White to move. How should White proceed?
1B6/1p3p2/8/P4N2/k4P1p/P6p/2K5/8 w – – 0 1
Bernhardt 1961
Daily Chess Improvement: A Classic Chess Tactic
About Author
White to move. How should White proceed?
1B6/1p3p2/8/P4N2/k4P1p/P6p/2K5/8 w – – 0 1
Bernhardt 1961
Beautiful!
1. a6!!.bxa6 forced. (If …,h2?. 2. a7!. h=Q. 3. a8=Q+ will win for white). (Similarly, b6/5. 2. a7!. And 3. a8=Q covers h1 square and white wins.
2. Nd6!!! Most difficult move to find. h2 what else. BK is practically immobile. If …Ka5. 3. Bc7+ drives back the BK to a4 and black has lost a move making it easier for white with the given continuation. Moreover if 2….Kxa3. 3. Ne4 and 4. Nf2 stops black h pawn from queening.
3. Ne4!!!. Threatening mate. Kg5.
4. Nf2!!! And white stops h pawns from queening and wins.
If 3……h1=Q. 4. Nc3+. Kxa3. (If Ka5. 5. Bc7#!!!) 5. Bxd6#!!!
The key squares are c3 and f2 for the WN which can be achieved from only e4. Also g5 needs to be guarded to trap the BK. Hence the right square is d6 from where the WN can move to e4!!!
3….Kb5 and b5 needs to be guarded, not g5
Beautiful indeed.
There is no time to unblock b8-h2 diagonal to stop B pawn so we have to build some net
to threaten the B King.
1. a6 seems natural to block a6 square 1… bxa6 (1… h2 2. a7 h1Q 3. a8Q+ Kc5 4. Nd6+ and
White will force BK to get to the a8-h1 diagonal and lose Q: eg 4… Kb6 5. Qa7+ Kc6 6. Qxb7+ and Qxh1)
However, I’m not so impressed with 2. Nd6!!! What after 2… f5!? It takes the key e4 square from white N
and discoordinates white pieces. I cannot find a way to trap BK now. Can you?
Therefore, the simple 2. Ne3 seems to be the only solution because it forces
2… h2 (otherwise Nf1) 3. Nd1 h1Q (otherwise Nf2) 4. Nc3+ Kxa3 5. Bc5#
Anyway, a beauty!
1… h2 2. a7 h1Q 3. a8Q+ Kc5 4. Nd6+ Kb6 5. Qa7+ Kc6 6. Qxb7+ Kc5 7. Qb5+ Kd4 8. Qe5#
1.a6 ba otherwise Wjhite queens first and mates
2.Ne3 h2 otherwise the N stops the h pawns with Nf1 e.g. Kb5 Nf1 Kc4 Ba7 Kd5 Kd3
3.Nd1 h1=Q White was threatening to stop the h pawns with Nf2
4.Nc3+ Kxa3
5.Bd6#
4…. Ka5 5.Bc7# sorry, I planned that but then for some reason I later had the Black a pawn on a5 in my head.
Sorry I said that White queens first if Black refuses to play 1…ba but of course White queens second but with check starting an interesting mating attack.
Stopping the pawns with Nf1 looks bad because the N isn’t guarding light squares so Nd1f2 allowing the pawn to run on a square might be better.
This is actually much harder than the lines given to date.
1. a6 ba is of course the only way for white to win, and black’s reply is forced.
2. Ne3 is correct, haded for f2 via d1.
2…h2 – black might as well do this as anything.
3. Nd1 K:a3 this is necessary for black to avoid the mate threats.
4. Nf2 and the fearsome h-pawns are stopped in their tracks.
Now the next phase of the ending is for white to eliminate the h-pawns while black goes after the remaining white pawn.
4… f5 (trying to keep the white B away from the h-pawns.
5. Bd6+ Ka4 6. Be7 h3 7. Bh4 Kb4 8. Bg3 Kc4 9. B:h2 Kd4 10. Kd2! white takes time out to keep the black K away from the f pawn.
10,,,a5 11. N:h3 Ke4 12. Ng5+ and the white K and N protect the f-pawn, while the B has time to get to the a1-h8 diagonal to stop the a-pawn.
Incredible endgame study! Thanks so much for sharing this, Susan. Who composed it?
It says Bernhardt 1961 and I think SP may have had problems with fans being too affectionate in the past so she has a bit of a wall up and isn’t going to make exceptions anyway but that doesn’t mean she doesn’t like you.
I remember this puzzle well because I botched it terribly, somehow convincing myself that 1.Ng3 could hold for white. Yikes! My files indicate it was February 18th 2010, but I couldn’t find it in the archives here, so maybe earlier that week. My notes also show CraigB was also involved in that discussion at the time- I copied his comment at that time to my files, it seems, though his comment today provides an even more thorough explication.
The issue still seems to revolve around white’s second move- I think maybe both 2.Ne3 and 2.Nd6 can win. Let’s first discuss Philosopher55’s suggestion about 2. …f5 taking e4 away in that line:
1. a6!! ba6 (forced)
2. Nd6 f5 (taking the e4 square away!)
3. Ne4!
Black can’t take the knight- doing so gives white the option of 3.f5 both stopping the h-pawns by opening the line to the h2 square for the bishop, and the white f-pawn runs free for a queen. Which solution is more elegant- 2.Nd6 or 2.Ne3? It is a matter of taste unless someone can show a forced mate shorter in one line than the other. I tend to prefer 2.Ne3 myself, though.
Wow, 3. Ne4! did not even come to my mind. :-(((
If so, the study seems to have 2 solutions (via 2. Nd6 and 2. Ne3). This should not happen
in a publishable chess composition. The composer would do everything to avoid such a grave
weakness. I suspect we miss something decisive.
If there is a difference in the two lines decisiveness, it is well beyond the horizon of my abilities. In neither line can I quickly mate the black king with the knight and bishop as long as black is careful to move his king at the right times. In both lines, white wins with brute force eventually. I only prefer the Ne3 line because it opens certain pretty variations shown above should black play 2. …h2. Other than that, I can’t identify which is actually stronger.