Tabiyas schrieb:
I downloaded the mostly functional COW 27 days ago - you get 30 days of the middle version as a 'trial'.

Observations:

COW:
I prefer how you can pull up individual openings to train compared to CPT. CPT's is more cumbersome.

It seems to load my GIANT group of games ( A00 thru A49 as I recall) faster than CPT...actually, CPT crashed and I could not load them and gave up...

I really like COW's 'flash card' implementation BETTER than CPTs...like how each new position moves around on the screen a bit instead of staying in the same place...makes it seem like completely new position (which it is!) compared to the board staying in the same place and the pieces just shuffling around. BTW, you can't actually do the training until you buy and register the product so some of what I say is based on the videos, not experience.


All that said...the version worth having is the (I think) $169 version and that is just WAY TOO EXPENSIVE!!

Mike at COW seems to update his program a lot more often than Stefan...who hibernates for a year or more at a time, making no improvements or tweaks at all. Being a subscriber...that irks the crap out of me. Sorry.
That said, maybe CPT is more 'stable' than COW (?) looking at the various updates COW has had.

If COW were say $59, I would get it...and try it out more. I might even like it more.

But, that said...I really think I like CPT better! It looks better. The training is (seems to be...see above) on balance a bit better (except the flash card item I mentioned). The comments area is better. I like how you can so easily reposition the different panes. But mostly...I guess I'm just use to it more.

That's perhaps not a 'ringing endorsement' of CPT over COW. So be it.

So far I haven't seen a reply post about how to make CPT help you to understand moves. I realized that "training" and "learning" are two totally different concepts. I could be wrong but I don't really believe in "training" a repertoire since you would be memorizing the moves and completely not understanding the reasons why your designated move is correct. Also, if you are "training" and get the move incorrect CPT does not provide the reason why your move is incorrect.

I think what I wrote previously about having a three function idea that doesn't make that much sense and is too difficult to implement.Plus it doesn't help you to understand the why when choosing your candidate move or the why when your candidate move is incorrect, which is very important.

I think CPT needs to be geared toward "learning" the opening. When we are "learning" the opening for the first time we first try to remember the ideas behind the opening. That way we know what to do in the opening. To make CPT better than COW Stefan could create several CPT ebooks not only showing all the moves of each opening, but also explaining the ideas behind each and every opening. I don't think COW explains the ideas behind the openings. I think in the future Stefan wanted to create CPT ebooks anyways so this would be the perfect time to do it! I know this requires a lot of work, but I guess you could have volunteers who are willing to do this task. Personally, I would be willing to volunteer to help Stefan with this task if he is interested in creating CPT ebooks. One thing to consider would be how many moves deep should the ebook opening should go. Another thing to consider is if the opening ebook should be free or be provided with CPT.

When we talk about "training" I think Stefan needs to include a separate functionality where you can play out the position against the engine if you get a move wrong. We have to define when a move is wrong/incorrect which CPT currently doesn't do that well since CPT says any move that is not in your repertoire is incorrect. Any move that is typically not an ebook line move or is negatively evaluated by the engine should be considered incorrect if you want to "learn" the opening. If we define incorrect this way then it seems that there would be no point in making an opening repertoire. Personally, I don't really believe in an opening repertoire because there are several moves that you could be playing that are correct but not the best moves and CPT will automatically count these wrong and give you a headache. The repertoire training feature can be kept if you really want to memorize moves, but I don't really understand the use for it. The reason why "repertoire training" doesn't really work is because it doesn't take into account move order and more than one candidate move in the position that could still be ok.

Personally, I want to "learn" openings, which might be possible to do with this program. What do people think about this?

You asked for what people think…so…for all I am about to say, do realize that I actually do like that you put forth ideas!

But, I am largely going to have to disagree with you. The name is Chess Position Trainer…it is not a learning program like say Fritz ‘sort of’ can be.

You speak of‘playing out a position against an engine’…I thought of this too, but is that really going to be of benefit? I mean, with engines being 3000+ elo strength and getting stronger, isn’t that rather a pointless exercise? How would it show your alternative move to be of any real benefit as – unless you are a GM - it will beat you like a drum before very many moves? What makes you think your next few moves are going to be good follow ups? Dumbing down an engine to only see, say 5 ply deep, might have a point but then if you beat the dumbed down engine, does that actually mean your alternative was really any good as the engine has been weakened?

