Two Knights' Defense 5.exd5 Na5 8.Bd3 =

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 d5 5.exd5 Na5 6.Bb5ch c6 7.dxc6 bxc6 8.Bd3

8.Bd3 is growing in popularity. Every variation I have looked at e.g. 8...Bc5 favors White, except for the critical variation where Houdini just doesn't understand long-term compensation. All too often proponents of the Two Knights' Defense will claim long-term compensation when Black is down a pawn with minimal pressure.

Here, in the starting position (see below) Houdini likes White for a very long time. However, when you let Houdini play both sides of the starting position at 1 sec/move, 5 sec./move and 10 sec./ Black consistently equalizes.

What is particularly alarming for players of the White pieces is that there are more ways for White to go wrong; in other words, at a practical level we tend to believe that there are more ways for White to lose than there are ways for Black to lose. Black's game, for the most part, plays itself. White, on the other hand, must be careful

This test below gets more to the theoretical evaluation. With best play is White equal or better? In other words should White play 8.Bd3?

To see if this was true at a slower time limit I decided to give the computer enough time to get out to a depth of 26. Here is a game Houdini 3-Houdini 3 at Depth=26:

8.Bd3 Ng4! 9.Ne4 f5 10.Be2 h5

11.Ng3Bc5 (This is the starting position---you will note that White can lose immediately with 12.O-O?? Qh4 -+)

12.Bxg4 hxg4 (What should be alarming for White is that Black has spent a great deal of time on a pawn storm but also has a lead in development! This is excellent long-term compensation)

13.d3 (What else?) O-O! It takes Houdini on the previous move a long time to find O-O. Many players looking at that great Rook on h8 would not want to castle, but you have to consider the fact that the Rook can always return to the h-file with Rf6/f4/Rh6)

14.O-O Nb7 (Houdini is beginning to see the long-term compensation and shows a steady drop in its evaluation)

15.Qe1 f4 16.Ne4 Bd4 17.g3 Be6 18.Nbc3 Bd5 19.Ne2 Bxe4

20.dxe4 Nc5 (Houdini is beginning to say that this is closer to equality than advantage White)

21.gxf4 Nxe4 22.Nxd4 Qxd4 23.Qe3 Qxe3 24.fxe3 Rad8 25.Kg2 Nd6 26.b3 exf4 27.exf4 Nf5 28.Re1 a6 29.Be3 Rf6 30.Kg1 Nxe3 31.Rxe3 Rd2 32.Rae1 Rxc2 =

In my opinion 11...Bc5 is equal, and, as indicated, there are more ways for White to go wrong so I believe that this is the critical variation in the 8.Bd3 line.

What if White plays 11.h3, looks a better move than Ng3.

Are you claiming an  advantage for White in the line 8. Bd3 Nd5 9. Nf3 Nf4 10. O-O Bd6 11. Re1 Nxd3? I would think that the bishop pair and white's doubled pawns would give Black compensation.

White seems to be better with the exchange sacrifice 8. Bd3 Ng4 9. Ne4 f5 10. Be2 h5 11. h3 fxe4 12. hxg4 Bc5 13. b4 Qd4 14. bxc5 O-O 15. O-O Qxa1, which I would mark as the mainline of Ng4 line.

I have to say looking over a few master games in a database is probably a better way to do opening research then having two computers play each other.

Some fresh stuff on that variation:

Oh, and of course white has played 11.h3 in ALL the games this position has encountered, but as usual Sloughter did not bother checking insignificant things like grandmaster games and databases... his Houdini knows better, doesn't he?

Well what the hell was the point of 13. c5?,it neither devolped a piece or dealt with a threat. Stelios also voluntarly surrended the Bishop pair with 15...Bg4 16. h3 Bxf3. 

How long will you keep that knight on a5, staring the rest of his army?

Black is actually OK in that game, his first mistake being 21...Kh8, and 22...Qc7? is a blunder - then he went quickly down (probably due to time pressure).

I think black is absolutely fine after 21...Qg5, which keeps the g3 knight restricted.

I was planing on recylce it via the b7 square at some point it would probably move to c5 and from there I don't really know. I probably should go back to thinking things over for a day before writing them down.

13..Bxb4 instead of Qd4 getting into the exchange sacrifice is an alternative? No actually it seems Qd4 may be best. Black probably still can hold on there with with best play?

The endgame (two pawns for the exchange) is almost certainly a draw, and it was well analysed since the game Stellwagen- L'Ami, Maastricht 2008.

Seems black was playing with a clear idea of controlling d4 while d2 and d3 seem to cut white's position in half.  His d6 bishop may look passive but if white has no pawn breaks and black has more space I'd think the worst black can do is a draw.  Clean looking game to me until black blunders.

Interesting to me if black tried 16...Bh5 if white has to play g4.  If so then Bg6 and f5 open him up in a bad way meanwhile Nd4 and f5 are coming anyway.

I'd looked at h3 in another post; Black has all the fun and even though the final position is equal, Black has better squares for his pieces. The position is equal

Jeez, you didn't care to look at existing games (again). 13.b4 is the critical line (instead of your 13.Nc3).

13 b4 was mentioned earlier in this same thread even?

13.b4 changes nothing i.e. 13...Qd4! 14.bxc5 O-O! and the mate threat gives Black time to castle. After 15.O-O Qxa1 Houdini has this as -0.07 at a depth of 25. Do you really think that White has any realistic chance of a plus?

White has a record of +3 =5 -0 from that position.

You cannot really do any serious analysis with that attitude, you can trust me on that.

19...Bf5 begs for 20.g4 Be6 21.Ba3, when you have granted white a useful extra move (g2-g4) over the more natural (and better) 19...Be6.

Either your Houdini is broken, or you have no clue how to use your rig.

Which pawn is White going to promote?

Errr, any. He has plenty of them, doesn't he?

19...Be6 must be a draw, but Black has to be careful.

As I indicated 8...Ng4 in the initial post is equal. More importantly Ng3 is superior to h3 when Black's position plays itself. All Black has to do is find in the "main" line is e3! & it's dead equal. At least with Ng3 White can complicate.