OK, our idea was 16.Nxe4 dxe4 17.Qxa5 Nxa5 18.Nd4
(also after 18.Nd2 e3 19.Nc4 Bc6 20.Rh3 Rxd1 21.Kxd1 Nxc4 and Rd8-Rd4 and we thought that Black couldn't be worse. The engines' evaluation confirmed that)
18...Ba4! this works due to some tactical tricks that we saw very quickly and the computer couldn't! This was the continuation of the game also.
19.b3
(19.Bg2 Rac8 20.Kb1 Rc4 is obviously confortable for Black)
19...Rac8 20.Kb2
(our idea was ofcourse 20.bxa4 Rxd4! my opponent had a slight smile on his face when he calculated this OTB. He had invested a big amount of time and energy up to this point and saiid to me than he now felt that he had the advantage. But.....)
20...Nc6
(we were expecting that! I was still following the "pub" preperation!!!!)
21.Nxc6 Bxc6 22.Bg2 f6! 23.Rxd8 Rxd8 24.Re1 f5 25.gxf5 exf5 26.Bh3 Rf8
(all these are logical moves up to this point. We thought that Black should be a liitle worse here but we then realised a crucial detail. My opponent felt insulted because i was replying in lightining speed untill now)
27.Bg2?! (i didn't like this move, but White has a week 'h' pawn no matter what!)
27...Be8! (simple chess) 28.Rh1 Bf7 and i now started to play real chess, not relying in my memory while my opponent was allready tired and in a bad psucological situation knowing that i had prepared all the moves up to this point. It's no suprise that i gained a preety much big advantage after 5-6 moved and when i gave him the chance to draw he couldn't find it. So, i won in the end.
I think that the final position is a little better for black but defendable for white. But as a conclusion i think that the concept we found improves the Black chances considerably over existent theory. Any thoughts?