Interview med Andreas Skytte Hagen
Interview med IM Andreas Skytte Hagen af IM Silas Esben Lund
Email interview with Danish International Master Andreas Skytte Hagen prior to the publication of Opening Simulator – King’s Indian Defence, Quality Chess 2019 (Esben Lund & Andreas Skytte Hagen). The book is set to be published at the end of July 2019, the interview was conducted shortly before the book release.
Co-writer and interviewer: Silas Esben Lund.
At the end of the interview I have added a KID game that Andreas played after the draft of the book was handed in to Quality Chess, with his own annotations.
SEL: Since this is your first chess book, chess players outside of Denmark would probably like to know more about you. Please give us a short outline of your life and chess career so far.
ASH: When I was a junior, I was good enough to compete for the Danish junior Championship and in fact won it in the 8th grade. In the 6th grade, I came into 3rd place in the Nordic junior Championship.
At around 15, the interest in the game faded, but later it reappeared. At around 2006-2007 and especially in my late twenties I started to become very interested in chess again and has been at it since then.
In 2011 at the age of 29, I got a big breakthrough by jumping from around 2270 to 2454 within less than 1 year (without taking time off from the job or anything) scoring 1 GM norm. I also got married in 2011, and I now have 3 children (born 2014, 2016 and 2018).
In my daily job I work as an actuary (got my master’s degree in 2010), and currently I test the calculation part of IT systems to the life and pension industry.
SEL: Your first chess book, and our first chess book written together, is close to publication. The title and the idea was yours – how did you get that idea?
ASH: I will mention 3 things that influenced me in such a way that the idea was born.
In my search for how to become a better player, I fell upon Jonathan Rowson’s Chess for Zebras. Over the years you know I have been quoting this book almost every time we met. I felt there was something to this book that was important. Notes were taken, and I was re-reading it. The key concepts in this book that triggered the idea for Opening Simulator was the concept of “know-how versus know-that”, and “reading and nodding”. The concepts are mentioned in the book, so I will not repeat them here in detail. But the basic point is that getting a lot of information and facts like endless variations and theory (know-that) will not make you more skilled, because what you need is know-how: the ability to play the position. Reading and nodding is a concept originally coined by Nigel Davies and covers that people can read a text and think they understand it (read and nod), but when it comes to actually applying what has been learned, they come up short. To sum up: know-that and reading and nodding does not make you a better player, but know-how does.
The other book that influenced me a lot was Colvin’s Talent is overrated. In this book, Colvin investigates the concept of “deliberate practice”. Deliberate practice consists of 5 elements, and these 5 elements must be in the training if you want to improve. Again, I got obsessed with this book, re-reading it and taking notes, and ever since I have spent a major part of my spare time thinking about how something can be practiced correctly in chess.
The above-mentioned books are what I would call “life-changing experiences”, because the training methods do not only apply to chess.
A third thing that influenced me was a conversation with the author of the woodpecker method, Hans Tikkanen, around 2010. He had also read Colvin’s book, but had interpreted it in the way that the exact same exercises should be repeated. I liked this interpretation (and Tikkanen’s results were compelling), so I implemented it in my own training. (For comparison, I solved tactics for half an hour every day at that point, but did not repeat the exact same exercises until I talked to Tikkanen. The reason was that one of the deliberate practice points was “It can be repeated a lot”, and here I thought of just solving exercises.)
Solving exercises are in general useful, but why not optimize the time spent training on an opening that you get often in your games? That was the question I asked myself, and based on my findings above, that was how the concept of “Opening Simulator” was born.
All this in fact happened because of you, since you recommended the two above-mentioned books to me, and it was through you that I got in contact with the Swedish players through your affiliation with the Lund Chess Club (LASK). Around that time, a lot of very ambitious players were playing in Lund, not only the two woodpecker authors, Tikkanen and Smith, but also Semcesen who made it to the GM title and Grandelius who is clearly the Swedish star right now.
Also, I was inspired by the lack of books with the approach I thought was correct, and since I couldn’t find books living up to these standards, I thought the idea should be published.
