Chess Tempo

Username:
Password:
/ Register

User Details

Username:
Blitz Rating:
Standard Rating:
Logout
December 02, 2008, 08:37:24 pm *
Welcome, Guest. Please login or register.
News: SMF - Just Installed!
 
Pages: [1]
Print
Author Topic: Reducing Repetition in Custom Sets  (Read 369 times)
newlook
Jr. Member
**
Posts: 69



View Profile
« on: July 30, 2008, 05:38:39 am »

I am starting to see a lot of repetition in my personal problem sets. (I recently saw the same problem twice in a row!) If the algorithm that minimizes repetition in rated sets could be applied, that would be great. I realize that I can create new sets with never-seen-before problems, but I'd rather not make a lot of sets with the same theme.
Logged

richard
Administrator
Hero Member
*****
Posts: 990



View Profile
« Reply #1 on: July 30, 2008, 10:00:17 am »

Hi newlook,

Applying the anti repetition code to the personal problem sets is a little tricky but certainly not impossible. I've added this to the todo list.  The current algorithm just takes a random problem from the set, so with reasonable sized sets repetition shouldn't be too much of a problem, but it could certainly be annoying on smaller sets.

Regards,

Richard.
Logged
drahacikfm
Sr. Member
****
Posts: 418


View Profile
« Reply #2 on: July 30, 2008, 10:10:17 am »

Newlook:  I saw in your history you are working on a problem set called 1300-1399.  I'm doing similar things, working my way up the standard-rated problems from the lowest to the highest.  Because when I do standard-rated, I never get problems below 1800, so I figure I will never see problems under 1800 unless I make a personal set and work on them.

So far I have done all problems rated under 1200 (4399 of them!).  When I make the personal sets, in the search I always click "Problems I have never got right".  Doesn't seem to make sense to choose any other option, because I only want to do each problem one time ( or once more if I got it wrong the first time).  Why would you want to make personal sets that allow repeats?  Richard already has it set up so you can avoid repeats.  I don't see why he needs to add complicated non-repeating code for sets a user creates without specifying no repeats, when it's easy to specify no repeats  Smiley  If you created the set 1300-1399 with repeats, you can just delete that set, and create it again, this time specifying no repeats.  I did that several times when I made a mistake in the set creation.
« Last Edit: July 30, 2008, 10:15:23 am by drahacikfm » Logged

FIDE Master Drahacik
newlook
Jr. Member
**
Posts: 69



View Profile
« Reply #3 on: July 30, 2008, 02:44:16 pm »

I thought that the problem sets were static -- so if you created a set with "problems I have never seen" below 1000, the "never seen" part was only true at the time of set creation. As drahacikfm says, those conditions are dynamically updated, so problems I see are removed from the set. Thanks for the tip. I will use that feature, and I take back my request for an additional non-repetition algorithm.
Logged

drahacikfm
Sr. Member
****
Posts: 418


View Profile
« Reply #4 on: July 30, 2008, 08:13:23 pm »

Yes, the sets are dynamic.  A "set" is not actually a saved list of problems.  It's the saved search parameters you used to create the set.  Each time you use the set, the search is done again with those parameters and one problem is shown to you.

Under the solving board, you see how many problems are left.  It's fun to see the number go down until you have done them all.  Then you get the message "The problem set is empty.  Please select another set" .. or something like that.  And you know you did them all!  But it's possible new problems will creep into that set later as the ratings of problems (or the tags) change.
Logged

FIDE Master Drahacik
richard
Administrator
Hero Member
*****
Posts: 990



View Profile
« Reply #5 on: July 31, 2008, 12:45:14 am »

Newlook: I plan to support static sets at some point, but I will have to restrict the number of static sets each user can have as it could get quite expensive storing the explicit problem number of each set (especially as the number of users increases).  I think in most cases dynamic sets are more useful, the area where static sets would be useful is in creating hand picked custom sets where you can add to and remove specific problems from a set.  I really need to look at writing some kind of manual to describe the site features in a bit more detail, but I've putting this off as much as possible as it is more fun to work on more features than writing manuals :-)

Regards,
Richard.
Logged
newlook
Jr. Member
**
Posts: 69



View Profile
« Reply #6 on: July 31, 2008, 06:17:32 am »

I really need to look at writing some kind of manual to describe the site features in a bit more detail, but I've putting this off as much as possible as it is more fun to work on more features than writing manuals :-)

Maybe you could set up a wiki for us to collectively write documentation.
Logged

richard
Administrator
Hero Member
*****
Posts: 990



View Profile
« Reply #7 on: July 31, 2008, 07:34:51 am »

Newlook: Great idea, are you aware of any wiki software than is geared especially towards user manual type wikis or do you think people usually use the mediawikia (same as wikipedia) stuff for this type of thing?

Regards,
Richard.
Logged
newlook
Jr. Member
**
Posts: 69



View Profile
« Reply #8 on: July 31, 2008, 04:56:13 pm »

I don't know enough about the different wikis to recommend one over another. I think I've seen MediaWiki used more often (outside of Wikipedia), and it seems a safe choice in terms of features and support. WikiMatrix has a lot of comparison information. I didn't see any mention of a wiki specific to writing a manual, but you could narrow the field based on other features (page history, kind of database, free and open source, etc.).
Logged

richard
Administrator
Hero Member
*****
Posts: 990



View Profile
« Reply #9 on: July 31, 2008, 06:43:39 pm »

Thanks for the link newlook, lots of interesting wiki software. Some of the structured wikis such as Twiki look like they are intended for the slightly more structured job of user docs. I'll try to squeeze in some time to look a bit deeper into this, but it probably wont be a short term task.

Regards,
Richard.
Logged
Pages: [1]
Print
Jump to: