool of positions available for selection in order to avoid a colour advantage or disadvantage, as incidentally always happens with starting position 518 (the classical one).

As far as the evaluation of such ‘mixed’ tournaments is concerned, it should in principle be assumed that the points won are equivalent, regardless of whether they were won with classical chess or in Fischer Random games. Nevertheless, it is clear that it would take a change in the FIDE tournament rules for such tournaments and the performances and ratings achieved to be officially recognized — in essence, it should suffice that not only the classical starting position, as laid down in the Laws of Chess of FIDE, applies, but also the starting positions drawn according to Chess960.

This will certainly not be the case overnight, but should one therefore bury an idea from the outset that is in itself a good one? In the long run, private initiatives have achieved many things that initially seemed unattainable. That should also apply here. Patience is needed until Fischer Random receives the recognition it deserves as part of the mainstream. 

The four positions

The four games

 


Links

Discuss

Rules for reader comments

 
 

Not registered yet? Register

arzi arzi 11/18/2022 06:23
I think these different variations of chess were created for fear of too many draws in tournaments? Like in this tournament.

https://theweekinchess.com/chessnews/events/st-louis-fall-chess-classic-2022

Only 21 wins in group A and 29 in group B.

If one person likes to play or watch football (soccer) and the other likes American football, let them watch or play those games, but don't try to mix them up or eliminate either of them. They are different games like Chess and FRChess. This simple thing shouldn't be too difficult to understand on a universal level or meta-level.
ICCF Grandmaster ICCF Grandmaster 11/18/2022 12:18
The contrast between classical chess = scientific and FRC = anti-scientific is complete nonsense. Everybody who has dealt with FRC for a longer time recognizes recurring structures and regularities despite the diversity of the starting positions, which can be seen as the starting point of a new theory formation.
There is by no means an anachy of randomness, but the experiences condense over time to valid insights. Compared to the prevailing highly specified opening theory, theory formation takes place on a more universal level. Perhaps one can say: on a meta-level. Otherwise it would be a pure game of skill and would hardly inspire world-class players.
lajosarpad lajosarpad 11/15/2022 12:31
@JNorri Yes, throwing away a significant part of chess culture is another aspect resulting from switching to FRC. If someone hates opening theory (more specifically the work needed to be invested into it), then the person has a legitimate position. Yet, those who like the semi-scientific activity of researching the opening and the culture arising from it also have a legitimate position. I fully respect that some people prefer FRC, even though I do not share their enthusiasm about that game. However, the article had the incorrect statement according to which chess and FRC are twins, which is historically false. @Mendheim reinforced this statement, so he/she might be the same person as the writer of the article, but, at least, the commenter and the article writer are like-minded.

I do not like the idea of injecting FRC into chess tournaments, but I have no problem with FRC tournaments. How would we assess the performance of the players if they arbitrarily change what game they are playing? How would we calculate the tournament points earned during the tournament if one is playing more FRC games than the other? Would we compute FRC games into chess Élő points? These are problems the article did not address, as it only focused on the preference of some players to play FRC and their lack of opportunities for doing so.
lajosarpad lajosarpad 11/15/2022 12:30
@arzi

One may argue that it is not always possible to determine the exact preceding position by just looking at the table position and that's true, so let's add the additional specification that having the moves from the starting position, one can reliably deduce what the preceding position of a ply was.

This means that if one looks at a chess position after 15 moves, the person will often what the main motifs were so far, even without knowing the moves, yet, if we speak about FRC, then looking at a position after 15 moves does not serve us with this kind of information, unless we carefully study what the original position was.

So, we can reduce positions to previous positions (and that's what retro chess problems are about, banking on the reducibility of the end of a bottleneck variation), but, in the advance forward, the variations diverge instead of joining. If you've meant that any given position is the result of branching out from previous positions and we can use the meet operation in order to go back to previous plies, ultimately reaching a well-defined starting position, then you are correct in saying so.
lajosarpad lajosarpad 11/15/2022 12:29
@arzi

"Prime numbers are the "REAL" numbers because all the other integers, whole numbers, are coming from the prime numbers, like Chess with it´s side games. :) "

It is an interesting way to put it. Basically, divisibility is a non-strict poset (Partially Ordered Set), because it fulfills the trio of the criteria:

1. Reflexivity: x is divisible with x
2. Antisymmetry: if x is divisible with y and y is divisible with x, then x = y
3. Transitivity: If x is divisible with y and y is divisible with z, then x is divisible with z

It is a partially ordered set, because there are incomparable pairs, like 5 is not divisible with 3 and 3 is not divisible with 5. And it also has a smallest element (1, because anything else can be divided by 1). It also has a largest item (0, because 0 can be divided by anything else).

