Stockfish Ruins Another Chess Puzzle 🤦🏽♂️
Вставка
- Опубліковано 15 лют 2023
- Puzzle FEN:
n3N3/3K3k/6R1/6P1/8/8/8/6r1 w - - 0 1
Puzzle Details:
Gorgiev, 1956
🍒 Most students gain 200+ points in months:
chessvibescourses.thinkific.c...
🐙 Play Chess Here:
www.chess.com/membership?ref_...
Links are affiliate links and help support the Chess Vibes channel via a commission. - Навчання та стиль
So Stockfish was like: "Actually that doesn't work because of this. However, if you do this instead, it does work, so nice puzzle anyway, you just got the wrong solution."
7:10 when you do that the rock would check the king after the king move to what ever you would simply eat the knight then if the king ate your rock then it would be a draw cuz you would also eat the rock :/
@@NotTured No after Re1 Kf7 Rxe8 you don't take back, instead you play Rg6 and black can't avoid mate
@@viktorsmets29 the rock is already at Rg6?
@@viktorsmets29 and even if you did Rg6 which is not possibe i will check you in RF8 you will be forced to eat my rock and ill eat your rock and cong its a drawn end game
@@viktorsmets29 i think i now know what you mean you mean Rg7 check then ill move my king to KH8 forced then you would do Rg6 right?
Stockfish out here delivering patch notes to old chess puzzles
Actual the tablebase did as chess is solved for 8 or less pieces on the board.
Imagine if someone refuted these puzzles before stockfish was invented
Would have been convicted of sorcery
imagine if they were refuted by an actual fish
@@lydiafromsemaphora Specifically one that does stocks
Someone actually made a stock fish, his channel is micheal reeves
I mean, the 7 piece endgame tablebase was computed in August 2012. Yes, Stockfish existed, but it didn't get it's first win as top computer until 2014. All 7 piece puzzles were either proved or refuted by then.
Refutations of old puzzles are the best chess videos. Thanks for posting and please make more
Second this, Nelson
All those dead people who are suddenly wrong. :D
The chess study by Tigran Gorgiev had long been cooked. Every chess study composer knows that ...Nd7 draws. The computer finds something like that in a second.
I found Nd7 because you said the move was crazy, I didn't know why Nd7 was good I just knew it was indeed crazy
Ye same
yep
Me too
i guessed goddamn rg4 cuz thats the most insane
same+1
Stockfish doesn't ruin chess, it finds moves and lines that weren't expected.
Thats why i am a fan of stockfish
and it's a fish
yes, of course it doesnt ruin chess, but it ruined this particular chess puzzle.
You must be fun at parties
@@17813_ he's just stating facts accept or cope
You don't even need Stockfish for this one. There's only 7 pieces on the board, so it's a tablebase position, and the tablebase will tell you right away it's a win for White and tell you every winning move too.
this puzzle was probably made before computers
Puzzle was made in 1956 :/
Aaaah - so it's actually tablebases that are ruining chess puzzles , not Stockfish 🙂
🤓🤓
That works only when you're not playing against Mittens
Actually, there are exactly 7 pieces in the board and we can use a tablebase. After Rg7+ Kh8, white has 3 winning moves: Ke7, Rg6, and Ke6.
(+ Rg6 repeats the previous position, so white has to move the king)
Idk why he keeps making these videos without looking more deeply into it
How you can check tablebase data? Is it same as lichess stockfish server analysis?
@@Alpha-Zero Yes because Lichess have a tablebase too at the analysis board.
The tablebase show you the moves for win, loss and draw. You don't have to activate stockfish for that.
I was thinking about the piece count and was like, hey, do we even need stockfish for this?
Nd7 was actually quite intuitive once you mentioned the stalemating idea.
Same reasoning
Not only Stockfish could find these other solutions. Looking at the 7 pieces Tablebase would have told us the same…….
Thank you so much for the video. That was a really nice solution. I like how Stockfish finds these lines and moves that we didn't notice when originally looking at the puzzle.
I did intuitively guess the knight to D7, based on what you said previously about taking away the pieces of black would lead to a draw. I didn't see it though of course, just since you asked that was the only move I could think would create a situation based on what you said previously
Beginning to feel sorry for the old masters. Computationally they didn't have the bandwidth to go through all permutations in one lifetime!
This was very cool its so much fun to see a refutation and new solution
Really happy that I found Ke7 at the beginning of the video lol. The whole video I was like "but what if you do Ke7 instead?? Pushing the pawn didn't seem best there..." and then you finally went back to that and I was like "yes!!!" haha
Stockfish gives a venomous sting to the puzzle, and stockfish gives the antidote.
