Stockfish on iPad versus Houdini on Windows

In my earlier blog "Can you believe your eyes ?", I showed that black is busted after the natural continuation 1.d4 d5 2.e4 dxe4 3.Nc3 Nf6 4.f3 exf3 5.Nxf3 Bg4 6.h3 Bh5 7.g4 Bg6 8.Ne5 e6 9.Qf3 c6 10.g5 Nd5 11.Bd3.

I showed the various line, and came up with a great white 18th move, ie 11...Nd7 12.Nxg6 hxg6 13.Nxd5 cxd5 14.00 Qe7 15.Bd2 f5 16.gxf6 gxf6 17.Bxg6+ Kd8 18.Kh1 !! Now reader "Anonymous" challenged my assessment and indicated "...After 18.Kh1 black has 18...Bh6 with only a slight edge for white."

Humm, That's strange, let's look at the line after 18...Bh6 (diagram).

My Stockfish app on my iPad immediately indicates 19.Ba5+ as clearly winning for white, whilst Houdini on my Windows PC only shows a very small edge for white ??  What's happening ? Could it be that Stockfish on my small iPad is better than Houdini on my serious Windows operating system ??

After 19.Ba5+ b6 20.Rae1, Houdini on Windows suggests 20...Rc8. Stockfisk on iPad answers 21.Bc3 (diagram). The disagreement continues as Houdini on Windows suggests a small advantage for white ( +0.23 ) whilst Stockfish on iPad sees a large white lead ( +1.5 ). Who is right ?

In this position, Houdini on Windows plays 21...Bg5, which is answered by Stockfish on iPad by 22.Bf5. Houdini on windows now defends with 22...Rc6 (diagram).

The disagreement continues : iPad says : 23.Re2 ( +0.8 ) Windows shouts : 23...Qf7. iPad plays 24.Bd3. Windows 24...f5 (diagram)

Houdini on Windows sticks to his earlier judgement and still sees a tiny white advantage ( +0.15 ), whilst the assessment by Stockfish on iPad has changed dramatically. From an earlier +1.5, it has now dropped to 0.5.

Who is right ? I guess Houdini on Windows is right, but actually... I dont care. I am happy that both engines, playing on such a master level, have problems assessing this position. I just shows the Blackmar Diemer is a serious chess opening leading to hugely complicated positions.

An alternative line

Last week, I showed that the line 1.d4 d5 2.e4 dxe4 3.Nc3 Nf6 4.f3 exf3 5.Nxf3 e6 6.Bg5 c5 7.Bxf6 Qxf6 8.Bb5+ offers no advantage to white as black can get an equal game after both 8...Bd7 and 8...Nd7.

So let's look for some alternatives. 8.d5 shows up immediately (diagram). Possible black moves include : 8...a6, 8...exd6, 8...Nd7, 8...Be7, 8...Bd6 and 8...Bd7

a/ 8...Bd7 9.Bc4 (+=)

b/ 8...Be7 9.Bb5+ (diagram)

...b1/ 9...Bd7 10.00
......b1a/ 10...a6 11.dxe6 fxe6 12.Ne4 Qh6 13.Bxd7+ Nxd7 14.Nfg5 (=)
......b1b/ 10...00 11.Qe2 (+=)
......b1c/ 10...Bxb5 11.Nxb5 (+)
......b1d/ 10...e5 11.Qe1 (+)
......b1d/ 10...Qh6 11.Qd3 (+=)
...b2/ 9...Nd7 10.d6 Bd8 11.00 (=)

c/ 8...Bd6 9.Bb5+ (+)

d/ 8...Nd7 9.Bc4 (=)

e/ 8...exd6 9.Bb5+ (+=)

f/ 8...a6 9.Qd2 (diagram)

...f1/ 9...Nd7 10.dxe6
......f1a/ 10...fxe6 11.000 (=)
......f1b/ 10...Qxe6+ 11.Kf2 (+=)
...f2/ 9...Bd7 10.000 (+=)
...f3/ 9...exd5 10.Nxd5 (+=)
...f4/ 9...e5 10.Bd3 (=)
...f5/ 9...Be7 10.Bd3 (=)
...f6/ 9...Bd6 10.dxe6 (=)

So we see that white doesn't have any improvement in 7.Bxf6 Qxf6 8.d5 compared to 8.Bb5+. What now ? Accept that the Blackmar Diemer does not win by force ? But then again, it took a worldchampion to figure out the great countermove 6...c5. 