Now, looking at your analogy, the numbers coming from prime numbers are results of the join operation (composite numbers) and reducing two joined operations is the result of a meet operation, so we can form a Lattice for divisibility.

If we view a chess game as a poset, where the ply index along with the position resulting after the given ply is a node/element in the relation, then a somewhat similar structure arises (a game tree), that is, the meet of different moves is the position preceding them in the previous ply and we can deduce any pair of positions transitively up until the starting position (as long as it's not randomized). However, we cannot join the moves, that is, a move is the start of a new branch in the game tree in the context of the position, so we cannot form a lattice of chess plies in a similar manner as we can form such a lattice for divisibility.
JNorri JNorri 11/14/2022 07:10
Without commenting on the gist of the article - the proposals for the future - I wish to express strong disagreement with this statement in the preface: "a basic insight (...) should by now be indisputable: Fischer Random or Chess960 is in no way inferior to classical chess and is at least as substantial, whether you like it or not."

In at least one very important way it is as good as categorically inferior. One only needs to read Polugaevsky's 'The Birth of a Variation', or indeed hundreds of treatises on the development of particular openings or variations, or the interrelations between them, some more scholarly than others, to appreciate that with only 960 a mountainous part of beautiful cultural history of chess would never have evolved. Of course it is difficult to convince critics and sceptics if they see no value in such work.

It is not inconceivable that some sort of shuffle chess had emerged the norm, considering the supposed dice controlled variants in India and Medieval Europe.
arzi arzi 11/14/2022 06:32
Lajosarpad:"With such an attitude prime numbers would have been thrown away as "dead art", instead of finding their use in cryptography thousands of years later than their discovery."

Actually, it would mean that the Chess would have been thrown away as "dead art". Prime numbers are the "REAL" numbers because all the other integers, whole numbers, are coming from the prime numbers, like Chess with it´s side games. :)
toposcar toposcar 11/13/2022 12:49
Personally I wish to go even further: why not a superfisherandom 960x960 where white and black have different positions (except one of 960 possibilities, obviously). I have tried it many times, and never never white has a huge advantage, the position is just heavily unbalanced!
lajosarpad lajosarpad 11/13/2022 01:36
@Mendheim

"In the end, however, chess theory is a dead art if it does not either feed its operator or lead to insights in other fields of knowledge. For 99,9% of all chess players we may exclude this. It is nice to occupy oneself a bit with "science", even if it may lead to nothing."

With such an attitude prime numbers would have been thrown away as "dead art", instead of finding their use in cryptography thousands of years later than their discovery. One may dislike opening theory, but others like the researching component of the opening and that's also a legitimate position.

@fgkdjlkag

"If you want to analyse the opening position, this can be done for all 960 starting positions."

The whole point of FRC is to do away with opening preparation. You can analyze a starting position, but if all your research will become useless in the next game (or the next year), then it will make much less sense to invest effort into it. If, for instance, a position is played for a year, then by the time you get used to it, you will have to start learning something else.
lajosarpad lajosarpad 11/13/2022 01:36
@Mendheim

"Comparisons of the twin pair of Fischer Chess/classical chess with table tennis/tennis or ice hockey/other ball games or milkshake/lemonade, on the other hand, are hugely misguided. Fischer chess is played according to the same rules as chess with the classical starting position, it is played with the same material, the same space and even under the same time conditions."

That's an incorrect statement, of course. Let's refer to the facts: https://www.fide.com/FIDE/handbook/LawsOfChess.pdf

Article 2 (page 3), section 2.3 states "The initial position of the pieces on the chessboard is as follows:" and a screenshot shows the starting position. Hence, the rules of chess specifies the starting position and using a different starting position violates rule 2.3. As a result, the rules of Fischer random chess and chess are different, and saying the contrary is an incorrect statement.

"Fischer chess, however, has in principle nothing less to do with "science". Rather, it is the zero hour in the creation of new knowledge, because of course new theory emerges with increased practice and research. "

If one can create new knowledge by investing zero hours, then the value of such new knowledge has no relevancy, because in the future, players will be able to avoid all previous knowledge by investing zero hours. That's the whole point of FRC. So, while I respect the will of FRC players to get rid of the effort of investing hours into opening preparation, it is an undeniable fact that FRC, if it is indeed effective in the avoidance of opening preparation, then it takes away the researching aspect of opening theory. While I consider the opening preparation to be a vital part of chess, players preferring FRC are happy to avoid it. So, instead of denying the fact that this research would disappear in FRC, we could admit that this is indeed the case and agree to disagree.
genem genem 11/13/2022 12:30
@Frederic wrote
{(1) Every December FIDE should declare a single Fischer Random position which will be used in all FR tournaments during the upcoming year. This would allow a certain amount of preparation,
}
I strongly agree with this proposal. We would learn more about chess in general by enabling a year's worth of focus and repetition, than by fully randomizing every start setup constantly.

Also, some of the 960 start setups are less interesting to play or review. Avoid setups that place bishops in corners. Prioritize setups where both of White's knights start on the SAME shade of square (to produce an important form of variety compared to the one classic setup, a variety we might name as "knight oppostion").
Michael Jones Michael Jones 11/12/2022 11:19
It shouldn't be a surprise that after 20 moves it's difficult to tell whether a position arose in FRC or standard chess - the whole point of FRC is that from the middlegame onwards it's essentially identical, the only difference is that it removes opening theory. Players will tend to follow the same general opening principles - control of the centre, develop pieces, king safety - it's just that the exact way those principles are applied will vary depending on the nature of the initial position. Once most of the pieces have left the first rank, and in particular once a few have been exchanged, the position will gradually become indistinguishable from standard chess. A No-Castling Chess position would be easier to tell from a standard one - the longer the kings remain in the centre, the more likely it becomes that the position arose from NCC rather than standard chess.

I find it interesting that FRC is mostly played at the elite level; I have never known a club-level tournament in it. Possibly because at recreational level, most players have little time or inclination to study openings in depth, so preparation plays a far smaller role than it does for super-GMs. I frequently find myself "out of book" at about move 5 and almost always by move 10, whereas a 2700 will often have prepared beyond move 20, and if they're having to think on the spot by move 15 it's a sign that something has gone wrong. Thus the main motivation behind FRC isn't relevant at lower levels of the game because there is little preparation involved in standard chess anyway. Perhaps there is a level where a player really "brings his opening variations and systems to the board like a life insurance policy", as the author suggests, but if so it certainly isn't the one I play at! I'd be happy to give FRC a go if I found that there was a tournament for players of roughly my strength, but if I did I doubt I'd find it much different from a standard chess tournament.
fgkdjlkag fgkdjlkag 11/12/2022 04:17
Elegant and interesting idea. I would play in a classical tournament under these rules. I suspect many GMs would, considering the wide interest.

@Mawin, how is it natural that the rooks start in the corners and have to wait until late in the game to do anything? There were many fascinating games in the recent chess960 world championship when rooks entered play immediately. The idea that rooks should be developed last is a holdover from chess518.

@Mendheim, I propound that chess518 theory does not lead to insight in any other fields. How durable and useful can such knowledge be when it becomes useless when you switch a bishop for a rook in the starting position? On the other hand, the creativity that chess960 requires, plus everything from the other chess, such as ability to concentrate, look ahead moves, etc, does translate to other fields.

