I've just added a new book to tacticsladder.com! It's a collection of 10,000 carefully analyzed mate-in-two problems. Each puzzle has been evaluated on metrics like legal moves, unique checkmating moves, engine nodes used, calculation time, piece count, king mobility, and much more.
If on buys the book is it possible to get the set of positions as a dataset?
I do like that the curation is attempting to be transparent, and I would be curious if I were to use those, for my individual experience or not, to be able to use those as some well defined dataset with some well defined question set. Is that dreaming? It seems that even with more definable (try to) restrictoin chess is still big, but concentrating on checkmates not far from terminal position seemed like a good place to start getting reall with notions of patterns.
But the book barrier to well defined data sets, might be in the way. Since as kindle book it is in my range, I wonder if I would have such access or if I would have to add personal experience as another barrier? I welcome such chess position dataset the more transparently curated (i.e. reproducible or in that direction) the better.
Excellent Martin. Thank you. I will soon have study time for a week or thereabouts. Have a Happy Holiday and Christmas yourself! All the best, John.
Thanks John, sounds good! Merry Christmas 🎄
If on buys the book is it possible to get the set of positions as a dataset?
I do like that the curation is attempting to be transparent, and I would be curious if I were to use those, for my individual experience or not, to be able to use those as some well defined dataset with some well defined question set. Is that dreaming? It seems that even with more definable (try to) restrictoin chess is still big, but concentrating on checkmates not far from terminal position seemed like a good place to start getting reall with notions of patterns.
But the book barrier to well defined data sets, might be in the way. Since as kindle book it is in my range, I wonder if I would have such access or if I would have to add personal experience as another barrier? I welcome such chess position dataset the more transparently curated (i.e. reproducible or in that direction) the better.
happy hollidays.. and good tidings.