More complicated Euwe defenses

In  June last year ( yes, already more than one year ago ), I covered some lines on the refutation of the refutation, occuring after 1.d4 d5 2.e4 dxe4 3.Nc3 Nf6 4.f3 exf3 5.Nxf3 e6 6.Bg5 c5 7.Bxf6 Qxf6 8.Bb5+. In my earlier blog , I indicated that both players have chances after 8...Bd7 9.Ne5.

Today, I will cover another, related line, when black blocks the check with the knight, eg 8...Nd7 and 8...Nc6.

After 8...Nc6, white simply develops with 9.00 (diagram) and the black queen finds itself in the firing line (+-)

Now that we have dismissed 8...Nc6, let's look a bit closer on 8...Nd7. I had considered 9.00 before, but it seems to me this brings nothing as 9...cxd4 10.Ne4 and black has the fantastic 10...Qf5 (diagram)

It seems that white has no advantage at all in this position, eg 11.Bxd7+ Bxd5 12.Nfg5 Qd5 13.Nxf7 Qxe4 14.Nxh8 Bc6 with complete equality. Now it is unlike that your opponent will find this great queenmove, but is there nothing better than 9.00 ??

Let's look at 9.Ne5 if this move gives an advantage.

It seems to me that 9...Bd6 gives black a completely equal game, eg 10.Nxd7 Bxd7 11.Bxd7+ Kxd7 12.Rf1 Qh4+ 13.Rf2

So I have to conclude that 6...c5 7.Bxf6 Qxf7 8.Bb5+ does not give any advantage to white as both 8...Bd7 and 8...Nd7 offer vlack a completely equal game. Maybe 8.d5 is better, but that's for another day

White to play and win !

Weaver Adams, an American chess player authored the book "White to Play and Win". In this, and other books, he tried to prove that white has a decisive opening advantage after 1.e4.

Now we all now that white has a decisive advantage, but not after 1.e4 but after 1.d4 d5 2.e4 dxe4 3.Nc3 Nf6 4.f3 exf3 5.Nxf3. I am not 100% sure when black does not capture on f3, but I am pretty sure white wins by force when black dares to capture on f3.

The proof ? Not necessary - just believe it and you will see ( I wont go as far as saying that playing the Blackmar Diemer is like religion, but it sure comes close for some... ).

Details on Weaver Adams can be found on Wikipedia. Unfortunately he did not meet Emil Diemer - imagine what weird theories they could have come up with if only they co-authored a book ?!

A subtle difference

In my earlier blog "Can you believe your eyes", I discovered that the line  1.d4 d5 2.e4 dxe4 3.Nc3 Nf6 4.f3 exf3 5.Nxf3 Bg4 6.h3 Bh5 7.g4 Bg6 8.Ne5 e6 9.Qf3 c6 10.g5 Nd5 11.Bd3 (diagram) is winning for white.

But, nearly the same position can occur through the Gunderam defense also, eg  1.d4 d5 2.e4 dxe4 3.Nc3 Nf6 4.f3 exf3 5.Nxf3 Bf5 6.Ne5 e6 7.g4 Bg6 8.Qf3 c6 9.g5 Nd5 10.Bd3 (diagram). Now this position is different from its Teichmann cousin by just the placement of the h-pawn.

Question. Is black also lost in this Gunderam line ( you would assume so, since the position is the same, no ) ? Answer. No ! Black has an equal game after 10...Qc7 11.Bxg6 hxg6. If white continues 12.00 as in the Teichmann variation, black has 12...Bd6 (diagram).

In the Teichmann line ( with the pawn on h3 ), white now has the basic combination 14.Qxf7+ Qxf7 15.Nxf7 and the bishop falls as both bishop and rook are attacked. But in the Gunderam, this is not possible as the bishop has an intermediate check on h2, saving the bishop from its aggressor, eg 13.Qxf7+ Qxf7 14.Nxf7 Bxh2+. Black then even ends up with the better game !


So 12.00 is not recommend and white should choose the alternative 12.Rf1 ( with an equal game after 12...Bb4 )


So we see that, at least in this line, the Teichmann offers more chances for white than the related Gunderam lines. Expert knowledge is obviously required to deal with this !