The common denominator for practically all opening books is that the reader is not forced to think. That makes them work more like dictionaries (know-that) or makes a reading and nodding approach too tempting for the reader (game collections and so on), and if the reader thinks his skill will improve he will be disappointed. I will mention below some approaches I found interesting.
The approach by Bellin/Ponzetto is a sympathetic one (read and play method), but thinking that the reader can just read (and nod), and then play well afterwards does not work in an optimal way if you ask me. We call our method “read, solve and play” because in order to be able to play well you have to increase your skill, and that is done by forcing your mind to think over problems (solving) that can occur in your games. The DVD “Tactics in the King’s Indian Defence” has a good idea, but it is implemented with a lot of exercises where it is not clear to see that the position arose from the KID.
Smirin’s King’s Indian Warfare is on to something, at least it has introductory exercises to the chapters, and together with comments to games from a great KID player, it makes this work valuable.
The work by Minev deserves special mention, which will be done later in the interview.
SEL: I have done all the writing of the book whereas you have done most of the research, and we have discussed central ideas at length to get both the white and black perspective. Also, we have both shared the burden of selecting 400 exercises on 5 levels. How does it feel now that years of work has finally come to fruition in the form of a published book?
ASH: It feels great to have an idea and see it realized. After the project started, I have thought a lot about different ideas and exercises, and it’s a relief that the project is done: now all that energy can be used elsewhere.
I wanted to be part of this because I thought the world should know about this idea of how to study an opening. Even though I don’t think it takes anything special to come up with the idea, I did not understand why anyone had not thought of the same before. Since I thought it would be something new, it would be interesting to publish. I would not have done it at this point in my life due to a lack of time – but since you offered to do the writing, I thought: why not?
As this is maybe the only chess book I will have my name on, I really wanted the product to be good, so I decided to spend a small fortune on every relevant KID book that has ever been published. The time spent researching and the money spent on this is something I better not think about. Some of the books were out of print, and especially one book was difficult to get a hold of, but I eventually found it in a Dutch antique book store not related to chess. It was the book by Minev mentioned in the bibliography. It turned out he had the idea of solving exercises in the opening you want to play, and to be sure he gets credit I would like to mention that he is the first I found with about the same idea. Some of the differences compared to our book is that he does not link it to science as we do (deliberate practice), and he gives exercises in positions where the reader can’t see that the position arose from the KID. In our book, the reader should be able to see that the position could have arisen from the KID – and if not, the solution should explain how it arose from the KID.
Also, in our book the reader should not be in doubt whether or not he has found enough to claim the problem solved. One of the points of deliberate practice is that feedback is continuously available. In our book, the reader will know exactly if an exercise is solved correctly, so there is no room for self-deception. Many books have incomplete or ambiguous solutions, and thereby not optimal feedback, and I wanted to make sure we did that part better. I have never seen any book where the solutions are so clear as ours.
A last difference I want to mention is that we divide our material into 5 different levels for pedagogical reasons, whereas Minev does not.
SEL: Our cooperation started sometime in 2007 if I remember correctly, where I blogged out that I was looking for a training partner. Your answered that call, and we started out when you were rated around 2250 Elo. Around that time, I was finalizing my IM title and in the spring of 2008, I made my first GM norm. Please reflect on these first years of playing training games together as well as doing other chess training.
ASH: I would describe the period as a search for the optimal way of training.
We played a lot of training games, and as I was significantly lower rated than you, I learned a great deal from the way you looked at different positions. I realized that my play lacked basic dynamic skills, and that was where my journey into dynamic and high-risk openings began, because my philosophy is that if I have to learn it, I first have to do it (the first times I then look like an idiot, but eventually things will be better).
I think we had a training game with opposite-coloured bishops, and I used my bishop to protect a pawn from a passive position. You pointed out that the bishop was “lame”, and the feedback hurt so much that I realized that I had misunderstood a great deal about what good chess is.
You had no problems with material imbalances and giving up material for something not so concrete (for example just for “position”), and that all influenced me and changed my perception.
