Best Engine for B + N vs K Endgame

I just adore practicing the bishop and knight vs king endgame.  (I don't know why, but I can practice it for hours.)

I have about 25 different engines that I play against in this endgame, but my favorite by far is Dragon (which I use under the Fritz 12 GUI).

Most engines that I know of (even the strongest ones) usually run for the edge of the board and the opposite color corner.  From there I can readily win the endgame.  But Dragon prefers to run around the central squares.  You can definitely not easily force it to an edge (at my skill level, anyway, which is pretty poor).

I was just wondering if anyone knows of other free engines that put up a particularly strong fight in the bishop and knight vs king endgame.  Thanks for any help.

Here's a link for downloading Dragon, if you are interested.  It is a UCI engine.:

http://wbec-ridderkerk.nl/html/download.htm

Amy

You can use any chess engine, just make sure to have Nalimov tablebases installed. Your engine will play perfect defence then.

Perfect defense isn't necessarily the toughest defense. I think the OP wants an engine that will put up the biggest fight, not the longest fight.

You got a point there. But do you have a better recommendation?

Amy,

I just adore practicing the bishop and knight vs king endgame.  (I don't know why, but I can practice it for hours.)

You writing that tells me that you are hardwired for chess.  You can become a very strong chess player if you really want to.

Sred's post is right on the money.  What you will learn from the Nalimov tablebase is that in order to drive the enemy King from the center in the fewest number of moves is your King and the Bishop.  That way you can tempo him when you have the opposition, forcing the enemy king to back instead of go sideways on you and remain on the rank or file that he is on, and not allowing you to drive him back to the edge of the board. Then Dragon will slowly have to let you drive it back to corner that has a square that is opposite the color of your Bishop.  I am sure you know that the reason the enemy runs to that corner is because if you have a dark square color bishop you cannot mate the enemy king in the two light square color corners and vice versa.

This is one of those endgames that bears out how mathematical chess is.  It has to do with concentric triangles.  If you are interested I will post a diagram that illustrates it so you can click on the notation and the pieces will move.

It is very nice to know that there are dedicated players like you out there.

Yes, I can usually beat the stronger engines using tablebases--not because I'm a wiz, but because I have practiced them from many different starting positions. (I don't know if the previous statement is a totally logical conclusion, so if it is not, please forgive me.)

Dragon seems to play more randomly and often throws me for a loop.  I can still beat it maybe 80% of the time.  But the other times I'll lose by the 50-move rule.

Usually I can win by using Deletang's Triangles, but there are times when I mess up and it escapes.

Amy

Interesting. I would have guessed that in this particular endgame the longest defence is usually also the most difficult to beat. There was a time when I suffered from a similar addiction to this endgame, but there were no strong engines around, so I simply played the opposite side, too, making the king moves that gave me the worst headache

@sred

If a chess engine like namilov's is playing it sees no difference between optimal and sub-optimal moves.  It is programmed with the correct continuation in either case.  With humans it very different.  Against humans it is better to play the sub-optimal move, because the human cannot memorize every possible continuation and many times makes the wrong move in response.  And, the 50-move rule is in effect.  That is why I described for qixel the King and bishop manouver which is part of the endgame technique for this endgame.  Computers have perfect memory, human GMs and the rest of us use endgame technique. This endgame is possible to win by force from any position within 25 moves, 30 moves with a couple of inaccuracies.

Thanks for the insight.  Yes, that must be what is happening.  I know a lot of continuations by rote.  But Dragon often forces me to think out a move on the fly.  That also can force me to spend a lot of time on a given particular move.  I doubt if I could beat Dragon in anything under 10 minutes.

It may have a weakness when I use Deletang's Triangles against it. It often allows me a strong king infiltration when the lower part of the large triangle is blockaded by bishop and knight.  It does not seem to be playing optimally in these situations.

Amy

33 moves.

It is possible that my memory is faulty, but it would be interesting to see Nalimov's

When you can win this endgame in your sleep no matter what the position is, that is when you will know it well enough.  Once you know the endgame technique it becomes rote.  That's when they will say about you what they use to say about Capablanca, "Look at her she plays chess like a machine, unbelievable."

Bwahaha.  Thanks, transpo, I like that !!! :)

@qixel wrote:

It may have a weakness when I use Deletang's Triangles against it. It often allows me a strong king infiltration when the lower part of the large triangle is blockaded by bishop and knight.  It does not seem to be playing optimally in these situations.

I am not sure but I think the Deletang's Triangles that you mentioned are maybe the same as the concentric triangles I mentioned earlier.  It has probably been programmed to play suboptimal moves in order to get the opponent to leave a hole(s) in the net(the triangle) in order to escape thru that hole(s).  The corraling of the king with the concentric triangles is another endgame technique in this endgame.  In fact the corraling technique is the same one that is used in all the basic checkmate endgames.

here is the aforementioned tablebase.  It is where I practice endgames. Easy to use after you get accustomed to it.

http://www.k4it.de/index.php?topic=egtb&lang=en