My plan was to play 14. Nxb4 all along (the position by 13. Na2 has already happened before once, and Black played 13. ...a5, which seems clearly inferior to 13. ...c5, so I had to give 13. ...c5 a look).

However, yesterday evening I got sidetracked by the possibility of playing 14. Nd3 instead. Black has to play 14. ...exd4. Now I can play

14. Bxd4 or 15. Naxb4

A) 15. Bxd4 Bc5. Now, the intuitive move is to gain a pair of Bishop, but 16. Bxc5 is interesting too:


A1) 16. Bxc5 Nxc5 17. Qa3.



My plan here is to force Black to eventually play ...b6, leaving him with a weak c5 Pawn after various exchanges. But this appears to fail: 17. ...Rac8 18. Nc1 Rae8?! (protecting the Queen and unpinning the Knight) 18. Bb5 Nxd3. If 20. Bxe8, then 20. ...Qxa3 21. Bxf7+ Kxf7 22. bxa3 (22.Rxa3? Nxc1) ...Nxc1 23. Rfxc1.



Black will slowly free his Bishop by playing ...Bg8 and ...e5, and my queenside pawns are weak. Black probably plays 22. ...Rxc1+ 23. Rxc1 Ne8!



If I get too frisky, I might end up in trouble: 24. Rc8 Ke7 25. Ra8 a6 26. Rb8 Nd6 and my Rook has no moves until Black moves his King, so he has the time to free his Bishop)

So I think White should play 20. Qxe7 Rxe7 21. Bxd3 (21. Nxd3? a6)



with a slight advantage, due to Black's passive Bishop.

17. ...Nfd7 does not change much: 18. Nc1 Rfe8?! 19. Nb3, Black takes a Knight, I change Queens and have a slightly better endgame.

So this seemed like a good idea, but then I discovered all these variations could give Black the idea of a more precise move order:

17. ...Rac8 18. Nc1 a6! 19. Nb3 Nfd7. Black's next move is 20. ...Rfe8, freeing himself.



All in all, the pin is not as dangerous as it looks.

Let's look at 16. Nxc5 now.

A1) 16. Nxc5 Nxc5. Should I play 17. Qa3, Black will free himself from the pin in the exact same way indicated above (...Rac8, ...a6, ...Nfd7 and ... Rae8).


So let us try 17. Qb5.



Now 17. ...Nfd7 looks bad because of 18. b4 (if 18. ...a6, then 19. Bxc5 axb5 20. Bxe7 bxc4 21. Bxf8). If Black play 17. ...Rfc8, then I still reply 18. b4 The sacrifice 18. ...Ncxe4 19.fxe4 Bxe4 does not seem to work bc White can placidly play 20. Bb3, parrying most threats:



Hence, 18. ...Ncd7 19. Qxb7 Rab8 20. Qa6.



My pieces are a bit over the place, and I hope I am not missing a tactic, but I should eventually be able to get the Queen back by Be2 and Qd3).

So it sure looks like Black has to play 17. ...b6. Now the obvious moves is to grab on c5, but 18. a5 is interesting: 18. ...Nbd7 19. axb6 axb6 20. Nb4 (obv not 20. Bxb6? Rab8) with the idea of 21. Nc6.



If I play 18. Bxc5, I plan to answer 18. ...Qxc5+ by 19. Qxc5 bxc5 20. b4. Sure, I am freeing Black of his weak Pawn, but as long as his Pawn is there, my own Pawn on b2 is also weak. However, Black does not have to exchange Queens: 18. ... bxc5.



His Queenside pawns are weak, but he is threatening ...Rb8 and ...Rxb2, so I have to play or 18. b3 (18. Qb3 Rab8 19.Qc3 Rfd8 and the White Queen's position harms my Knight a great deal). Black will then be able to play Rfd8 and place a Rook on d4, conquering the d-file. Sure, his pawns are weak and his Bishop sucks, but I'd rather not let him have that counterplay. Here's a possible variation: 18. b3 Rab8 19. Qa5 Rfd8 20. Rfd1 Rd4




So 17. a5 looks ultimately better than the obvious capture in c5.

B) 15. Naxb4.

15. ...dxe3 16. Rfe1 (if I play 16. Nc2, then 16. ...e2, and I have to play 17. Rfe1 anyway) ...Nc5.



Now, neither 17. Nxc5 Qxc5 nor 17. Qa3 seem very favourable to me. I admit I was not very thorough while analyzing these positions; this is probably due to the fact that the best position I can hope to achieve here are not better than what I have in the 15. Axd4 variation.

B) 14. Nxb4. Ultimately, I decided to choose this rather than go for this promising position:



The only reason is that 14. Nxb4 seems to give me an extra pawn.
After 14. ...cxb4 15. Nxd7 Nxd7 I plan to play 16. a5, preventing Black from protecting the Pawn in b4 with his peer.



I am now threatening 17. Ra4, and I do not see a good defence against this threat, though it is possible that I am missing smth really, really, obvious. Plans involving ...e5 and ...Nc5 do not seem to work; for example, 16. ...e5 17.d5 Nc5 18. Qxb4, and Black does not have time to play ...f5 because the Knight is hanging.

Black might try to play 16. ...b6, trying to open the a-file, but I can simply play 17. a6 Rb8 (now I cannot play 18. Ra4 bc of 18. ...b5) 18. Bb5 (and definitely not 18. Bf4 e5).



Now 18. ...e5 appears, again, not to work: 19. Bxd7 exd4 (19. ...Qxd7 2. Qxb4) 20. Bxd4 Qxd7 21. Qxb4.



Sure, opposite Bishop have drawish tendencies, but his Bishop is very passive.

Looking forward to see if I am missing smth really obvious.
Also wondering if anyone has bothered to read this

I took another look at the positions and I noticed that after 15. ...axb4 16. Nxd7 Nxd7 17. a5, rtspurs can play 17. ...Nb8 with the idea of 18. ...Nc6. Not sure how I missed it, but here we are. Anyway, I guess I can reply with 18. d5, which should shut down the whole idea. So no harm, no foul?