The culmination happened in 2011 where everything clicked: Rowson, deliberate practice, my daily training, training games with you. It catapulted me from around 2270 to 2454 at age 29. Even though that might not be a remarkable achievement, I can proudly say that it was done without taking time off from work (I have never taken time off in order to prioritize chess).
SEL: Later, when you became a stronger player, I started testing you in sharp endgames – and the material and thoughts behind it later turned into a book published by Quality Chess (Esben Lund: Sharp Endgames, 2017). Tell the readers of your view of this whole process where you asked me to implement deliberate practice to optimize the training.
ASH: I remember we had a conversation about implementing deliberate practice to endings. What I had intended was just applying deliberate practice to known endgame theory, but I realized your intent was to implement it to the step before that: the sharp endgames leading up to the known theory. We agreed upon you training me in your material, and I liked the idea of using deliberate practice in this training as it worked as a safety net for not wasting time. That meant that whenever I had made something wrong or suboptimal, you would present me with the same position after a while (when I had forgotten the solution). It was a kind of brainwash or positive torture, but I think it was great. And by the way, I was amazed with how much interesting material you had accumulated, and I am glad you decided to publish it.
SEL: Sharp Endgames again: I gave you a lot of exercises to solve, including all of the exercises that ended up in the book. Please reflect on what you learned from this. It is always difficult to suggest in terms of increased rating, but perhaps it changed your perception of endgames (basic and sharp) and your decision-making.
ASH: It changed my perception of the game. Previously when I had a better or equal position, I would be afraid of changing the nature of the position for fear I could not control the consequences. This is no longer such a big problem for me, but old habits die hard, so it is still something I have to work with.
It may seem strange that I have no problems entering an all-out attack in the KID, but at the same time have a problem with changing the nature of the position in an endgame. But there is a big difference: In the KID, I know that even though my attack is maybe dubious, my opponent will give me 1-2 or even 3 chances to get back and even win because defending an attack on the king in the KID is very difficult over the board. In a sharp endgame, decisions are more fatal: a miscalculation of a tempo and the opponent queens before me. If I have made a miscalculation in the KID, I throw the 2nd or 3rd wave at him.
Also, I had some problems with my understanding of the knight (for example: quickly seeing knight routes), but after the training my familiarity with this piece improved a lot. If I had not received this training, I would not have been able to beat Peter Heine Nielsen in a knight endgame in a deciding last round game of the Nordic Championship in 2013. Admittedly, Peter Heine Nielsen had to take a lot of risk in a worse position to be able to win the Nordic Championship, and if that had not been the case, I don’t think I would have won. But when he took huge risks I was ready to quickly establish the knight routes under time pressure to claim the win in that game, and instead Axel Smith took the Nordic Championship title.
SEL: Deliberate practice has been an interest for you after coming across the book by Geoff Colvin (Talent is Overrated. What really Separates World-Class Performers from Everybody Else). Deliberate practice is the common denominator for both Sharp Endgames and Opening Simulator – please elaborate on your interest in and experience with the concept and how it has worked for you.
ASH: When I got acquainted with the deliberate practice concept some things became clear to me. I now felt I understood why I did not seem to improve. I also realized that you do not get very far if you just say “I have no talent for X”. A lot of people like the talent explanation because it makes them not responsible for what is happening. As they have no “talent” they can worry about other things (like relaxing) – at least that is the story they tell themselves. If the theory behind deliberate practice is correct, you follow the 5 underlying points, do the suffering and improve. It means you are 100% responsible for your results, and I am very attracted to philosophies like that.
When reading Colvin’s book I understood that top performers brutally isolate the skill that is needed to perform, and then work intently at it via deliberate practice. When the specific skill is increased, it is on to the next. Our book is trying to isolate a skill that is needed to perform well: The skill of playing a position arising from the KID. Solving a book on general tactical exercises is of course a good way of training, but I felt like we could isolate the skill even more by making this book.
Of course my own jump in 2011 is not a proof the deliberate practice works, but I have no doubt that it does work, because I felt the effects: Over a long enough time period deliberate practice will prove its worth.
