Great site! A few questions...

[tmr]

Thank for the great site.  One question, have the problem ratings changed in the last day or so?  Seems as if quite a few mate in two problems are rated fairly high, higher than it seems they've been in the past.  My rating has been rising on these problems and I'm not too sure it is because I'm getting better. 

Also, how random is the problem selector algorithm?  I've seen quite a few repeats so far and I've only done less than 400 problems.  This seems strange with the large problem set you have.  Of course I'm no statistician.

I've read a few posts here a bit critical of your problem selection.  Personally I like the wide variety of problems you've included and while I find some solutions frustrating it is normally because I've overlooked the best line.  On those problems that ask for the quickest mate, like 24693, I think you should retain these despite the critisism.  It is fairly clear that when the problem didn't end after the simple capture that the fastest mate is going to be the objective.  While perhaps not real world, I think these still have value in honing ones analytical capabilities.

Lastly, a suggestion on the tagging of problems and other information displayed during problem solution.  I haven't noticed many tags but when present, they inevitably give away the solution.  It might be helpful to hide the tag until requested (similar to the show/hide comment buttons).  Similarly with other information such as the average time to solve, the problem rating or even problem quality.  Often times these will give me a hint that can direct me toward the proper solution.  Such clues would not exist in real life.  I suppose ideally one would have the option of showing this information during or after problem solution.  Of course I can choose to ignore them but I'm weak and often find myself grasping at any little thing that will give a clue to the solution.

Any way thanks for the great site which I find quite superior to the other options out there.

[richard]

Hi Tmr,

As the method for choosing problems is based on a bell curve around your current rating, the chances of getting a repeat are much more likely than the total number of problems in the problem set would suggest. While it is theoretically possible to see any problem in the problem set, most problems you see will be clustered around your current rating.

I generally agree with you on the problem selection issues, however enough users found the "winning alternative" and longer mate issues annoying enough for me to do something about.  I have an updated problem set that removes a large number of mates with these issues. Hopefully when the update goes live it will not detract from the experience of those who were already happy with the current problem set and remove some of the frustration other users have been experiencing.  The main downside has been the removal of many of the longer mates, however the problem set was probably too heavily weighted with mates in any case.

I agree with you on the tags.  What will be come the next release of the UI already has that problem addressed.  Not only tags, but also the rating, quality and comments for the problem are not visible while the tactic is live. They become visible after you complete the problem. This will be the behaviour for all users. Unfortunately I can't make it optional as for the ratings to be fair, everyone needs to be operating under the same conditions. 

Regards,
RIchard.

[tmr]

Thanks for the update and all of the work on the site.  I look forward to the new update.

TMR

[chipschap]

I have an updated problem set that removes a large number of mates with these issues. Hopefully when the update goes live it will not detract from the experience of those who were already happy with the current problem set and remove some of the frustration other users have been experiencing.

I confess to being annoyed at times, especially when I find a "good enough" vs. "best" solution.

However, when I feel this way, I completely miss the point.

The point is, rather than saying "I'm annoyed because...." I should instead ask, "Is the experience on this site improving my chess?"  And of course, the answer is an emphatic "yes."

[richard]

Hi chipschap,

My opinion on the "good enough" vs. "best" has shifted a bit over time.  Originally I believed that it was mostly a non-issue and that the positions that still had "good enough" alternative moves were still instructive and therefore belonged in the set.  However I've come to the point where I probably rate "over the board" reality as a more important factor.  Forcing users to find moves there would be no reason to find in a real game might be counterproductive to effective training.  Finding the "best move" probably still has some value in these situations but overall I think the problem set is best without the positions with "good enough" alternatives.  The new problem set does manage to weed out a large number of the "good enoughs", however there are costs.  There are very few longer mates left so the problem set has got a bit easier (overall). I'm hoping I've left enough challenge in for the higher rated players, but this is difficult to assess without having the set live for a considerable amount of time.

Anyway, I'm going to roll the dice on the new problem set soon, as a last resort I can always revert back to the older set if there is a consensus that the new set is a step backwards.

Regards,
Richard.

[chipschap]

Richard, I of course leave it to you to decide what problems to include or exclude; my point was simply that, as things stand, I still benefit tremendously from practicing here, and I need to concentrate not on being annoyed by the occasional problem, but on what I'm learning.

Over the board I am currently just above USCF 1300, an extremely modest rating (I've been as high as almost 1400 but I went into a slump)---- but a rating at which tactical practice is by far the highest-payoff item.  Since starting to practice seriously both here and at CTS my over the board play has improved noticeably, and I can hardly wait to see how that works out in my next tournament over Memorial Day weekend.