I fail to see how chess960 reduces scientificity. If you want to analyse the opening position, this can be done for all 960 starting positions. All having 1 starting position does is increase accuracy of play for that position.
Mendheim Mendheim 11/11/2022 09:56
None of the critics here puts forward arguments that cast doubt on the quality of Fischer Random Chess. Only *lajosarpad* criticizes a loss of scientificity in favor of a randomly selected starting position, which minimizes theoretical preparation.
But it is precisely the growing excess of preparation that is the problem of the aging traditional chess, which is why Fischer Chess has become a real alternative for the future.
Fischer chess, however, has in principle nothing less to do with "science". Rather, it is the zero hour in the creation of new knowledge, because of course new theory emerges with increased practice and research.
In the end, however, chess theory is a dead art if it does not either feed its operator or lead to insights in other fields of knowledge. For 99,9% of all chess players we may exclude this. It is nice to occupy oneself a bit with "science", even if it may lead to nothing.
The two earlier proposals by *frederic* meet the need for manageable structure and containment of chaos. It is quite interesting that Garry Kasparov favored #2 (10 positions carefully selected by FIDE as a theme for one year [why not a little longer?]). A considerable compromise that could pave the way for the Fischer Random in the long run.
Mendheim Mendheim 11/11/2022 09:55
Comparisons of the twin pair of Fischer Chess/classical chess with table tennis/tennis or ice hockey/other ball games or milkshake/lemonade, on the other hand, are hugely misguided. Fischer chess is played according to the same rules as chess with the classical starting position, it is played with the same material, the same space and even under the same time conditions. Only one of many components differs and evaporates in the course of the game; in this respect, the two are just infinitely more closely related than any of the other comparative pairs.
anthonyy anthonyy 11/11/2022 09:25
and then we shall need 960 ELO, one for each starting position, times 3 or 4 depending on the time allowed
hansj hansj 11/11/2022 03:45
If people want to play fischer random, let them do that.
I do not even master classical chess.
lajosarpad lajosarpad 11/11/2022 10:31
I find Fischer Random Chess less interesting because of its randomness. One of the main reasons as of why I strongly prefer chess over card games is that it is completely lacking randomness and it is a semi-scientific activity, Fischer Random Chess is specifically (partly) taking the scientific aspect away by reducing preparation, even though subsequent analysis is also possible in Fischer Random Chess. So, if there is a Fischer Random Chess tournament, then, (if at all) I only watch the games superficially.

If chess and Fischer random are twins, then Fischer must have lived 1500+ years ago. If not, then let's not call them twins.

I also disagree with the term "neoclassical" applied to Fischer random. Lots of years need to pass before Fischer random chess will become a classical activity.

Finally, I do not see table tennis tournaments where players may choose to play tennis instead (or the other way around), affecting the overall result of the tournament. I rather see separate tournaments for table tennis and tennis.

It's understandable that Fischer random chess players feel a lack of playing opportunities, but, if there is a widespread feeling of this, then FRC tournaments will eventually be organized by sponsors.

Let's keep milkshakes and lemonades as they are, instead of mixing them.
Frederic Frederic 11/11/2022 09:55
I proposed, well over a decade ago, two variations: (1) Every December FIDE should declare a single Fischer Random position which will be used in all FR tournaments during the upcoming year. This would allow a certain amount of preparation, which has been part of chess for over a millennium. It would also eliminate players sitting there thinking intensely about move two: "Can I play 2.c4, or does it lose instantly?" Also: players know that the preparation they invest will be obsolete in the coming year.

(2) The second option is to select ten positions for the upcoming year. They are chosen for their reasonable starting positions and eliminate all the ugly, unnatural ones. Half an hour before each round of a FR tournament, the organisers (or the general public) select the one the players will get.

I discussed both options with Kasparov and Fischer. Garry liked the second version, Bobby was sceptical.
arzi arzi 11/11/2022 06:15
Arno Nicke:"If classical chess and Fischer Random are basically twins, why can’t they appear together in a tournament? To put it quite concretely and simply: in any chess tournament I would like to have the freedom of choice to play a Fischer Random game with my opponent, if he ‘ticks’ the same way as I do, instead of a game with the classical starting position."

Indeed, BUT, why can't and should not these two twin games be played as separate tournaments? After all, ice hockey and some ball sports have their own so-called twin sports, but you don't want to play or watch them as alternatives to each other, do you?

I have a shocking proposal that will blow the minds of the weak and the ugly. Let's call Chess as Chess and Fischer Random Chess as FRChess, and not mix them up, right?
Mawin Mawin 11/11/2022 05:44
I propose a relocation variant. In Relocation Chess the players can, before play begins, swap places of the king/queen and another piece except the rooks. This method gives 64 of the Fischer Random positions, which are all natural because the rooks remain in the corners. There are also variants that give only mirrored positions.
http://mlwi.magix.net/bg/relocationvariants.htm
PamH PamH 11/11/2022 12:15
The chess game as it was conceived is immensely difficult in all its aspects and we still see these people draw a chess with pieces in random positions, without meaning.
He is famously said, the faces won't get a rating like normal chess and we want to create confusion.
1