I also tried the deliberate practice training methods on a very ambitious 1700 adult player who over a time period of a couple of years raised his game to almost 2100 as far as I recall. We did 6-hour training sessions where I sat next to him when he solved easy tactical exercises, and when he thought he had a solution I asked him about if he had seen an answer to this and that move, constantly trying to stretch his horizon. It also meant he had to play out positions from Dvoretsky’s Endgame Manual against me in order to prove the authors evaluation of certain positions. And if he failed, I would return to the position at some point.
Working with children, I would recommend skipping the last two points of deliberate practice (“it is mentally hard”, and “it isn’t much fun”). It is my impression that children work in another way, but if you are a committed adult and really want to improve on something, I recommend all 5 points of deliberate practice.
SEL: Once we had the concept Opening Simulator ready, we had to agree upon a well-suited opening. As I recall, it didn’t take us long to end up with the King’s Indian Defence. Tell us about your fascination with the King’s Indian Defence and why you decided to add it to your repertoire. I know that you have done a lot of thinking and research on the opening, give us some insights into this too.
ASH: One thing was certain: It had to be a big opening complex where the configuration was easily identifiable with the opening. Openings like the Najdorf, Ruy Lopez, QGD, Nimzo Indian, King’s Indian and Grunfeld Indian came to mind.
Over the years, I had experimented a lot with different openings – because how could I know enough about an opening to conclude anything if I had not played it? The experimentation led me towards the Grunfeld Indian because I really liked the dynamics and search for the win. White should not commit many mistakes in that opening before the advantage to Black becomes serious.
Even though I was very satisfied with the Grunfeld, I thought about taking even more risk by experimenting with the King’s Indian. I quickly became fascinated. To me, the KID is the opening you want to play if you want to take maximum risk to play for the win, but in a theoretically approved way. In the KID, you get to study the games of some of history’s most exciting and uncompromising players like Fischer and Kasparov.
One of the main arguments I have heard for not playing the KID is that it is so risky with a lot of opening theory, and that it won’t be possible to play the opening without this vast opening knowledge. This argument does not correspond with my own results. Yes, there is a lot of theory and it is risky, but it is possible to play with good results without knowing a lot of theory. My games against Pelletier and Prizant come to mind where I had absolutely strategically losing positions from the opening but even managed to win both games despite a huge rating difference in both games. Even if you get an objectively inferior or losing position in the KID, White cannot slowly strangle you in many of the variations. Getting slowly strangled is something I can’t stand. The Mar del Plata variation is a good example of a variation where even though White breaks through on the queenside obtaining a winning position, he will have trouble withstanding the Black attack on the king. The practical aspect is in Black’s favour: If you have a losing position, you can always throw the kitchen sink attack: Let White respond to that over the board.
Another argument I have heard for not playing the KID is that White objectively has a better position. People refer to Fischer not employing it in the World championship match against Spassky (or against Petrosian) and Kasparov quitting it while he was an active player.
I have one argument against this that I think is very strong: If White was better, why is the KID so popular in correspondence chess? Correspondence chess is as close to the truth we can get, and if correspondence players think the KID is great, then that must be a better guess. That was the theoretical discussion. In the practical discussion, I can say I have great results, and I love playing the KID. That is enough for me.
I became very fascinated with the Mar del Plata pawn races. White wants to break through on the queenside and Black wants to checkmate White. A lot of times I saw big advantages to White given by the computer, but doesn’t the Black attack look dangerous? What if I play g4-g3 (confer Taimanov-Najdorf) and sacrifice this pawn even though the computer says it is bad for Black? White takes with hxg3 fxg3 Bxg3, but then h5-h4 Bh2 Nh5. Still, the computer says White is a lot better. But then playing some more natural moves I saw Black has full compensation and a huge attack, in practice very difficult to withstand for White over the board. This happened repeatedly when I studied the KID: The computer claims White is better but further study revealed this to be not true. Then the computer evaluation drops, and it recognizes the Black attack.
Besides the above arguments for choosing the KID, I think the KID is the opening that should be chosen for the book based on ours, the authors’, experience. The KID is the opening I have played the most against 1.d4, and I have also played the White side quite a lot, as have you. I thought the KID was the best opening to combine our knowledge and experience.