Perhaps simply being able to switch on a powerful engine within training mode? It would show a line (or multi-pv) many ply deep and you can use it to quickly see (if as I mention below, you could tab thru the line) if your alternate is even playable. That would seem like an easy solution – to let the engine give a quick verdict of your alternate move.

Stefan making e-books? We know that it takes him ages to make updates to the program, do you really think he wants to be writing e-books, which will go out of date…and which may not be for lines you like to play?

I think you might could subscribe to Chesspublishing.com. They have oodles of ebooks by actual experts in each opening. Use those and tweak them to your hearts content within CPT. I do!

“The repertoire training feature can be kept in if you really want to memorize moves, but I don’t really understand the use for it”. I really can’t believe you said that! That is THE primary purpose of CPT! Please don't think me harsh, but as much as I like reading your ideas, I sometimes I think think you might consider a totally different training tool. Please don’t ask Stefan to change the very concept of CPT!

I really think one should just work on their repertoire outside of training sessions. With some suggestions I've given, one would be able to do that more easily within CPT! Doing so, you can incorporate any new ideas into your repertoire only after you have decided am alternate line is any good. I take great pains to find suboptimal lines and remove those branches from ‘my’ repertoire, and use the engines to also look for ‘weak’ moves my opponent might play that I can take advantage of - things I might not spot Over The Board (OTB), but because I spotted it in CPT, I am ready should I see that in a real OTB game!

Me, I would like to see mostly what I think might be simple things to implement: being able to hit the ‘x’ key as I can while an engine is running within Chessbase to make it ‘pass’ a half move in the current position so it can tell me what threats are there which I (or my opponent) might not see. That would also be great for editing lines…and give you an idea of what comment to append, like ‘I have to play Nge2 because black is threatening to sac on c3!’. If you have the engine 'pass', you can have it show you those threats!

Another would be to be able to ‘freeze the analysis position’ and have buttons below it so that you could play out the line the engine is showing. Other programs do this. Right now I have to actually put the line into CPT to see what the engine is aiming for...then go back and delete them if they are not to my taste.

Another…changing the ‘flash card concept’ to be random and a bit like COW…nothing proprietary there and it is a great COW feature…even the ‘cards’ move around slightly from one to the next, better than simply having a stationary board where the position changes. But that’s a minor thing.

Thanks for listening,
Brian

Hi Fischer2008
Hi Leavenfish
(sorry for my bad English)

I dont want to make a debate. But as it was asked for I must say I agree with Leavenfish. Dont make change the concept of CPT.

Yes Fischer2008 your ideas are intresting. And I must say that what you said about diffrence between Learning and Training is exactly what I realised or re-realised just a week ago. I have started to use Spaced Repetition in ChessTempo.com and I frequently must learn endgame and some combination motives outside that program. Same here. Chessbase now have simular Opening Training but still lacks the Spaced Repetition and Schedule. So Learning I Think as Leavenfish best outside CPT. Training or rather memory is what CPT are for.

I got CPT for one reason and one reason only and thats the Spaced Repetitions Schedule. I Think that is as good as the Idea of Saving Positions as important. Thats the soul of the program.

But a function Fischer2008 (or was it someone else?) that when you make a move that aint in Repertoire it would be great with some directly feedback controlling your move against the computer and it says something like "bad move, hang your pieces" or "an ok move, but not in repertoire" or simular.

Still I kind of agree with Fischer2008 as I Think what his main critisism is it that the program are hard to use. Yes its hard to use. One must create ones repertoire. Thats hard. Really hard. And every help is needed. And here I got an idea. Maybe the A00-E99 idea or simular is kind of a simular idea. Ok here, is my idea:

1) Have one CPT-book in the program as default. And that should have instructions of why moves and variations are played. Lika a book as New in Chess´s Chess Opening Essentials 1, 2, 3, & 4 by www.newinchess.com/catalog/product/view/...-essentials-1-2-3-4/

2) Instead of selecting repertoir (one as black; one as White) by selecting moves. One maybe can do it the other way around by selecting move to remove.

In that way It would be easier I figure to get started. I remove all but 1.c4, 1.d4 and 1.Nf3. And as black I select 1.c6 against 1.e4 and the Slav against 1.d4 just as an example.