Normally I'd say "it's a win for the intended side, just not the intended way" is not really ruining a puzzle, but a pawn checkmate is pretty cool...
I wasn't in the mood to really analyze today, but after you showed the difference between Ke6 and Ke7 (that white can reach the f8 square), Nd7 was the first move that came to my mind. So I found it, but didn't verify it as winning.
Thanks for this study by Gorgiev. It’s not just 2.Ke7 that wins but also 2.Ke6 or even 2.Rg6. So the problem is actually savagely dualised. If one uses Stockfish with 5 parallel searches, I suppose it will pick these up, but easier is to use a tablebase like Syzygy. Question really is how to fix it, if possible.
I haven't played chess in 100 years and I was never more than average. But I love this channel! You really do a nice job at presenting these ideas.
Composer🕵️: Hey everyone! Look at this chess puzzle I made! It is white to play and win and ends up having a "smothered mate" using a pawn♟️
**Stockfish joined the game*
Stockfish🐟: You sure about that? Checkout this knight🐎 sacrifice to draw the game.
Composer🕵️: Oh-- **sad noises*
Stockfish🐟: Don't worry, there is another move for white to win the game, just use and move the king🤴 already closer to the black's.
Composer🕵️: --oh ok
this pesky fish is so smart lol
always interesting to see the solutions to a problem
I actually saw the Rg7 Kh8 Ke7 line and assumed that was the intended solution.
The mating pattern with the knight on f6 and the king on h8 is too strong to avoid, you have to sac the rook and rook + 1 vs knight in the corner is an easy win
I actually saw the move stockfish had for the right reason. Im proud of myself
Nice one! I was able to guess knight d7, because you showed lines where whites king wants to move to f8, and that blacks knight is untouchable because of the stalemate trap.
Saw knight to d7 the first time you showed the position and was like "thats what I would do, to force a draw"
I found the king move instead of the main line at the start - lol. The smothered pawn mate was cool, though.
I did see ... Nd7 but only after understanding white's ideas about getting to Kf8. I wouldn't have found that move otherwise. I did originally see Rg7+ Kh8, Ke7 because I felt the king needed to support the rook as the knight and pawn weren't able to do so, but I couldn't see the end of that line. This is probably why I'm still rank 900 and why I enjoy your puzzles and your other videos that help me see play evolutions.
I saw knight to d7 pretty fast when you asked for the move. Prior you showed that if black loses it's pieces it's a stalemate, so yeah, makes sense that getting rid of your pieces asap was the way to go.
Where are u getting these puzzles
where can I find a record of the original puzzle?
I thought of king e7 already before pawn g6, feeling like we should move the king to safety from opponent's knight and protect own rook for checkmating. I didn't think of anything else further though.
The puzzle and the refutation are both amazing but I did want to note that he didn't explain that after Ke7 Re1+ Kf7 he did not comment on Rxe8!? removing the defense of the rook but white has Rg6! saving the rook with tempo of mate and attacking the rook. Still it was a fun episode.
As far as the refutation I missed Nd7 and thought it was Nc8+?, the idea is that if white tries his Kf8 idea Rf1+ Rf7 Rxf7!? you cant take with the king because of stalemate, so gxf7 Kh7 and I thought the white king cant move from in front of the pawn because the knight on c8 covers e7 and white's knight blocks e8. Additionally if white plays Nc7 black can play Nd6 taking the pawn before white can move it and insufficient material. Finally I thought if white plays Nf6+ Kh8 and if Ke8?? Nd6+ forks the king and the pawn, but of course the fun is spoiled again as white can play Ne4 covering d6 allowing Ke8 and promotion. Although it is an interesting thought if it was white's turn to move when the black king started on h7 after the trade of rooks it would be a draw because of zugzwang after Kh8. And you just draw with one tempo.
for once stockfish agrees with me, Ke7 make more sense then g6
Did not find Nd7 although it does make sense blocking the f8 square
Great puzzle and good job stockfish.
I found the move that draws the game but I only saw the idea of taking away the f8 square for the King and to lure the King away
this reminds me of the SF refute of the famous Tal chess puzzle!
At 7:19, would rook takes knight be possible? If you take the rook then the black king takes your rook and pawn
Stockfish: Your solution is wrong, but you can still win it.
i found the move cause i though "ok getting to that squares important, how can we block it" and the knight is what i saw first lol
Nc7 was actually my guess for the fish move, but I didn't have the whole line worked out. It seems so clear once it's explained. Is there partial credit in chess? 😂
Thanks Nelson!
