Thanks for your insight. I use the two resources you mention but feared I had been missing out all these years on a wonder weapon known only to the .ctg book experts.
> feared I had been missing out all these years on a wonder weapon known only to the .ctg book experts
There is no such thing, although many of the .ctg makers would want you to believe otherwise. Untill 2-3 years ago, it was a good idea to have armed yourself with 100 thousands of latest Playchess games, as you could find there interesting ideas. Now not any more, I can't remember the last time I saw anything original from those.
Databases often consist of:
Weak players against good players,
weak engines against good engines,
or weak opening books against good opening books,
or weak against weak,
or good against good,
or time runs out and game is lost in a winning or drawn position,
or luck decides a game,
or games from slow time control,
A move that is often played and has a winning chance of maybe 60%+ doesn't mean it's a good move.
So in my opinion databases and statistics aren't reliable.
I guess even quality databases have many flaws in it, and to find them out you have to analyse every game and every move in it. A very time and power consuming task.
Maybe the best thing is to begin analysing from the 20 possible start positions with a good engine,
and going further through the chess tree choosing only the best 5 to 10 moves in every position and so forth for example.
The requirement for such a project is to have the right analysing method, because analysing with 30+ plies doesn't make a move a good move, after move 15 you can be sure you have chosen a good path, ...
After about one million analysed positions (with 5-10 alternatives per position) and many months of hard work you can be sure to have a picture of many good moves you can pick from
and mark them in appropiate colours (green, blue, black, yellow and red - if it's an Aquarium tree). A small problem with great impact could arise -> The engine can ponder a very weak move, it alters the minimaxed tree to the worse, and to find it costs more time.
Such a project is a very time-, power-consuming and costly task. And you need severe discipline, stamina, ultimate perfection and will power to bring such a project along,
but you can be very sure to have very good moves in your opening book derived from analysis and not from statistics.
The one with such an opening book should be highly rewarded for it's work when he may release such a chess jewel, shouldn't he?
But i know the world wants everything for free, so such an opening book might not exist in the near future. Maybe! ;)
>Maybe the best thing is to begin analyzing from the 20 possible start positions with a good engine, and going further through the chess tree choosing only the best 5 to 10 moves in every position and so forth for example. The requirement for such a project is to have the right analysing method, because analysing with 30+ plies doesn't make a move a good move, after move 15 you can be sure you have chosen a good path, ...
>After about one million analysed positions (with 5-10 alternatives per position) and many months of hard work you can be sure to have a picture of many good moves you can pick from and mark them in appropriate colors (green, blue, black, yellow and red - if it's an Aquarium tree).
You'll be long dead before you can start playing with you crayons, because as soon as ply 11, you're going to have a tree with billions (thousands of millions) of positions.
But you don't need most of them...
Here's a small video for you about my work i do for Chess.
BTW, in the Fried Liver, what's up with spending so much time with 5… Nxd5, instead of 5… Na5?
> Watching that video, I can tell you couldn't avoid the temptation, of getting the crayons out of their box, well before meeting your guidelines. In the Berlin Wall, there's just two alternatives analyzed, by move 15.
This video i have made only for you in just an hour of work. It's only an example line. The opening book isn't deep it's wide.
And to examine the example line it would cost me about 2 weeks to analyse it thoroughly.
At move 11 for example there are so many possibilities... But, i have it in my queue right now, because i don't want to disappoint you. ;)
> BTW, in the Fried Liver, what's up with spending so much time with 5… Nxd5, instead of 5… Na5?
That was a project in this forum some time ago, where some guys wanted to prove Mr. Hyatts statement that Black wins easily, and we found out that he is wrong.
Don't worry i have 5. Na5 in my queue. ;)
Have you investigated the 9th move of the Fried Liver more exactly?
Then you would have seen that the Hiarcs Opening Book does a weak a3-move for White that gives Black too much advantage...
I have spotted many such mistakes in other opening books. Anyway.
The video should only show you how complex my work is, how thorough i analyse and i have to follow many lines.
I have no automatic script, i do it by hand, i put all lines in the queue by hand, and all moves are marked by hand position per position.
Call me insane, 'cause this is what i do all day long, but i am open for new thoughts.
You can give me interesting lines and i may go through it when i have time.
Hyatt, I seem to remember something of the kind. You guys give too much importance to his statements, he was probably using Crafty to analyze Nxd5!
I don't have access to the Hiarcs book, but I don't miss it. I'm not talking about this a3 move in particular (which looks weird from my 18 hundred vantage point, why on earth does white allow a check and gives black the initiative?). I met a guy in the EM, who always uses that book, and his performance is run-of-the-mill, even thought he has decent HW. But that's a problem with all commercial books; some pay little money to the bookmakers, and get an accordingly cheap product, while in other cases, the person responsible just isn't up to the task. The best one is probably the Sicilian, but even they have been unable to win more than a single EM, which isn't that much, considering they're a bunch. Private books rule.
I don't think that spending your time, in doing what you like, constitutes insanity. Addiction? Maybe.
It will take me about 2 months, maybe 3 months, to fully analyze it.
I only have to analyse the good paths and the good sidelines and hit the Syzygies (Endgame Table Base).
The rest is up to the engine when it comes to play those lines.
Every game is a draw so nothing special and no wonder weapon will occur.
I am surprised how good this book is, so i respect the work of Mark, but he has to go through it to find the weak spots like in the Fried Liver Attack move 9.
The Rybka 4 book (R4) is a small book where most of the good moves, when i pierce through the tree, are marked red. I don't know why?! You can spot them when you watch my small video.
The Houdini 2.0 Book (H2W) is only bells and whistles. It's an outdated book with some good lines in it, but not very deep and thorough. The most positions derived from statistics (> 60 million?!?).
The Junior 13 book (J13) i haven't gone through very much, but it's a nice book to have it in my repertoire.
A nice book is also packed with Deep Fritz 14 (DF14). Not very strong but it has some good moves in it that i follow from time to time.
Perfect 15t (PF15) is also nice. With this small 8mover i can see where Sedats opening book competitions lead to.
Mytree is an old tree analysed with Houdini 2 to 4. I have started the ESGTB-project with Houdini.
RS is Rising Star with very few marked moves in it, but it's going (statistically) very deep. A real database book.
8mover GM is the old Stockfish Fishcooking book.
So i have bought all the commercial books to see what they have in their repertoire and i use all of them only as a reference.
When i go through my little tree i see this and that and after analysing so many positions i can check and compare it with other opening books.
That's all i wanted to have.
From time to time i flush all the queues and set new tasks for not getting stuck on some unimportant sidelines.
Have fun watching.
Btw. The tree has grown to more than 1.2 million analysed positions. In the video you will see that i am in the Berlin Defense / Berlin Wall Opening,
and it's getting further really fast. I am on my way up to move 15 in the main lines, so i think 2 weeks was a very good estimation. ;)
(If i would like to have all the sidelines then i would need about 2-3 months. And consider! I use more than 5-10 moves per position!)
I would be very pleased to hear some really good openings from you that i could check as well.
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill