Tactics server

Hi wormwood,

Richard from Chess Tempo here. I'd like to address a few misconceptions you appear have regarding the Chess Tempo problem set.

Firstly, there are ZERO problems on Chess Tempo that are taken from CTS. They are all extracted from publicly available PGN records of real games (as are the CTS problems), so there will naturally be some overlap (which is perhaps the source of your misconception). I doubt the overlap is that large, but I've not done any calculations to estimate the number. I've never asked to use the CTS set, and had started writing Chess Tempo well before I bumped into CTS (although I did notice CTS before releasing the first public version of Chess Tempo ,and borrowed their excellent idea of rating problems against people using glicko).

The initial Chess Tempo problem set did contain many ambiguous problems where a second best (but still winning answer) would get you marked wrong. The idea that finding 'the best' move should be good enough and that all other moves fail was clearly flawed, and the site was changed to deal with this well over a year ago (June 2008). Ambiguous problems are now either removed at the generation stage or alternative winning lines give the user a 'try again' message, allowing them to continue to look for the 'best move', without failing the problem.

You were certainly one of the first Chess Tempo users, and I imagine the original problem set probably left a bad taste in your mouth. To my knowledge, we have only ever interacted twice, once on the CTS message board where I posted to refute similar incorrect statements about ambiguity (to which you didn't reply), and an earlier solitary post you made on Chess Tempo reporting what turned out to be a UI bug (not a problem set ambiguity issue), just after the site started over 2 years ago. I'm a bit confused about these other conversations you say we had, is there a link to them somewhere so I can refresh my memory?

It's not uncommon for CTS users to blame the problems instead of their flawed solution when they first start using Chess Tempo. Chess Tempo tends to target harder problems than CTS, especially for CTS users who are solving for accuracy instead of rating on CTS (and therefore get very low ratings), so the obvious move is much less commonly correct on Chess Tempo, usually because of a non-obvious opponent response that refutes the more obvious , but incorrect move.

In any case, ambiguity is currently pretty much a solved problem on Chess Tempo these days. Sometimes you can find a move that might have deserved a 'try again', especially in early endgame positions where the engine can't see far enough ahead to notice that the move would be winning, but these positions are rare and much more common on CTS (where the engine appears to only use 5 seconds per position compared to many times that on Chess Tempo). If you have doubts about the level of ambiguity in CTS problems, I suggest you look at the CTS problems linked to in this post:
http://chesstempo.com/chess-forum/site_feedback/alternative_solutions-t484.0.html;msg2880#msg2880

CTS has hundreds of problems where there are moves which are clearly winning, that if played will fail the problem. That thread lists a small number of them, my sampling indicates that around 5% of their problems have these issues, I suspect users don't notice them 5% of the time, partly because problems with these types of ambiguities tend to get higher ratings (I know, because on Chess Tempo when I had these types of ambiguous problems, they tended to become quite high rated). The 5% is only a rough estimate, as it was based on a fairly small sampling of problems, it could be quite a bit smaller or larger (although I did try to be fair on my sample and include problems across a wide rating range).

I'm happy to debate the merits of CTS versus Chess Tempo, they both certainly have their pros and cons, but please try and stick to the facts. Your claims that I've essentially ripped off the CTS problem set are both offensive and false.

Regards,
Richard.

Originally posted by wormwood
like I said, I was there. you were not. and it's silly that you insist on things you know nothing about.

from what I understand, it turned out to be you whose claim was simply wrong. strange stuff.

Originally posted by richardchesstempo
Hi wormwood,

Richard from Chess Tempo here. I'd like to address a few misconceptions you appear have regarding the Chess Tempo problem set.

Firstly, there are ZERO problems on Chess Tempo that are taken from CTS. They are all extracted from publicly available PGN records of real games (as are the CTS problems), so there will naturally be som ed off the CTS problem set are both offensive and false.

