# chess online

GameKnot related: Chess Tactics
« Back to forum
FromMessage

Chess Tactics

Tactical exercises #38445, #36576, #37593, #32099(there is 2...Qxh2 3.Rh8#)

The list goes on... How can these questions be tactical exercises? And please don't use
Gameknot game database for questions. How can l guess or know some crap players' moves?
jonjmo
20-Aug-10, 05:04

Just came in to the forum to ask about #36701, doesn't make any sense at all to me.

*spoiler*
Why wouldn't 2. Rc5 lead to Qxc5 from black?
 kingdawar20-Aug-10, 05:18

How can l guess or know some crap players' moves?

You actually have to guess the program's moves. Only the positions from the players are taken. From there the deus ex machina pops up and you have to play out the computer sequence.

GK does not understand this offering positions and then finding of program's "best" moves sequences has little to do with "tactics".

There is not mate with 3.Rh8. Because 3...Qh3.
 kingdawar20-Aug-10, 05:27

It is a pity wrong examples are given
32099 is absolutely correct, 2. ... Qxh2+ 3. Kxh2 Rh8+ 4. Qh3

But 37593... yes what to do after white's fourth move. Everything wins.

36576... not good

38445... again, no good
 kingdawar20-Aug-10, 05:31

jonjmo, after 2. ... Qxc5 white has an in between check with 3. Qa6+.

A good tactics method would have presented you this "most demanding" move. Alas, this one is just one week old, give it some time, this feature only plays the computer's best move.

The played moves in the exercise you give are correct and unique, for now we can be happy with that.
 kingdawar20-Aug-10, 06:39

Here's a more interesting example
34874

So how exactly can this position after 1. Bb1 Rxd2 2. Nxd2 be assessed "objectively"? At a certain point it becomes a matter of preference (subjective) and no hard case can be made for or against any move, as long as it is followed up properly. Here I tried capturing an extra Bishop via 2. ... Qe1+ 3. Nf1 Qxb1 and mate to follow in any case. Of course, it was wrong. And even if it had been right, it wouldn't have been the only way to continue.

So who is ever going to successfully solve this exercise? What will he have learned from it? It's really just a junk variation. The first move is OK, it is clear, you can take the Queen because 2. Bxe4 Rd1+ 3. Ne1 Rxe1#. But after that?
 kingdawar20-Aug-10, 07:05

Dubious tactics
34178

Keep 'em coming
ericaalders
20-Aug-10, 07:34

Deleted by ericaalders on 20-Aug-10, 19:07.
 kingdawar20-Aug-10, 07:40

34880
The whole sequence is flawed?!
ericaalders
20-Aug-10, 08:02

Deleted by ericaalders on 20-Aug-10, 19:07.
 Gameknot.com20-Aug-10, 16:15

I understand that you guys are all trying to help here, but unfortunately simply saying that something is dubious, or bad, or crap, isn't helpful at all. 99% of all complaints about tactical exercises turn out to be the result of the person overlooking something, or simply not understand the goal of these exercises. We'd be glad to looking into each problem, and possibly find an explanation for you, if you could please actually explain what the problem is in each particular case. Please be as detailed as possible.
 Gameknot.com20-Aug-10, 16:39

I can see one of the supposedly bad exercises has already been explained here. I'll try to address another, which heinzkat complained about -- #34874. Moving the pawn allows your Bishop to come out and help your Queen get to the checkmate much faster. Capturing the Bishop on b1 doesn't help in any way, as it is on a white square, and is of no real threat, but it delays the checkmate by at least 3 moves, as far as I can tell. You'll probably say that it doesn't really matter how fast you can checkmate as long as you win, right? But that would be an odd thing to say for heinzkat, who is an absolute King of the puzzles section, where you live or die by a single extra move to checkmate.
 kingdawar20-Aug-10, 17:58

Hey GK
Yes, I already hoped someone would mention that.

It's a matter of stipulation. The stipulation that comes with the mating problems is "mate in x" and you know in advance what to look for, and discard options that fail to fit within the given amount of moves.

The stipulation coming with these new tactics is quite undefined. The stipulation is "make the moves the engine has come up with"; not the "best" moves or the "winning" moves. There is not really any objective definition of what users are supposed to achieve. Just... play out the computer vs. computer moves from a semi-random position. With carefully selected problems this could possibly work out OK - but really, preferably the comp. opponent shouldn't play "best" but instead "most demanding" moves.

Also, there is no real point shown with these mating positions. In the puzzle section, the positions are added because they have a ring to it and it can be conclusively shown that all variations lead to mate, without too much possibility of aberration (although sadly, many add such positions anyway). Opposite to these random positions; nobody really cares much about the computer sequence leading to mate in 8 moves more instead of 10.

It must be said all the mates in 3-4-5 tactics that have been filtered from the GK games, I do think the program has filtered out pretty sound examples - for the mate section, and I would like to see them posted in the mate section by GK_com crediting the games they came from. But with all those mates in over 8 moves vs. mate in 10 or mate in 12 vs. +20; those positions simply burst out of their junctures.

And by the way, maybe you could rename the "Puzzles" section in order to make the difference between the Tactics & Problems more clear. "Mate problems" or "Directmates", something like that. Problemists never like the denigrating term "puzzle" anyway
 Gameknot.com20-Aug-10, 18:49

heinzkat, you said:
"The stipulation coming with these new tactics is quite undefined. The stipulation is "make the moves the engine has come up with"; not the "best" moves or the "winning" moves. There is not really any objective definition of what users are supposed to achieve. Just... play out the computer vs. computer moves from a semi-random position. With carefully selected problems this could possibly work out OK - but really, preferably the comp. opponent shouldn't play "best" but instead "most demanding" moves."