Did find Nd7 based on what you showed us earlier happens if the knight is taken and the fact that it denies the f8 square
The solution for stockfish after the position on 8:13 starts with Nxf6 (has to take because of the threat of Rg8#), then Rxf6 (If the pawn takes on F6 instead of the Rook, since your king is blocking the pawn and the pawn can't control the H6 or H7 squares, Black can more likely save, or at least prolong, the game by checking you with their rook and forcing your king to block the pawn, but if Rxf6, White wins faster). The black rook has to go to h1 to defend Rh6#, then you play Rf4 to keep threatening checkmate on H4 but giving your pawn and king more breathing room. If Rh7+, it's forced mate in 20, with a possible sequence being something like this: Kf8, Rook somewhere along the 7th rank, like Ra7, Rh4+, Rh7 is forced, Rxh7, Kxh7, Wkf7, Bkh8, Wkg6, Bkg8, Wkh6, Bkf7, g6+, Bkf8, Wkh7, Bke7, g7, Bkd6, g8=Q, Bkc5, Qb3, Bkd4, Wkg6, Bkc5, Wkf5, Bkd4, Qb4+, Bkd5, Wkf6, Bkc6, Wke6, Bkc7, Qb5, Bkc8, Wkd6, Bkd8, and finally Qd7 or Qb8 is mate.
I love it when the moves I think about are the correct answer, when the composer had an entire other solution.
I found the move that refutes it, but I don't feel like I can really take credit. Just from your explanation of the intended line I knew it had to be that move.
You emphasized both that the square the knife is controlling was important, and that taking the knight drew the game, so that move immediately jumped out to me when you told us the position the move was in.
I was so surprised, I kept thinking it was king e7 at the start and then it was haha
Couldn't find it; was thinking I could suicide the rook into the pawn but it never worked out. Very interesting. Thanks for sharing this!
Even given that the author didn't see the refutation, I wonder how they concluded that the king move was not also winning (that one that we know was in fact the only winning move).
Composer: take that I win
Stockfish: Yes, but actually NO
7:10 I was thinking Ke7 the whole game because i thought that he could repetitive check and/or sac a knight and rook to get stalemate
5:30 Found! It was not that hard because basically the knight has only 2 places he can go which allow him to get active and defend the position: d5 and d7. d5 has already been analyzed. So it remains d7 and moving the rook but basically there are also only two options for the rook: checking the white king which has already been analyzed too and taking on g6 but after rook take rook you understand it's just a useless sacrifice. So, knight to d7 is the only valid option that has not been analyzed yet in this position. However, I would have never found such a moove if i had not been told it exist.
I found Nd7 almost immediately, actually when you were showing the winning line for the first time at around 2:46
I always wonder why the move Rf1 doesn't played to cut off the king into going to the f file in many position. Do you guys have any answer?
2. (g6 Nb6+ 3. Ke7 Nd5+ 4. Kf7 Nf6 5. Rg8+) mate in 4 for white
"Stockfish ruins another chess puzzle"
Alternative title: "Stockfish mourns Mittens in the only way it ever can"
Got a challenge for you. Pacifist chess. Can you beat Martin without ever capturing any pieces? (Not sure if you’ve done this, but sounds like it would be fun to watch)
Nd7 was easy to find
There is an interesting line after 1. Rg7+ Kh8 2. Ke7 Re1+ 3. Kf7. Black plays 3. ... Rxe8! white can't take the rook 4. Rg6! with Rh6# threat actually wins the game.
5:44 What if White walks up to the Black Rook and lets it check him but drags the rook back to a square where the Knight can take it and the King or Knight don't protect the White Rook but can protect the pawn after the Black King takes the White Rook? Might that be a way out of the stalemate trap?
Nd7. Same as ke6 nd5. N holds f8 f6, can be taken
5:28 I've found it, but only after i was told to search for it
I don't get why the nxd7 move was so special. Move the rook to skewer the knight with rxe1and if the king takes it, it's a draw since you lose the rook in the process, with the king having position on the pawn, as well as their knight still, too. If the king doesn't take then you've got a free piece and advantage, potentially even a win for black. it's surprising that nobody noticed that... i'd love to know how a king/pawn/rook manages to beat a king/rook/knight if that is actually how it plays out from here.
Isn't that within the capacity of a tablebase? AIUI, all positions wit 8 or fewer pieces have been solved.
1:14 actually an easier thing for white would have been trying the Arabian Mate. Nf6 and Rg8 would have been mate, if we protected the rook with the king
i directly saw Rg7 but couldn't find a good follow up.
As you showed the pawn moving up i instinctively hated that move it feels useless to constrict a King that is already in the Corner and can't move and stockfish proves you don't need it (and it's even wrong if black is strong enough).
I think the Composer started at the finale with the rook sacrifice and tried to get ppl there.