Regards,
Richard.

thanks for taking the time to explain where CT currently is. like I said, I've always known you had only good intentions with writing your site. I have no problems with people preferring CT over CTS. if anyone likes it better, by all means knock yourself out.

however, you did ask on the CTS board if the problems were available for use, and shortly after that had a huge problem set which 'miraculously' looked exactly like the CTS set, only with the aforementioned problematic continuations. I always assumed you took them from CTS based on that, and you never said otherwise. we even discussed many of the problems with exchange sacs, comparing the CTS continuations with yours. at that time you never mentioned anything like "what a coincidence, it's the same exact problem as in my set". we had many discussions like that over the first months when you tried to get your problem set working (after we kept complaining about the errors). at some point you changed your handle, and we went at least twice over the same things again. there's also been a couple of discussions exactly like this one, prompted by the same reasons as this was. all of them ended with you saying "but now I've fixed the problem set", then I once again go check your set and hit a faulty problem within a short session. (and if you think I decided they were faulty just by my own patzer tactical judgment without double checking, well I didn't. I'd never think a problem was wrong simply because I got it wrong, that would be just stupid and counterproductive to chess improvement.)

I admit a lot of the reasons why I don't like CT stem from the early years. but the thing is, when you go to a site again and again, and always find the same old problems existing, after a while you get fed up with it. and it's my right to do so, just like it's your right to not to like CTS. you say once again that you've fixed the problems, well good for you. but I'm not gonna bother once again.

I understand why like to remember things differently. CT is your labour of love, criticism towards it is difficult to receive without emotional bias. but do you think it's maybe possible you're not being completely objective about your own creation either? I remember it this way, you remember something very different. it's been years already, and I certainly don't carry any backups for ancient CTS boards, so I can't really produce any direct quotes from you here. but this is how I remember it. if I got something wrong, it's not intentional. but I'm quite sure this is how it went.


all the best for you and your project. many people like it, and for that alone it has its place. it's not for my taste, but that's just a personal preference.

Hi wormwood,

It is possible you are confusing me with someone else, but I assure you I've never inquired about the availability of CTS problems (and I'm 100% sure on this because I'd never posted on the CTS board before Chess Tempo was released). In fact I'm fairly convinced you must have me confused with someone else, as I've never had the discussions you mention on the CTS board (which username do you remember having these discussions with? ). I've had lots of in depth discussions on the CT forum regarding the composition of the problem set, but none of them with you (unless you used a name other than wormwood back then), perhaps you followed these discussions without participating, and have mis-remembered them as occurring on CTS (btw you can use archive.org to look at some of the older CTS message board threads, I assure you that you will not find me discussing any of these matters on CTS (other than the single reply to your post that I've already mentioned, and happened more than a year AFTER CT was launched)).

It's also not accurate that I "suddenly" had a huge problem set. You can see in my older news articles when I added problems, it took quite some time before chess tempo reached the number of problems that CTS has, the initial set was considerably smaller. It also doesn't make sense that I'd take the problems and make them worse, the reason the CT problems were initially problematic is because the first version of the chess tempo generator only had limited ambiguity checks. Recent versions are much better and the ambiguity checking is now much stronger than what has been applied to CTS problems (as shown by all the ambiguous problems on CTS posted in the link I gave). CTS problems do have comparative strengths, but right now, lack of ambiguity is not one of them.

In terms of your most recent experience on CT, it's possible you were unlucky with one of the problems you were served, I'm not saying the problem set is 100% perfect, but it certainly deals with ambiguity considerably better than CTS does, and has done so for at least a year.

CTS is an excellent training tool, especially for drilling large numbers of patterns, I'm glad you've been able to find it useful, and I'm sure you're a much stronger player after the 100K+ problems you've done there (a fantastic effort by the way). I have no problem with you preferring CTS over CT, and I'm not interested in winning you over, however I do have a problem with you suggesting that I've ripped off CTS's problems without any actual evidence other than your memory. This is not just an issue of your memory of events versus mine, some of the things you describe simply never happened. If your memories of these events are accurate, then archive.org should be able to find some evidence of them. Please don't continue making these claims unless you have some actual evidence to back them up, as I interpret them as an insult to the 1000s of hours I've spent working on CT and building the generator that produced all its problems.