As I remember Stefan was not against making CPT-books to sell. Yes hi had not done it yet and nobody else has done it. Yes the program is great but I Think many is like me having a hard time getting started. Its like we would have Chessbase program but that the Company doesnt sell databases. The idea of Chessbase is the database. Its in the name.

Here we have Chess Position Trainer, but no database with Positions. Maybe its good like a barrier against people (like me) that had not yet created theire repertoire. For as long as the repertoire brings better positions (or save time) against unprepered opponents as long we have a reason to use the program. The day everybody has it we are doomed and have to look at other ways to keep a cutting edge.

Last The moving around of the board on screen like in COW I also would Think should be implemented. As logical as board sounds.

Now everybody have a nice summer!

Tabiyas schrieb:
Mostly what I do is import an already well thought out repertoire by an expert. My very first repertoire was the pgn file from Everyman Publishing, a Sicilian Kan book by Hellsten. Most recently I made one based on a new book on the English opening, Killer Dutch and Smerdon's Scandanavian. Again, Chesspublishing.com is great for all sorts of well thought out repertoires.

Before I do any training, I go in and tweak to my hearts content - sometimes throwing out various lines. Usually though, I ADD LINES repertoire books leave out...usually early lines designed to avoid your repertoire.

An example: The reason I got the Kan book was actually that I had played the Sicilian O'Kelly for years with great results: 1. e4 c5, 2.Nf3 a6. Here, if someone plays 3.d4 you are in great shape. But what if they play 3.c4? I discovered playing 3. e6 was very good as when white then plays 4.d4, I can play a Kan...with the kicker that it is a sideline in the main Kan which is less good for White and something an Open Sicilian player would not normally play against a straight Kan! I've seen Kamsky do this as Black. So, now I have a more 'homemade' O'Kelly teamed up with an experts Kan for a nice fully thought out repertoire against the Open Sicilan!

Is it hard...well, to me no, to someone else maybe it could be. The manual though is excellent. But it is also quite fun building and maintaining (then training) a well thought out repertoire. Currently I only play OTB, but figure when I return to 'Correspondence Chess', CPT will be invaluable.
Tabiyas schrieb:
I don't know what if anything Stefan will end up doing, but perhaps a fairly simple solution would be to be able to 'switch on' an analysis engine within training - you can already do that while just 'stepping thru' your repertoire (as I actually do more than use training mode, because I like to edit my repertoire), to see if the engine evaluation is indicative of how 'bad' your move might be. If it were not 'bad'(one should be able to self-define if say going from 0.00 to -.33 or -2.25 is 'bad enough')and if it were not too bad, give you the option to flag that move to go back in and investigate later outside of training.

As I said, you can already do that...just not within training mode. For me, training mode is essentially about REMEMBERING what I have already decided is the best course. At 53, my memory tends to fade, thus I have CPT!

Anyway, I hope my use of CPT and input is helpful to someone.

Thanks for your input Leavenfish and Tabiyas.

I wanted to ask, how would you say you remember openings? Do you rely on CPT to memorize the moves? If so, then don't you get annoyed when you think you play an ok alternative move during training but CPT counts this automatically wrong since it is not in your repertoire? The move could still be a playable move. Worst off you don't know why your alternative move is incorrect unless you play out the position with the aid of an engine I think. Yes the issue maybe that the engine is very strong and doesn't explain to you where you go wrong in all the variations. Any thoughts on how to improve this specifically?

When you are making a repertoire for training and rule out "other moves" that you have included in your repertoire don't you still have a tendency to play these other moves when you are training your repertoire? How do you unlearn these tendencies? Simply repeating training over and over does not seem like a very efficient process.

Leavenfish you mentioned using a totally different program other than CPT to "learn" openings. I don't think there is any that is why I wanted to make suggestions for CPT. Fritz currently is not really a "learning" program either since it only lets you play against the engine and you can create an opening repertoire by playing against the computer. Besides I thought Stefan wanted to make an opening database or CPT ebooks in the future. If he does then it would make sense to somehow be able to "learn" the openings through a learning program. Learning would be treated differently than training since you wouldn't have an opening repertoire but could make one later if you wanted to.

When creating a repertoire for learning (not training) it would be great if CPT recognized moves such as Tabiyas mentioned and I previously mentioned as ok moves, not in repertoire moves, and incorrect moves. Ok moves would be anything that is in an opening database. Incorrect moves would be anything not in the opening database or negatively evaluated by an engine (like chesshero does).