You can't be farther from the truth. The goal for these exercises have been defined: to improve your score/position/advantage in the shortest number of moves. I really don't see how it can be stated any more clearly. No, we don't say beforehand whether you have to checkmate or to gain material advantage, but that's intentional -- otherwise it would make it easier to solve them. It's part of the exercise to figure out which is which. Just like in a real game -- is it the right time to sacrifice your Queen leading to a checkmate or not yet, and you need to keep building up your material advantage?

You certainly don't have to guess "the moves the engine has come up with". That would imply that the engine makes random moves, which isn't true. All moves (except for the very first one of course) are the best moves possible by a wide margin (as you can see from the provided scores for the solution move and the second best move).

Yes, I realize that some of the problems that require you to figure out the difference between a move leading to a mate in 10 vs. a move leading to a mate in 8 are very difficult. Quite possibly, they are out of reach of most humans. But once again, they are no less valid than deciding between mate in 3 vs. mate in 4. They are just very very very difficult problems and they will eventually get a very high rating reflecting that.

Are you simply proposing that we should remove all difficult problems? Or just the ones you failed to solve?   I'm kidding of course, but it does seem that you have a particular vision of what the "Tactics Exercises" should be like, and it seems anything that doesn't fit your vision you label "dubious" or "bogus".

 kingdawar20-Aug-10, 19:07

As I have tried to make clear more often than not, it is not a matter of difficulty or easiness at all that I am concerned about (also in the mate section), it is a matter of soundness. I would like to see more genuinely difficult problems. A lot of the exercises I have solved I didn't approve of either. Because those moves aren't the unique way to win that they are claimed to be.

OK, maybe with the deep mates there is some sort of " tangible evidence" the requested move is better based on well, yes, the difference between forced mate and no forced mate in that amount of moves. But with the positions in which you have to choose between +14 and +10, there is no real difference present at all. Look a few moves deeper in the +10 line and everything on the opposition's side falls just as much as in the +14 sequence. Hmm.

Yes I am maybe too critical and I should take this new feature as it is... I would just like to see the exercises more like the CTS has set this up. There it is do-or-die. Only one move leads to a real advantage; all other moves are clearly inferior. And not because the computer says so; it is clear that one move succeeds and the rest fails.
 Gameknot.com20-Aug-10, 19:52

"But with the positions in which you have to choose between +14 and +10, there is no real difference present at all. Look a few moves deeper in the +10 line and everything on the opposition's side falls just as much as in the +14 sequence."

If that was the case, the chess engine eval would have been +14 for the second move as well. That's the thing with computers -- they are precise and they don't miss anything. If an engine says one move is +14 and another is +10, that's because they are exactly that. Of course to arrive at the same conclusion it might require you to analyze the position 12-20 moves ahead, but that just means it's a very difficult problem, not that it's a "dubious" problem.

"I would just like to see the exercises more like the CTS has set this up."

Ahh, therein lies the problem! You seem to be prejudiced against our tactics problems simply because they are not "like CTS". There is nothing wrong with CTS, but that would be like calling Chinese chess dubious simply because it doesn't use FIDE rules.
 kingdawar20-Aug-10, 19:58

Please don't take it personal. The feedback isn't meant that way.
 easy1921-Aug-10, 00:36

just a small note
so far i have no complaints about the tactical exercises. i like it and i spend much more time on it then i wanted in the first place..

So what would be a improvement for my joy and makes it easier for me to come up with a solution...?

That is actually simple. I want to know what kind of exercise it is..
am i looking for a mate in 3 to 10
am i looking for a tactical sequence with forks pins etc..
am i looking for a big advantage
am i looking for the best moves in a lost position...

I spend most time figuring that out. ( but on the other hand finding that out is a exercise 2 )

Easy19...
... has put his finger on it, I think.

I treat these exercises as though they were from actual games. You begin the exercise as if you had just played a move, and are waiting all agog for your opponent's reply. Having observed that respomse, then the exercise is this: how best to continue?

Now, I can see some difficulty when it's a case of the move you choose is clearly winning, but some other move is 'winninger', and the actual solution to the exercise is 'winningest'.

Back in OTB days, I had a tendency to inefficiency in closing out games, as the following example will show:

w
I'll begin this earlier on than where my real point begins, with Black (my opponent) having offered the exchange of knights on f5.

23.Nxf5 exf5
24.Rg3 ...
Threatening to snip off the h-pawn

24...Kh1?
25.Bxh6! ...
Anyway!

25... Ne6
26.Bxg7+ Kh7
27.Bf6 Nf5
Now, here, with Black about to administer a family check on e2, White decides to play absolutely safe:

28.Rg7+? ...
This wins, of course, but it is nowhere near as good as 28.Rh3+!! Nxh3+ 29.Qxh3+ Kg6
30.Qg3+ Kh3 31.Qg7+ Kh4 32.Qg5#. I even kind of 'saw' that the check on h3 was good, but didn't check it through and just looked for absolute no risk. Immediately after the game I ran though this line and confirmed that it was absolutely risk free. I ought to have been confident of that in play. It was easy enough to calculate.

28...Kh8
29.Rg4+ Kh7
30.Rxf4? ...
Uncompromising, but still a long way second best. 30.Rh4+! Nh5 31.Qg3! Kh3 32.Qg7#. This one was frankly an oversight, having seen that the text move won easily.

30... Kg8
31.Rh4 ...
(1-0 - mate next)
You will note that the third rate sequence I found was no slower than the 'best' - or even the second best - line. It's just that it fell short aesthetically. It lacked elegance.

Now, I will put the win ahead of elegance any day, but still, I've ever since felt a sense of dissatisfaction over my finishing in this game.

That is why I am not altogether on the side of heinskat in this argument with GK, though I can see his point. These exercises are as much a test of imagination as anything else.