Regards,
Richard.

Time to raid the forums

Thread 118904

Originally posted by richardchesstempo
Hi wormwood,

It is possible you are confusing me with someone else, but I assure you I've never inquired about the availability of CTS problems (and I'm 100% sure on this because I'd never posted on the CTS board before Chess Tempo was released). In fact I'm fairly convinced you must have me confused with someone else, as I've never had the disc ...[text shortened]... ding the generator that produced all its problems.

Regards,
Richard.

well now that you claim you didn't have a CTS handle before the new one, I know you're simply making things up. I can't remember what your original handle was, but I certainly remember when you started using the new one.

which made me wonder a bit, so I popped on your site. and just like I suspected, you've made it commercial now. 35$ a year for a CTS ripoff with stuff you got for free. small wonder you're insisting on these things. I thought you were better than that. my mistake.

follow the money.

Originally posted by wormwood
well now that you claim you didn't have a CTS handle before the new one, I know you're simply making things up. I can't remember what your original handle was, but I certainly remember when you started using the new one.

which made me wonder a bit, so I popped on your site. and just like I suspected, you've made it commercial now. 35$ a year for a CTS ri ...[text shortened]... ting on these things. I thought you were better than that. my mistake.

follow the money.

it's sometimes a lot easier to say "I was wrong," you know. Plus you wouldn't look like an 8 year old.

Originally posted by heinzkat
Time to raid the forums

Thread 118904

hah, that's almost word for word the discussion we had with richard when he began coding chesstempo, all the way to the +1.75 cutoff value, which then was about exchange sac vs computer's 'positional' evaluation. he didn't get it then, and based on what I've seen today I'd be surprised if he gets it now. some things seem to never change.

Originally posted by philidor position
it's sometimes a lot easier to say "I was wrong," you know. Plus you wouldn't look like an 8 year old.

I'll say it once more fanboy, you don't know what you're talking about.

Hi wormwood,

As I've already stated, these discussions you refer to never occurred. There was no pre-release discussion about chess tempo on CTS, there was some discussion on chess tempo about how to best improve the problems (with some users who were probably also CTS users, you can still read these if you'd like), but that was after the site was released, and didn't occur on CTS. You present zero evidence that anything you say is true, there are over 30K problems on CT, and it is very obvious that they haven't come from the same program that generated the CTS problems. The poor quality of the original problem set should be enough proof of this. On the one hand you are claiming that I've stolen the CTS problem set, on the other hand you are saying that the CT problem set is rubbish, without actually explaining how this contradiction can be resolved. It seems you are not interested in actually presenting any evidence of your claims (which isn't surprising). I've stated the facts, unless you're going to present some actual evidence of your wild claims, there doesn't seem to be much point in continuing the discussion.

Best wishes,
Richard.

Originally posted by richardchesstempo
Hi wormwood, CT is the BEST

Any comments on my laments in the other thread?

Hi heinzkat,

The complaint that some CT problems have 'computerish' moves is a reasonable one. In some positions you are expected to find moves which are quite hard for humans to see, and sometimes the main line will have the opponent throw a major piece to avoid mate, where perhaps it might have been more human to keep the piece and make the player prove they can see the mate (there have been some improvements made to the generator in this respect in the few days, but it will take a while for the entire problem set to reflect these changes). CTS is probably better in this regard than CT, however CTS has other issues, such as positions where you can win a large amount of material with a move that is second best and get marked wrong (CT has second best winning moves, but you are asked to try again instead of failing if you play one). CTS also tends to finish problems early far too often, CT has this issue with some problems also, but it's not as severe as it is at CTS. CT does use some lower rated players games, but most of these are around the 1600 level and a large number of the tactics come from 2000+ games. The position phili posi showed (the first one) appears to be from these two players:
http://ratings.fide.com/card.phtml?event=4633440
and
http://ratings.fide.com/card.phtml?event=1603914
As you can see , neither of these guys are 1300 level patzers (although I don't know what their rating was at the time the game was played, it might have gone up a lot since then).