[richard]

Cool. Good luck at the tournament, (let us know how you go!). 

As I'm currently adding a few features to the site ,I'd be interested in your feedback as a user of both here and CTS. Which areas would like to see Chess Tempo improve in (especially in relation to CTS)? Feedback from other users on this topic is also more than welcome.

Regards,
Richard.

[chipschap]

I don't know if I can give a proper answer as I use CTS and CT for different purposes.  CT has the "standard" mode and in general provides me with harder problems.  CT, therefore, is the place to practice calculation in a situation unconstrained by time.  It is closer to a tournament situation.  I might do 10 problems on CT for every 50 or more on CTS, but I am thinking rather than reacting.  This is a very strong advantage for CT; on CTS everything is constrained by the need to solve a problem in just 3 seconds to obtain full score.

On the other hand I use CTS to do pattern drills.  I'll do 50 problems at a sitting (whereas 10 or so is typical here) and simply react to each one.  Of course, this has over time (I've done about 10,000 CTS problems) increased my "memory store" and my average CTS rating is slowly rising (I look at the long-term trend to even out daily fluctuations).

On the other hand, CT has a blitz mode but I've used that very little.  The problem set seems less suited to blitz mode due to its overall harder nature and existence of duals and alternate solutions---- CTS is pretty clean in that respect.  And therein is the main advantage of CTS: a mature problem set which has stable and pretty meaningful problem ratings.

As far as being user-friendly both sites are good, with CT seeming to have a few extra features (somewhat better stats for instance, with a 1000-item graph rather than a 200-item graph).

So, bottom line, where should you go with CT?  I'd say it's really quite good at this point and very, very useful.  Continuing to weed through the problem set may have the best payoff.  And, of course, growing the user base ---- that helps all around especially in getting stable and reliable problem ratings.

Hope this is at least of some use!

[richard]

Thanks, chipschap.

That was certainly useful , especially the comparison between how you use CT and CTS. I'm keen to find out if the new problem set (when made public) gets closer to CTS in overall quality without, throwing out too many of the more legitimately difficult problems.

Regards,
Richard.

[oblonsky]

Hi Richard

There is something that you seem to have right here that may be missing over on CTS.  Your solutions are not ticked as correct unless the line is complete.  CTS has the unfortunate habit of accepting the first move and assumes the rest of the line is understood by the user and thus does not have to play through it.

The reporting here is slightly better than on CTS, though as many have pointed out, the problem set does not give as much credit for tries that are almost as overpowering.  I'm sure appropriate pruning of the problem set will take care of this.

The most important difference is the standard option.  It is so much more like the real world to take your time and actually play through the variations of a position without the imagined pressure of the clock ticking.

CTS is designed for quick responses so I have to admit i find it better for training my intuition or to get a feel for what kind of tactical motives I miss out on.  I think this is an issue with the problem set rather than the interface.

I think the blitz option would have, for me, an advantage over CTS simply because you have to do more to win credit if the problem set was that little more stable.  The stats page confuses me and is slow to sort. 

The point of statistical analysis based on some ratings criteria is to judge your relative position within the group.   There are no position numbers on the stats.  I appreciate all of this is vanity, but I do like to know how bad I am compared with the other users and then pick off targets to gauge my relative improvement.

These are minor details for a site I like to use as part of my training regime.  It may not be possible to implement many of the features of CT-Art as Convekta, or it ChessOK?, have more resources to play with.

I'm sure you are aware of Convekta products so I won't bore you by attempting to describe them here.

[richard]

Thanks oblonsky,

I also like the way CT usually lets the tactic reach its conclusion, while not perfect it does seem to get this right more often than CTS.  It's hard to guess without seeing their code but I suspect CTS's early pruning is done to avoid showing what might be ambiguous moves after  that point.  When I detect ambiguity I prefer to throw the problem out but CTS has done a better job (at least so far :-) ) at not presenting ambiguous positions to the user. Hopefully the upcoming problem set update will keep the ability to see tactics play out and at the same time remove most of the ambiguous positions (mostly mates with alternative material wins).

I assume when you said that the stats were slow to sort that you are referring to the users list page and not the stats tab? I've completely reworked the users list display, the new (unreleased) version  has ranking numbers in the left column and does all of the sorting on the server so it should be much faster (at the cost of a small delay when paging through the pages in the user list).

I've only read about CT-Art's features, not actually used it (it doesn't run on my operating system of choice).  If there are CT-Art features that you think might be useful to have on CT, then by all means let me know and I'll add them to the todo list.

Thanks for the comments.

Regards,
Richard.