I analyzed the following position with both Stockfish 10 and Komodo 13.2. The 2 monster engines evaluated it at a dead draw at 0.00. Is this really the case or does either side have some sort of an advantage? Notice that in the position black is up 2 pawns although they aren't pawns to feel happy about. White has the bishop pair and a rook on an opening file. Compared to black's rook which is undeveloped. I would also be interested to know what is a neural network's(Leela) evaluation of the position is any of you have it downloaded.

My take: White is down 2 pawns, but he has the bishop pair, and black's pawn structure on the queenside is deteriorated. White also has a very active rook which can maneuver itself between the 2 open files to exert pressure. However, the main idea that I noted was the there are 2 open files so black can play King f7 and bring the rook over one of them. Perhaps the biggest compensation that white has is the massive space advantage due to the strong center. My take is that it should be equal, but can someone explain it better because the engine might not be right?

White to move

 [FEN "4k2r/p1p1n2p/B1p1bpp1/8/4PP2/B7/6PP/3R2K1 w k - 0 1"]

1 Answer 1


My initial impression was that in a practical game, white would win.

If it were not for the plan that the computer found, my impression would be that Rb1-b7, and winning back one pawn, with the remaining weak pawns, white would then likely win back the second, and just be clearly better. White is also ahead in development, and has an incredibly strong bishop pair, as you mentioned. Lastly, if the black king went to f7, as soon as it did, if white could take on c7, the Rh8 will be forced to go to e8, and it will be hard to break that pin. This would be very bad. I had a position once as white that could have ended up in this dream scenario, but my opponent played differently, and I ultimately won tactically instead. That game definitely influenced my initial, and shallow, assessment of the position. My initial, cursory, assessment was wrong.

Now, that I turn on the computer, it took the wind out of my idea immediately, and it is a likely draw for one reason: It is a positional standoff, and immediately, which white should force. With the Rb1 plan, which Stockfish 11 thought was best, black threatens to play Kd7 and Nc8, which defends everything. As soon as black plays Kd7, white needs to play Rd1 forcing the king back, and then can return to b1 with the same threat forcing a repetition. If black does not find that plan, white likely wins, but frankly, it would not be hard for a strong player to find. My initial thoughts were just from a quick glance.

White can try to move around, and induce a black mistake; but there are also many more pitfalls for white than black, and white could also make a mistake. You can never lose sight of the fact that you are two pawns down, and must be careful that there may be a defense that will leave you begging for a draw…if you are lucky.

The evaluation of the position is such:

Black has two pawns, and that is really the major redeeming quality of the position from black’s perspective. It helps that the weaknesses are close to home, which usually makes them more defendable, and the king is close enough to help with that. The white bishops are a bit unfortunately placed for a Bc5xa7 idea due to the move Ra8 winning, so it will take at least an extra move for that plan to work. Black has the possibility of Kd7 and Nc8 holding it together if Rb1 by white, and even with the potential loss of a7, black is up a pawn with a compact structure. White must be careful not to overplay his hand, or the extra pawns could become decisive. The Nc8 is the hero of black’s defense.

White has the two bishops, and a very weakened black queenside to go after. The Rd1 is more active for now, and the Rh8 is out-of-play for now, but might get out in time. The Kg1 is going to be the more active king. It is very important that the Pe4 takes the squares f5, and d5. especially, away from the black pieces. Black is, clearly, on the defensive.

With such a material imbalance, it comes down to very concrete calculation, which computers are particularly good at, and I believe the evaluation; however, there is room for both sides to play on, but to do so leaves the game on a razor’s edge. Black could slip up, and white’s activity could be decisive, but if white is not accurate, the material could have the last word, and from the lines below, this is the more-likely scenario.

The obvious plans for white are Rb1-b7, Bc5 and trying to capture on a7, and bringing the K to the center or queenside via d4 can c5 (it turns out that this is a fantasy though). Some of these might combine with a plan of Bb7-c6. Black’s plans are mostly reactionary to whatever white plays, but some are based on f5, securing the d5 square, or Nc8-d6 or b6.

After seeing the variations below, white easily has enough activity to compensate for the two-pawns deficit, but as the variations show, mostly due to the incredibly well-placed Ne7-c8, no more. To play for more is risking having to play a long defense for white.

In conclusion, I found many positions that white was either worse, or losing outright, but none that white was better. After analyzing this in depth, it is my opinion that with best play, white has absolutely nothing, and has a significantly higher risk of losing than black does, which I found counterintuitive and surprising.

After looking at the lines, as long as black sees the Kd7 and Nc8 plan in response to Rb1, which is fairly easy, he has no risk of losing. After looking at the lines, based on my years of experience, between two strong players, I did not see any moves that were particularly difficult for black to find. White should force the draw.