I'm not saying that CT is perfect by any means, there is certainly still room for improvement , and there are things that CTS does better, however a large number of people seem to find CT useful in helping them improve, and there are at least a significant number of members who consider CT their primary tactics training resource, despite being fully aware of CTS (I give them credit for the glicko idea and link to them in my FAQ).

Regards,
Richard.

I'm up 1-0 against the White player (in a rapid game though years ago), the second one I do not know 😛

What about the Qb8/Qb7 fiasco? That really is an element of CT that bugs me

Originally posted by richardchesstempo
...On the one hand you are claiming that I've stolen the CTS problem set, on the other hand you are saying that the CT problem set is rubbish, without actually explaining how this contradiction can be resolved...

well, one thing I'd like to make sure right now, as it seems you've misunderstood it: taking the CTS problem set was never the problem, I never thought that you stole. you can't steal what's, as far as I know, freely available. at least the CTS creators have never had any problems with people taking the set for their own use, and results of that have always been readily available for anyone interested.

about the 'contradiction' between CTS & CT set quality, I have done nothing but tried to explain the difference for YEARS, which was your recalculated engine solutions, and specifically your blind trust to numerical evaluation in the absence of understanding chess. you obviously still don't get it, which also explains why the problem persists.

as for proving we had our talks, I really don't care enough to go looking for some internet babble from years ago, it's just not worth the time. if someone doesn't believe me, big deal. all that is irrelevant to begin with, and I'm sure the CTS oldtimers remember it anyway.

and still, regardless of our disagreement of what happened, I don't see why anyone who likes CT over CTS shouldn't train there. it's a personal preference. good luck to you and your now commercial site.

hehe, Nice win in that game (it's a pity you weren't the other player in that position shown, that would have been REALLY amusing :-) )

In terms of the Qb7/Qb8, I agree that having to choose between those moves can annoy some people, and it's obviously annoyed you :-) I can also understand the source of the annoyance, either move is winning, the two pawns extra in one line over the other isn't really that interesting given how won that position is (after either of those moves). The Alternatives and 'try again' prompt are designed to try and reduce this annoyance. If you see a move that wins a queen and you are sure it wins then it is completely safe to play that move without further thought. If there WAS a better move then you'll be given another chance to find it (and I agree that 'better' can be controversial). I probably have enough problems in my problem set now that I could start to think about getting rid of some of these, however there are two reasons why I don't. The first is that real games have positions where you need to choose between two good moves, completely cleaning the set of these means you are reducing the number of patterns you are being exposed to (even if the choice itself doesn't matter). The second is that because of the strict ambiguity checks that are in place now it can be hard to produce a large number of high rated problems, so keeping these (as long as the 'try agains' are in place) allows a larger pool of problems to be kept. The 'try agains' in these situations are not to everyone's taste, and you wouldn't be the only one who didn't like it, but at the moment most users are happy with the compromise. Most users agree that the time spent looking for the better move is not wasted as it's just more calculation practice. This is especially true when you play a move which gives you a 'try again' and then fail when you can't find the other move, and I may consider changing that behaviour to be a little less harsh in the future (again CTS would just fail you there, although to be fair their gap from best to second best is a bit larger so the equivalent situation on CTS would be trying to take a rook when the best move wins a queen).

BTW to re-inforce your own point ,that game was:
[Event "It (open)"]
[Site "Banska Stiavnica (Slovakia)"]
[Date "2006.07.12"]
[Round "5"]
[White "Machalova Veronika (SVK)"]
[Black "Kantorik Marian (SVK)"]

And black played the move that didn't win the extra pawns, as I said, this is not one of CT's best problems :-)

Anyway I hope I don't come across as too much of a rabid defender of CT, obviously it's impossible for me to be completely objective, but I try to listen to constructive criticism, as it helps me improve the site .

Regards,
Richard.