ChessMint: Chess training sheet generator based on lichess puzzles

Thank you very much for the ideas! Making a list of possible features / improvements, so this helps.

@b4c4d2 I like all of your suggestions, and they should be relatively easy to make happen. Will have to take a look at what piece set options there are, but yeah, more control over the look of the resulting PDF would be great. And I agree that there is some room for improvement w.r.t. the spacing/general look of the resulting PDFs. And yeah, haven't even considered en passent, you're totally right :D

@aeroplanino1986 agreed, that would be nice to have for blindfold. Could be tricky to implement, because number of pieces is not data that I have ready to work with, and I probably don't want to do any heavy data pre-processing, but maybe there are some options (sorting by FEN string length before sampling or something along those lines perhaps). Will keep this in mind as a possible feature!

@TotalNoob69 Not sure I understand exactly what you mean, but defining new, complex measures in order to make puzzle selection more specialized is not something I will do anytime soon.
For now, I find the rating range and theme selection quite nice already (on lichess, you can't define an exact rating range for example). But in general I agree that something like this could be interesting, even just to think about conceptually. (What makes a puzzle "interesting"/"surprising"/etc.? Not easy to define/quantify)

Thank you very much for the ideas! Making a list of possible features / improvements, so this helps. @b4c4d2 I like all of your suggestions, and they should be relatively easy to make happen. Will have to take a look at what piece set options there are, but yeah, more control over the look of the resulting PDF would be great. And I agree that there is some room for improvement w.r.t. the spacing/general look of the resulting PDFs. And yeah, haven't even considered en passent, you're totally right :D @aeroplanino1986 agreed, that would be nice to have for blindfold. Could be tricky to implement, because number of pieces is not data that I have ready to work with, and I probably don't want to do any heavy data pre-processing, but maybe there are some options (sorting by FEN string length before sampling or something along those lines perhaps). Will keep this in mind as a possible feature! @TotalNoob69 Not sure I understand exactly what you mean, but defining new, complex measures in order to make puzzle selection more specialized is not something I will do anytime soon. For now, I find the rating range and theme selection quite nice already (on lichess, you can't define an exact rating range for example). But in general I agree that something like this could be interesting, even just to think about conceptually. (What makes a puzzle "interesting"/"surprising"/etc.? Not easy to define/quantify)

TotalNoob69

Not sure I understand exactly what you mean, but defining new measures in order to make puzzle selection more specialized is not something I will do anytime soon.
For now, I find the rating range and theme selection quite nice already (on lichess, you can't define an exact rating range for example). But in general I agree that something like this could be interesting, even just to think about conceptually. (What makes a puzzle "interesting"/"surprising"/etc.? Not easy to define/quantify)

I meant the opposite: in order to be puzzles, the moves are already interesting. So we could use them to calibrate an algorithm for how interesting a move is.

> Not sure I understand exactly what you mean, but defining new measures in order to make puzzle selection more specialized is not something I will do anytime soon. > For now, I find the rating range and theme selection quite nice already (on lichess, you can't define an exact rating range for example). But in general I agree that something like this could be interesting, even just to think about conceptually. (What makes a puzzle "interesting"/"surprising"/etc.? Not easy to define/quantify) I meant the opposite: in order to be puzzles, the moves are already interesting. So we could use them to calibrate an algorithm for how interesting a move is.

NaperCoach

unable to generate any Blindfold puzzle sheets. Is it a working module?

IM iakov98 edited

@NaperCoach You‘re right, thanks for letting me know. Will look into it.

EDIT: Should be fixed now!

@NaperCoach You‘re right, thanks for letting me know. Will look into it. EDIT: Should be fixed now!

Your network blocks the Lichess assets!