Siderite/lichessTools

Investigate Filesystem API and how it can be used to allow various epic features

Closed this issue · 5 comments

Maybe we can finally enable large game collections.

Also together with the IndexedDb API

if i was a pro player in tournaments, it has a "dossier" feature... you choose an opponent that you are playing the next day, and it will give you a kind of assassin dossier on them. See: en.chessbase.com/support-kb/(X(1)S(xoi2evfhwaovicizptxnlsae))/content/details/1155/Player_dossiers?AspxAutoDetectCookieSupport=1
but "prepare for an opponent" could also do with getting stuff from the fide website... i saw a great free lecture video about this here: www.udemy.com/course/the-grandmasters-secrets/?couponCode=ACCAGE0923 - see the lecture preview of "How to prepare for the game?" (18 mins) - could you somehow automate some of that to get key gmes - i think this feature would really help the active pro players.
maybe also even for lichess opponents too... a kind of "prepare dossier for player X"
Some are not convinced by CB prices here: www.reddit.com/r/chess/comments/17v3hmy/chessbase_pricing_and_new_features/ - if all they did is a new file format.. that doesn't seem particularly lucrative. I know CB 14 is quite stable - so another reason for me never to upgrade
FM ZibbitChess reckons you cant get the same results without chessbase: www.reddit.com/r/chess/comments/12fa0j5/is_chessbase_really_that_necessary/
possibly the file storage being optimised helps CB a lot for larger DBs but the value must also be tied into the database itself.
I guess if you could optimise the file storage to have say a million games occupying minimum megabytes that would be cool - that seems a major background technical feature of CB. And one could just pay to use TWIC's database for lgames played in recent years
in summary two things so far: Optimal file storage.... and Prepare for Opponent facility... me personally im not updating my CB - CB 14 seems okay for my purposes. I'm not usually referencing the BIg DB - i like to just download PGN from Chessgames.com generally.
The chessbase proprietry file format (.cb*) means that not only less storage space compared to normal PGN but also faster retrieval speeds and also adds more features to the data too potentially. I'm not quite sure if you want to stick to .PGN though - that is standard and somehow have a version which speeds up the .pgn maybe like your own compressed .pgn format. Searching say a million games for a start sequence hopefully won't take hours. And if search speeds are maximised, then you could have more sophisticated ways of searching for games or positions within games.
They have mentioned a new one here: https://help.chessbase.com/CBase/17/Eng/index.html?database_formats.htm
"do-it1 CBONE format

ChessBase now supports a new database format with the extension *.cbone. This format is listed in the dialogs for opening and saving files. The advantage of this format is that in contrast to the CBH format the whole database is stored in one file. This makes the format suitable for backups and for sending to other people. If you use a large number of small files you should consider using this format. It is possible to exchange games between CBH, CBF and PGN databases. You can create databases in any of these formats and copy games into it." sounds good actually - they had multiple files before
for me after creatign a course, i use Export to PGN to give PGN on my courses
so the "working format" is a Chessbase format usually
well if i need speed... actually often i just stick with .pgn though

  1. Players Database Updated
  • Search for real players, very important
  1. Access to cloud engines

databases: Caissabase, MegaDatabase, Yottachess, TWIC...

I just want to use everything in Lichess :)

I think it is relatively feasible to create a book of experts for a given position from a chess opening
It might look something like this:

  1. Click on the icon of another book. 2. We choose a color.

  2. After processing, instead of moves, we see surnames (or nicknames [though these can hardly be processed efficiently in huge databases, even the small Lichess Elite Database])

  3. We add additional important indicators
    (1a. number of games in a given position 2a. player rating 3a. winning percentage [ In theory this is important, often professionals just play the same thing without practicing, simply because they are used to it] 4a. Average rating of opponents)

  4. Here's how I see it roughly
    W Carlsen, Magnus 281(games) 2840(rating) %67 (percentage of results) 2690 (average rating of opponents)

  5. When we click on a player, we already get a tree of moves from his games.
    This will allow us to find a successful and strong player and copy some of his repertoire. Why do what has already been done for me? A cynical and reasonable question.

Much later, this idea could be developed into a feature to find grandmaster copies of a player. While this feature would be cool, it's probably globally not worth the conditions it needs.

Implemented in v2.3.150