вот это я так понимаю что это за зверь такой же как и в розницу по поводу 💋💋💋💋💋💋💋💋💋💋👄👄👄👄👄✋✋✋✌👆👎☝👍👇💪🏃🏃💪💪💪💪💪
Can there be more than 1 solution to a puzzle?
Sort:
josi_10 пишет:
josi_10 пишет: вот это я таклллллл понимаю что это за зверь такой же как и в розницу по поводу 💋💋💋💋💋💋💋💋💋💋👄👄👄👄👄✋✋✋✌👆👎☝👍👇💪🏃🏃💪💪💪💪💪
вот это я так понимаю что это за зверь такой же как и в розницу по поводу 💋💋💋💋💋💋💋💋💋💋👄👄👄👄👄✋✋✋✌👆👎☝👍👇💪🏃🏃💪💪💪💪💪
forked_again пишет:
forked_again пишет: Numquam wrote: The tactics trainer relies heavily on the evaluation of an engine. There are lots of problems with alternative clear wins. Some line which is evaluated as +2 and an easy win for humans can still be wrong according to the tactics trainer, because the 'best' line is evaluated as +6. That doesn't mean the move is better, only according to the engine. A move which wins faster isn't better. You don't get extra points for winning in less moves. You only need to see a clear win and try to avoid complications if you are winning. This is also the reason self-learning engines may go for slower wins.Exceptions are moves which waste tempos or lead to repeating the same position. I agree such moves should not be correct.All good points. Engine "best moves" often involve a sacrifice for a positional win. It is easy for a human to miscalculate something like that. If you see how you can increase a material advantage and eventually force a checkmate, then why look for anything faster that ighhhbbbhn to be a good time ⌚ I am not sure if I can get a free to cal I am not sure what to expect from you soon 🔜 to get a 😗😗😗😗😗
Numquam wrote:
The tactics trainer relies heavily on the evaluation of an engine. There are lots of problems with alternative clear wins. Some line which is evaluated as +2 and an easy win for humans can still be wrong according to the tactics trainer, because the 'best' line is evaluated as +6. That doesn't mean the move is better, only according to the engine. A move which wins faster isn't better. You don't get extra points for winning in less moves. You only need to see a clear win and try to avoid complications if you are winning. This is also the reason self-learning engines may go for slower wins.
Exceptions are moves which waste tempos or lead to repeating the same position. I agree such moves should not be correct.
All good points.
Engine "best moves" often involve a sacrifice for a positional win. It is easy for a human to miscalculate something like that. If you see how you can increase a material advantage and eventually force a checkmate, then why look for anything faster that is riskier?
In this position with white to move, you might think that capturing the queen is the correct move. But it isn't. That's because there is a better move.
EDIT: Put a white king.
Feb 19, 2019
0
#26
TitanChess666 wrote:
Chess.com tactics usually have more than one winning line. I think it's unfair to have people lose points for finding a winning, but not best, line. Thus I wonder, do chess.com tactis have More than one solution to them?
The point is to find the BEST winning moves and I'm ok with only getting points for seeing them. What drives me crazy is when I get a ton of points for an easy mate-in-two but then am penalized huge points because I made the wrong 3rd move, seeing a 6 move win instead of the "best" five move win.
Trexler3241 schreef:
In this position with white to move, you might think that capturing the queen is the correct move. But it isn't. That's because there is a better move.
In my opinion that is a bad problem even though I can see the mate in like 2 s. Better examples can be found in high rated tactics problems. Queen+knight (and king) win very easily vs Rook and 2 pawns and king. It would be more difficult if you lose knight.
And please also put a king somewhere 
Trexler3241 wrote:
In this position with white to move, you might think that capturing the queen is the correct move. But it isn't. That's because there is a better move.
EDIT: Put a white king.
Great example of the type of horrible problem chess.com has.
Feb 19, 2019
0
#29
Numquam wrote:
Trexler3241 schreef:
In this position with white to move, you might think that capturing the queen is the correct move. But it isn't. That's because there is a better move.
In my opinion that is a bad problem even though I can see the mate in like 2 s. Better examples can be found in high rated tactics problems. Queen+knight (and king) win very easily vs Rook and 2 pawns and king. It would be more difficult if you lose knight.
And please also put a king somewhere
You can take the queen and not lose the knight.
forked_again schreef:
Numquam wrote:
Trexler3241 schreef:
In this position with white to move, you might think that capturing the queen is the correct move. But it isn't. That's because there is a better move.
In my opinion that is a bad problem even though I can see the mate in like 2 s. Better examples can be found in high rated tactics problems. Queen+knight (and king) win very easily vs Rook and 2 pawns and king. It would be more difficult if you lose knight.
And please also put a king somewhere
You can take the queen and not lose the knight.
Exactly the problem requires some serious modification, because as I said queen+knight win too easily.
This is a good tactics problem and you can see that after the incorrect move, the position isn't winning, but a draw.
