Cascla: Responding to points from ChessTalk posters so far

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Cascla: Responding to points from ChessTalk posters so far

    I am delighted to see information, ideas and constructive suggestions from posters following this thread! This is what I was hoping for.

    1) Notation:
    This was raised by Mr. Tessier and Mr. Bonham. My ultimate plan and hope is to have this Cascla idea discussed by FIDE, which would then decide on a final notation for it, to be used worldwide. Certainly, we have some issues with notation. For one, I have used K, Q and N to denote castling locations and / or properties. These letters, as abbreviations, are not the same in all languages. I knew this, and have an alternative system, which I will outline now. It draws upon the languageless Chess Informant's symbols: they use >> for King's side, and << for Queen's side. Keeping the 'N' for cases of no castling, as I have used so far, we could have these three to describe properties. So, in the game I provided yesterday, Hughes -- Dixon, the very first ever published anywhere with a Cascla, Black won, castled first on the Kingside, and White castled on the Kingside. That was: (K-/K'+). It would be: (>>/>>'+). I believe my system presented yesterday is superior that offered by Mr. Tessier, simply in terms of how the eye reads information. Having / to separate the sides is easier to comphrehend, and it furthers the idea of the Cascla entity.
    2) Denoting which side castled first:
    Mr. Tessier and Mr. Bonham also raised this point. I described a ' to denote which side castled first; this is from mathematics, where this symbol denotes a 'prime' in the sense of a differential, for calculus. I have a background as a mathematician and statistician, and worked extensively as a geophysicist, dealing with large amounts of data. So I have some technical credentials and insight, to go with my chess background, and have combined all of this to create this Castla concept. I didn't actually have this idea incorporated for quite a while after I began working on the system, which has been several years in the creation to this stage. Not using it leaves us with 13 classifications. My main reason for using it is that chess is a reactive, dynamic game. When we play, we are reacting to what our opponent does on the board. As the game unfolds, players are having to make decisions based on their opponent's choices. Castling is a fundamental aspect of strategy, since King safety is a requirement for successful outcomes. Checkmate ends the game, so if we wish to win, we need to take care of King safety. Questions arise such as: should we castle at all, should we castle before or after our opponent castles, should we castle on the same side as our opponent? I determined that aiming to understand which side has castled first, and whether this impacts the game's properties in this regard, WOULD in fact be important in statistical determination and analysis. I am arguing from the standpoint of a trained professional. If it winds up being 'noise', as Mr. Bonham asserts, then that will come out as well, but we should analyze it first.
    3) Castling 'lag':
    Again, I didn't originally have this feature. It may wind up being 'noise', as Mr. Bonham believes, but including it now leads us to the possibility of making that conclusion AFTER analysis, not BEFORE. From the scientific standpoint, using it makes eminent sense.
    4) Data analysis methods:
    Mr. Brodie and Mr. Bonham follow this point. Mr. Brodie states that my Cascla idea is NOT currently trackable on ChessBase, which is significant. I believe the conjoined aspect of the castling property is, thus, new, and very significant from the standpoint of future application and understanding. Incidentally, for those who don't know, Mr. Brodie is the creator and maintainer of the 'CanBase', collecting the largest set of Canadian games in the world.
    Mr. Bonham took a lot of time and effort, drawing upon his knowledge and experience, to itemize many constructive suggestions, and for this I thank him sincerely! I want to take more time to analyze his material in detail before responding.
    5) Why ChessTalk?
    I want to explain that I have thought a lot about how to launch this idea. Chess "N Math Association is itself a groundbreaking organization, founded by IA Larry Bevand in 1985, and it has been very successful on several fronts. I've worked with Larry on a number of tournaments, and hold him in high regard. ChessTalk is available and accessible. It attracts many talented and passionate chess people, from grandmasters to other strong internationally-titled players to arbiters to coaches to journalists to average players to organizers to sponsors to chess fans. At its best, it is an astonishingly effective forum for chess discussion on a variety of topics. I felt that it would be an optimal platform for launching this concept. We are already seeing it work well. :)

  • #2
    Re: Cascla: Responding to points from ChessTalk posters so far

    Originally posted by Frank Dixon View Post
    I am delighted to see information, ideas and constructive suggestions from posters following this thread! This is what I was hoping for.

