Puzzle Responses

[dlester]

I am just starting to participate here and haven't read all the history, so if this has been discussed already just ignore me.

There are times, rather frequently in fact, that I am viewing one of the puzzles and I see two, or maybe three moves that look good.  Obviously there is only going to be one 'most correct' response and even if there is another good response it is significantly 'less good' than the correct move (horrible grammar).

Anyway, since you are still apparently looking for content to differentiate subscribers from free members (and assuming your database structure could easily support it) you could keep track of the incorrect responses people make for each of the puzzles. 

Based on some criteria you could set (maybe five people have selected the same wrong response for example) then you could have your engine store the analysis for that move with the puzzle and offer to display the refutation for that move whenever someone plays it in the future.  Over time this could develop into a pretty impressive resource that could show a subscriber not only the correct move, but why their move wasn't the best.

[richard]

Hi dlester,

There is a new problem set almost ready for release which either removes positions with alternative winning moves, or where it can, records the alternatives so that users are not marked wrong when selecting them.

The premium feature set will also allow users to see all the mistakes made on a problem, how many times the mistake was made and the average rating of users making the mistake. This feature is in testing already.  I also intend to use the data as input to the engine for further analysis of the mistakes but this is a little further away (I also need to save the analysis I'm doing on the main line so that I can present that analysis to the user).

Regards,
Richard.

[d-man]

very nice, richard. I had this idea today that is a little similar, that you maybe could allow users themselves to add refuting lines (users can be very zealous when allowed). Or maybe the users could add some kind of hints for the puzzle (colored squares, arrows, like in CT art). Users could turn the "hint" on after a failure.

I know this might seem extreme or out of the vision of the site, but i thought When you fail a problem, you don't learn anything much unless you know why your move doesn't work.

Besides adding fen copying, maybe you could add pasting the position in the pgn viewer or somewhere where you could fiddle with the pieces around.

at any rate i am definitely considering signing up for this site, mainly because it is much more user friendly than any others i've seen. CTS makes me very frustrated for some reason.

[richard]

Hi d-man,

The hints idea is a good one, I have a few other ideas that also require support for arrow markers in the chess board UI so when I get around to this I'll see how easy it would be to add user based hints, my main issue would be how to avoid "bad" hints without spending a lot of time examing hints one by one..perhaps allowing a vote for best hint feature would suffice.

I'm hoping that a combination of the alternative lines and post-hoc analysis of common mistakes will make the refutation idea no longer necessary, but this depends on how good the new generator (and any subsequent versions) is at identifying all reasonable alternative lines.

An idea similar to hints and refutation lines is to allow users to enter the type of analysis currently often see in problem comments, but in a way which allows easy viewing on the board.  This would be particular useful where the idea behind the tactic is not clear after the computer opponents move.

An "analysis board" to look at the problem is something other users have also suggested.  I think it is a reasonable idea but it is lower priority at the moment.


Regards,
RIchard.