Feb 19, 2019
0
#31
Numquam wrote:
forked_again schreef:
Numquam wrote:
Trexler3241 schreef:
In this position with white to move, you might think that capturing the queen is the correct move. But it isn't. That's because there is a better move.
In my opinion that is a bad problem even though I can see the mate in like 2 s. Better examples can be found in high rated tactics problems. Queen+knight (and king) win very easily vs Rook and 2 pawns and king. It would be more difficult if you lose knight.
And please also put a king somewhere
You can take the queen and not lose the knight.
Exactly the problem requires some serious modification, because as I said queen+knight win too easily.
This is a good tactics problem and you can see that after the incorrect move, the position isn't winning, but a draw.
In your new position you can still force the smothered mate.
forked_again schreef:
Numquam wrote:
forked_again schreef:
Numquam wrote:
Trexler3241 schreef:
In this position with white to move, you might think that capturing the queen is the correct move. But it isn't. That's because there is a better move.
In my opinion that is a bad problem even though I can see the mate in like 2 s. Better examples can be found in high rated tactics problems. Queen+knight (and king) win very easily vs Rook and 2 pawns and king. It would be more difficult if you lose knight.
And please also put a king somewhere
You can take the queen and not lose the knight.
Exactly the problem requires some serious modification, because as I said queen+knight win too easily.
This is a good tactics problem and you can see that after the incorrect move, the position isn't winning, but a draw.
In your new position you can still force the smothered mate.
Yea, the smothered mate is the ONLY winning move in my position and that is exactly the point. Nxe8 draws. That is the difference between a bad tactics problem like yours and a good one.
Feb 19, 2019
0
#33
Numquam wrote:
forked_again schreef:
Numquam wrote:
forked_again schreef:
Numquam wrote:
Trexler3241 schreef:
In this position with white to move, you might think that capturing the queen is the correct move. But it isn't. That's because there is a better move.
In my opinion that is a bad problem even though I can see the mate in like 2 s. Better examples can be found in high rated tactics problems. Queen+knight (and king) win very easily vs Rook and 2 pawns and king. It would be more difficult if you lose knight.
And please also put a king somewhere
You can take the queen and not lose the knight.
Exactly the problem requires some serious modification, because as I said queen+knight win too easily.
This is a good tactics problem and you can see that after the incorrect move, the position isn't winning, but a draw.
In your new position you can still force the smothered mate.
Yea, the smothered mate is the ONLY winning move in my position and that is exactly the point. Nxe8 draws. That is the difference between a bad tactics problem like yours and a good one.
It wasn't my bad tactics problem LOL!
But chess.com is full of tactics problems like that. Guess the correct mate in 3 and you get 8 points. Pick a more clear mate in 5 and you get minus 14 points!
Feb 19, 2019
0
#34
Of course there are sometimes many variations to a win. I had some rude troll argue with me about chess tactics on a particular puzzle, who didn't seem to understand anything of the point I was trying to make. See comments in: https://www.chess.com/puzzles/problem/36271
A lot of tactics are "Engine based" so there is only one "perfect move," Despite there being sometimes many winning moves. Sometimes it also could be, that the puzzle wants you for find a particular tactic/technique instead of a mere good move.
Checkmates come to my head when you say more than one solution. For example, if you're checkmating with a king and a queen, sometimes the computer analysis that there is a better move even no matter what (if you know your checkmates) it will be a mate in 5.
Feb 19, 2019
0
#36
I think the analysis engine is good, but I usually get -10 if I get 2/3 problems. That really annoys me.
مرگ بر آمریکا
💩
Tactics trainer gave me an example of a bad problem today. While I found the solution, I first found an alternative solution which I most likely would have played in a real game, because it is a clear win. The only reason I didn't go for it is because I knew tactics trainer likes solutions where you win a full rook rather than go for an endgame where you win pawns, because your opponent's rooks are too passive.
https://www.chess.com/puzzles/problem/38886
The alternative solution is Qa6 Qe1+ Rxe1 bxa6 when a5 is the most convincing win. (Re8+ also wins)
Qe1+ is forced, the queen can't be taken due to Rb8-Rb7 checkmate
The tactics trainer relies heavily on the evaluation of an engine. There are lots of problems with alternative clear wins. Some line which is evaluated as +2 and an easy win for humans can still be wrong according to the tactics trainer, because the 'best' line is evaluated as +6. That doesn't mean the move is better, only according to the engine. A move which wins faster isn't better. You don't get extra points for winning in less moves. You only need to see a clear win and try to avoid complications if you are winning. This is also the reason self-learning engines may go for slower wins.
Exceptions are moves which waste tempos or lead to repeating the same position. I agree such moves should not be correct.
All good points.
Engine "best moves" often involve a sacrifice for a positional win. It is easy for a human to miscalculate something like that. If you see how you can increase a material advantage and eventually force a checkmate, then why look for anything faster that is riskier?