 [FEN "4k2r/p1p1n2p/B1p1bpp1/8/4PP2/B7/6PP/3R2K1 w k - 0 1"]

 1. Kf2 {It is the Kf2 plan that offers white the least losing chances.} (1. Bc5 {The Bc5 plan.} Kf7 $15 {For example.} 2. Rb1 Rd8 3. Kf2 (3. Rb7 $2 Rd1+ 4. Kf2 Rc1 5. Be3 (5. Bxe7 Kxe7 6. Rxc7+ Kd6 7. Rxh7 Ra1 8. Bf1 a5) 5... Rc2+ 6. Kf3 Bc8 7. Rxa7 Bxa6 8. Rxa6 Rc3 9. Ke2 f5 10. Kd2 (10. exf5 Nxf5 $17) 10... Rc4 11. exf5 Nxf5 12. Ra7 Ke6 13. Rxc7 h5 {And although equal, black is to be preferred.}) 3... f5 $1 4. Rb7 Rd7 5. e5 Nd5 6. g3 Nc3 7. Ke3 Nb5 8. Bxb5 cxb5 9. Rxa7 (9. Rxb5 $2 Rd5 $15) 9... c6= {But why would white want to play for this?}) (1. Bb7 $2 {The Bb7 plan immediately is just too slow.} Bc8 $1 (1... Kf7 {Is also good.} 2. Rc1 Bd7 3. Ba6 Re8 4. Bc5 f5 5. exf5 Nxf5 6. Bxa7 Ra8 7. Ra1 $1 Nd6 8. Bf2 Be6 $15) 2. Bxc8 Nxc8 3. Bb4 {Only move, and try to figure out why OTB.} Rg8 {Preparing to activate the rook, and black is better. There is no need to continue the analysis as it is clear that black is much better placed than at the beginning when white still had the two bishops, so white should have just forced the repetition.}) (1. Rb1 Kd7 2. Rd1+ {And white can force the repetition.} (2. Rb7 $6 {First, the Rb7 plan, which is trying to push too hard, and black is already at least equal.} Nc8 (2... Ra8 {Is also possible with the idea of Nc8-d6 or b6.} 3. Bc5 Nc8 4. Rb2 {And white is already forced to retreat or see the Rb7 get shut out of play, and maybe later, trapped.} Nb6 5. Rd2+ Ke8 6. Bb7 Rb8 7. Bxc6+ Bd7 8. Bxb6 Bxc6 9. Bxc7 Rb1+ 10. Kf2 Bxe4 $15) 3. Rb8 {Already having to bail out.} (3. Bc5 $4 Nd6 4. Bxd6 {Almost forced since e4 hangs.} (4. Rb4 Re8 {And black is winning now, and free. E4 is very weak.} 5. e5 fxe5 6. fxe5 Nf7 $19 {And d5 is now available for the Be6.}) 4... Kxd6 5. Rxa7 Rb8 6. Be2 Rb2 $17 {With a huge advantage that is almost winning.}) 3...Rd8 {Forcing white to bail out into the worse side of the draw, for example...} 4. Bxc8+ (4. Kf2 $2 Nb6 $17) 4... Rxc8 5. Rxc8 (5. Rb1 $2 Re8 {Again, aiming at e4.}) 5... Kxc8 6. Be7 f5 7. e5 Kd7 8. Bc5 a5 {Still equal, but why go for this when there is much easier?}) 2... Ke8 3. Rb1=) 1... Kf7 (1... Nc8 2. Rc1 Kd7 3. Rd1+ Ke8 (3... Nd6 $4 4. e5 $16) 4. Bb7 Bd7 5. Bc5 $2 Nd6 $1 6. Bxd6 cxd6 7. Rxd6 Ke7 8. Rd1 a5 9. Ra1 Rb8 10. Ba6 c5 $1 11. Rxa5 $4 Ra8 $19) 2. Rb1 (2. Bc5 $2 f5 3. exf5 Nxf5 $17 {And black has solved all his problems.}) 2...Rd8 3. Rb7 {There are too many lines to give them all, but black has no problems.} Rd7 4. Rxa7 f5 5. e5 Nd5 6. g3 Nc3 7. Bb4 Ne4+ {Or Nb5 or Nd5.} 8. Ke3 c5 9. Ba5 c6 10. Rxd7+ Bxd7=
  • 2
    I like your answer, but I am going to not accept it at the moment because I am thinking to bounty it so that more people are willing to help. I hope it's ok.
    – SubhanKhan
    Mar 5, 2020 at 5:11
  • 3
    It's for the purpose of expanding the community.
    – SubhanKhan
    Mar 5, 2020 at 5:11
  • 1
    @PhishMaster I just have one small thing to request for your answer which is that I could only use the 2 engines that I had access to, so if you have other programs like Houdini or Leela could you put their evaluation, as well as I, know that Houdini is a very strong defensive engine while Leela is very strong with the initiative. If you can then just put them in the comments.
    – SubhanKhan
    Mar 5, 2020 at 22:01
  • 1
    No problem! Your answer was very helpful anyways.
    – SubhanKhan
    Mar 6, 2020 at 3:03
  • 1
    @TonyEnnis I think that both are hanging by a thread, and to push it for either side, risks losing. How does black return the c6-pawn? Thanks. Mar 7, 2020 at 15:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.