Odd Openings

       There exist a lot of openings that you never see in master practice, but not all of them are unsound; in some cases they just aren’t popular while in other cases they may be, if not positionally suspect, just not dynamic enough to generate any interest.  I remember GM Robert Byrne writing an article where he discussed what he called, if memory serves, twilight openings.  He gave the Budapest Gambit as one example stating that it’s playable even if rarely seen.  In a recent correspondence game I met a 2400+ opponent who played an unbooked opening as white against me.  Engines rely on opening books because so far with their books turned off, none of them have succeeded in discovering a “new” opening that GM’s consider worthwhile.
       In the game in question, I am assuming my opponent wanted to get me out of the book and at the same time, get into a position where it’s likely his positional judgment will outweigh mine and/or an engine. The game opened:

1.e4 e5 2.Qf3 Nc6

A game Dimitrov,P (2404)-Mazi,L (2374) Zadar 2008 continued 2...Nf6 3.Bc4 d6 4.h3 c6 5.d3 Be6 6.Bb3 Nbd7 7.c3 g6 8.Ne2 Bg7 9.0–0 0–0 and what can you say?  The position is fully equal.
3.c3
Also good is 3.Bc4 Nf6 4.Ne2 Bc5 and now black can play 4...d6 5.Nbc3 Bg4 6.Qg3 Qd7 7.d3 h6 8.0–o as in Varavin,V (2399)-Tolstikh,N (2425) Alushta 2004 or, for that matter, any number of engine recommendations all of which look OK for black.
3...Bc5 4.Ne2 Nf6 5.d3 d5
And now Black seems to have full equality, if not slightly the better game.

       That got me to thinking maybe I ought to try something different in one of my own games.  So in the following game I decided to forego my usual standard solid, mainline stuff and play something crazy, but not totally unsound.

1.d4 d5 2.Bg5 Qd6
Entering uncharted territory.  This move looks like an old time engine move from days gone by when they didn’t have much of a book.  Obviously it violates opening principles and isn’t something you’d want to play against a GM, but the move is not totally new.
3.e3 Qb6
This really takes us out of the books! Black played it safe in the game Urbina,C-Elflow,K Mallorca 2004 continued 4.Bd3 g6 5.Nf3 Bg7 6.0–0 0–0 7.Nc3 Nc6 8.Qe2 Bg4 9.Qe1 Bxf3 10.gxf3 e5 11.Nb5 Qe7 12.e4 dxe4 13.Bxe4 Nxd4 14.Nxd4 exd4 15.Rd1 Rfd8 16.Bxb7 Qxe1 17.Rfxe1 Rab8 18.Bc6 Rxb2 19.Re7 Rxc2 20.Rxc7 Rd6 21.Rc8+ Bf8 22.Bh6 Nd7 23.Bb7 Rxc8 24.Bxc8 Bxh6 0–1 Urbina,C-Elflow,K Mallorca 2004.

While Ivanova,D-Vladimirova,D, Ruse 20073...Bf5 4.Bd3 Qg6 5.Nf3 h6 6.Bxf5 Qxf5 7.Bf4 and it’s obvious White has a comfortable game. 
4.Nf3 …
There is lots of room for analysis at this point. The engines prefer 4.Nc3 when 4...Qxb2? 5.Nxd5 loses for black.
4...Qxb2?!
There wasn’t much point to playing 3…Qb3 if this isn’t the followup! This is safe now.  By safe I mean there is no immediate disaster as a result of the move, but after a likely continuation of:
5.Nbd2 Nf6 6.Rb1 Qc3 7.Rb3 Qa5 8.c4 c6

It should be pretty clear that black’s position is not one you would find many player’s willing to accept over the board, but in correspondence play it might be defensible.

In seems that in these two examples there isn’t any clear refutation and these lines might be worth looking into with the idea of generating some engine analysis and creating your own variations.  You probably won’t want to play them against Anand the next time you meet him though.