If "I bet you're great at parties" was a person
When I first seen it thought win for white. Pretty much always right. When see an over all of board. Whether it’s a win for white, black or draw.
But don’t and struggle to fine winning move. Was thinking Rock to F7, but more a think about it. King to E7 would make more sense
6:56
Puzzle author: What do you think of my super challenging puzzle?
Stockfish: Your solution is wrong.
I found Nd7, but only because you said it was a crazy move and it was the first crazy move I could think of
And in the Stockfish line, after R31+, there is Kf7 RxN, Rg6 winning.
I found Knight to d7 but simply because you said that Stockfish found a crazy move.
What if from the position at 1:54:
black rook to e1+
The point is if the white king stays next to its Knight the black rook can take the knight, and if the king takes the rook then the other king can take the white rook and its a draw. It'd be a draw because the white king can't defend the pawn and so the black king would just take the pawn on the next turn leaving just the lone white king against the black king and knight.
I don't know what would happen if the white king doesn't take the rook.
...Re1+, Kf7 Rxe8, Rh7 mate.
The only reason I found Nd7 was because I was just looking for the most ridiculous move. Lol😂
😂
i never wanted to move the pawn in the first place
How did i find the true winning move for white before he even showed the rest of the possibilities, cause i thought why move the pawn its guarded by the rook lets bring in the king to prevent the check instead
I'm not good at chess but I actually found the move. I was like wait what?
I must be a robot cuz that king move was my first thought
I found the stockfish move, not cause im good at chess, more cause im good at throwing pieces away
Is it only stockfish or woildn't any old mephisto with 16 half-moves depth would find the solution?
Curiously I did not see the flaw that Stockfish saw, but my first instinct was the Stockfish sanctioned solution.
Great channel
Whats crazy is that i didnt see the pawn raise move, i thought the best move was to move the king
At 7:21 instead of R f1 check, what if R takes Knight on e8?
After Rxe8, White has Rg6, with the idea Rh6X.
So stock fish found the actual solution for the puzzle.
I seen that the pawn move was incorrect when the video started
Black is also drawn in the end of the main line as long as they don't play nf6
hide the evel bar
it is small on the right of the bord
it kills the magic
I actually found knight d7.
Great puzzle!
is it possible to beat martin with only Book move, Great move, and Brilliant move?
When someone tells you “There’s a better move” then it’s pretty easy to spot it.
that's one of the problems with puzzles the solutions often times are way easier to see when you know theirs a solution or crazy play when usually in a game the hardest part is knowing when their is a crazy play more so then finding it.
in 7:17 is there rxe8 taking the knight?
In the "improved" version what happens if rook sacks itself for knight? That looks totally drawn
that's what I'm wondering. seems to me that everybody is missing the easiest possible solution. the only way it could get weird is if the king actually didn't take the rook after losing the knight.. almost certainly a win for black or draw for white, at best.
Feel really smug since I wanted ke7 from the start.
1 question... Why @7:15... Rook can give a check from "e rank" and capture white knight.
If king captures rook, black captures white rook.
If pawn push or white king move... Rook keep giving check to white
...Re1+, Kf7 Rxe8, Rg6! threatens Rh6 mate which Black can't stop without losing the rook.
oh! Does it count as seeing it if I correctly guessed what it was, but didn't think through enough to see why? I mean, that wouldn't cut it in a real chess game with no hints that this is a puzzle to solve, so I suppose not.
I found it watching chess puzzle make my 700 elo brain sel turn like stock fish lol
Even before stockfish, I thought they had 7 piece endgame tables? Just look it up in the table to verify if the composer's solution worked/if there exist other solutions.
I got the knight to d7, my idea is just trying to sacrifice and get a stalemate, and it was actually correct :), lucky me
knight d7 crossed my mind but i thought it was stupid lol
At 7:19 can't black go re8 capturing the knight, and if white goes ke8 then black can go kg7 capturing the white rook?
After which white can't defend the pawn leading into black successfully getting a draw.
same bro i was wondering
i tried calculating if its still winning after white plays g6 but we can sacrifice the rook again for a second time by playing Re7! after which white has to take and black takes the rook which is obviously a draw
7:09 that was the move I was originally going for. More luck than skill, but I'll take it
So that means I found the stockfish AND the intended move!!!!
I agreed with the composer, which means stockfish schooled me too. I didn't find Nd7. I used Re1 check to capture the knight. If white king takes black rook then the Black King captures the white rook and the pawn. If white declines the sacrifice as they should, then black can repeat rook checks to get the desired rook trade in multiple ways. Edit: This doesn't work for trapping the white rook if you use stockfish strategy of leaving the pawn at g5. Still, the loss of the knight removes the easy checkmate options.