    1) Notation:
    This was raised by Mr. Tessier and Mr. Bonham. My ultimate plan and hope is to have this Cascla idea discussed by FIDE, which would then decide on a final notation for it, to be used worldwide. Certainly, we have some issues with notation. For one, I have used K, Q and N to denote castling locations and / or properties. These letters, as abbreviations, are not the same in all languages. I knew this, and have an alternative system, which I will outline now. It draws upon the languageless Chess Informant's symbols: they use >> for King's side, and << for Queen's side. Keeping the 'N' for cases of no castling, as I have used so far, we could have these three to describe properties. So, in the game I provided yesterday, Hughes -- Dixon, the very first ever published anywhere with a Cascla, Black won, castled first on the Kingside, and White castled on the Kingside. That was: (K-/K'+). It would be: (>>/>>'+). I believe my system presented yesterday is superior that offered by Mr. Tessier, simply in terms of how the eye reads information. Having / to separate the sides is easier to comphrehend, and it furthers the idea of the Cascla entity.
    2) Denoting which side castled first:
    Mr. Tessier and Mr. Bonham also raised this point. I described a ' to denote which side castled first; this is from mathematics, where this symbol denotes a 'prime' in the sense of a differential, for calculus. I have a background as a mathematician and statistician, and worked extensively as a geophysicist, dealing with large amounts of data. So I have some technical credentials and insight, to go with my chess background, and have combined all of this to create this Castla concept. I didn't actually have this idea incorporated for quite a while after I began working on the system, which has been several years in the creation to this stage. Not using it leaves us with 13 classifications. My main reason for using it is that chess is a reactive, dynamic game. When we play, we are reacting to what our opponent does on the board. As the game unfolds, players are having to make decisions based on their opponent's choices. Castling is a fundamental aspect of strategy, since King safety is a requirement for successful outcomes. Checkmate ends the game, so if we wish to win, we need to take care of King safety. Questions arise such as: should we castle at all, should we castle before or after our opponent castles, should we castle on the same side as our opponent? I determined that aiming to understand which side has castled first, and whether this impacts the game's properties in this regard, WOULD in fact be important in statistical determination and analysis. I am arguing from the standpoint of a trained professional. If it winds up being 'noise', as Mr. Bonham asserts, then that will come out as well, but we should analyze it first.
    3) Castling 'lag':
    Again, I didn't originally have this feature. It may wind up being 'noise', as Mr. Bonham believes, but including it now leads us to the possibility of making that conclusion AFTER analysis, not BEFORE. From the scientific standpoint, using it makes eminent sense.
    4) Data analysis methods:
    Mr. Brodie and Mr. Bonham follow this point. Mr. Brodie states that my Cascla idea is NOT currently trackable on ChessBase, which is significant. I believe the conjoined aspect of the castling property is, thus, new, and very significant from the standpoint of future application and understanding. Incidentally, for those who don't know, Mr. Brodie is the creator and maintainer of the 'CanBase', collecting the largest set of Canadian games in the world.
    Mr. Bonham took a lot of time and effort, drawing upon his knowledge and experience, to itemize many constructive suggestions, and for this I thank him sincerely! I want to take more time to analyze his material in detail before responding.
    5) Why ChessTalk?
    I want to explain that I have thought a lot about how to launch this idea. Chess "N Math Association is itself a groundbreaking organization, founded by IA Larry Bevand in 1985, and it has been very successful on several fronts. I've worked with Larry on a number of tournaments, and hold him in high regard. ChessTalk is available and accessible. It attracts many talented and passionate chess people, from grandmasters to other strong internationally-titled players to arbiters to coaches to journalists to average players to organizers to sponsors to chess fans. At its best, it is an astonishingly effective forum for chess discussion on a variety of topics. I felt that it would be an optimal platform for launching this concept. We are already seeing it work well. :)

    Hi Frank, I'd like to respond further to some of this. Regarding notation and language, aren't the pieces notated "KQBNRP" in all pgn games worldwide? And also in all FEN diagram notation (with the Black pieces being lower case)? I could be wrong on this but then all the web sites with pgn and FEN parsers would have to handle different languages if there were such differences.

    If what I just described IS the global standard with pgn and FEN, then I think you should conform to that standard, because eventually your notation may be added to both pgn and FEN.

    Further regarding "castle lag".... I was thinking, why not capture which actual move each side castled on (if they castled at all)? Then the lag could be computed from that. The thing is, lag is only one piece of information, whereas the move number that actually was castling would add additional information. What if, for example, someone wanted to analyze Cascla notation to ask from a particular pgn database of games: "What % of games did White win where White castled Kingside and White's castling was done on or before move 6?" Your current notation would not be able to answer that query -- it can only determine lag, as you call it.

    Quick question Frank: if I were willing to donate Hadoop programming (including Spark analysis) to your project free of charge, would you be willing (and able) to get yourself plus some Queens University chess players to play some correspondence games of Option Chess by email, or even (preferably) hold a weekend tournament in Option Chess with usual time controls?

    Option Chess is my own gift to the chess world. If you go to chessbase.com and do a search on my name plus "Option Chess" you will find an article from 2014 I authored describing it. It is relatively simple, not some 10 x 10 hugely complex monstrosity, but rather a small extension to standard chess, designed in fact to be an eventual successor to standard chess in the long term. You could become part of history, for Cascla and for Option Chess. Or not... But Louis Morin in Montreal will tell you that Option Chess is fascinating to play.
    Only the rushing is heard...
    Onward flies the bird.

    Comment

    Working...
    X