SEL: What is your work with chess these days? I know you have a daughter who is doing well with chess and excelled at a young age, tell us more about that.
ASH: Most of my work with chess these days is done in relation to my children. Frida (now 5) learned to play when she was 3, and Julian (just turned 3) can play chess well without the checkmate concept, i.e. the aim of the game is just to kill all the pieces. My other daughter, Vilje, just turned one, and she likes to play as well, but not with the correct rules as we know them: She tries to imitate how we move the pieces, and then looks at me for acknowledgement, and no matter what she does I always say it is a good move. If she moves rook from a1 to d4 I applaud, and she applauds herself, because at this point I realize that teaching her the correct moves is not useful (it would be in the panic zone, i.e. it is a task she can’t solve). 3 children hungry for knowledge (not just chess related), and I feel I should give them that. That leaves little time for my own chess work, but right now I see no other solution.
I have no pedagogical background, but based on my experience with my own children, I feel I know enough to give the following general advice when teaching chess to toddlers:
Children have a natural curiosity towards new things, and it is possible to build the interest for chess upon this curiosity. I view it as the adults’ responsibility to communicate the game in an exciting way. What is an exciting way depends on the child. If you know the child well you know what interests the child, so use that as an entrance to make the game exciting. Remember, every topic can be interesting or boring – it all depends on how you communicate it. This implies choosing tasks that are in the learning zone of the child at the specific moment, again referring to one of the points of deliberate practice (it is designed to improve performance). If you choose something that is too difficult, the child will lose interest.
I recommend removing distractions from the room, especially the really tempting brain dead distractions like the iPad, and turn off the television. The mind of the toddler is easily distracted. Always make the child get the feeling that it is competent, and if the child does not really get it, and just wants to play around with the pieces, just go with the flow and let the child decide.
It is difficult for me to give concrete advice about activities, because I think it all depends on the specific child, but I feel like the above general recommendations are useful. Remember, as the adult, you are the designer of the practice activity.
My experience is that the game can be taught to children at least from the age of 3 (if done correctly), and the pieces can be introduced whenever the child gets control of the basic senses. If my 3 children will continue with chess when they grow older is up to them. The key point I want them to take away from the game is that if they want to improve at something, it can be done (deliberate practice), and the confidence they get from that. I also want them to value the importance of critical thinking, and chess is very instructive in teaching a child that. A child will think up wonderful plans like Nf3-g5, Qd1-h5, Qh5xh7 checkmate, but the child has to learn that after Nf3-g5 the move h7-h6 ruins this plan. Thus, chess makes you critical (this is a positive word in my vocabulary), not just towards the world, but also towards your own fantasy plans.
SEL: Last question – anything else from our years of cooperation that comes to mind that you want to share?
ASH: One thing springs to mind. Besides the common denominator between Sharp Endgames and Opening Simulator KID, I see a common theme between The Secret Life Of Bad Bishops (Quality Chess, 2014 – SEL) and our book too: The double-edged bishop mentioned in your book exists in the KID as the bishop on g7. Sometimes I hear that this is Black’s “bad bishop”, and that Black should find a way to exchange it, especially in the Mar del Plata pawn structures. I think calling it a double-edged bishop is more to the point, because its potential can be realized if handled correctly. For example: It can be activated through h6 in the Mar del Plata and the Petrosian Variations, and it can play a crucial attacking role in the Mar del Plata on h6 after the Taimanov-Najdorf pawn sacrifice g4-g3. Another theme to realize its potential is to play Nh5-f4 and sacrifice a pawn after Bxf4 exf4 Qxf4. The bishop’s strength typically easily outweighs the loss of a pawn. Also, using it to hold the Black structure together in the Mar del Plata pawn races via f8, protecting d6, can be an important role.
I think you mention the KID bishop as a DEB in your book.
SEL: Yes, I do mention it briefly in the introduction to the book (page 14). Thank you for the interview, and good luck with all your chess activities.
Below is the KID game with Andreas’ own annotations.
Jakob Aabling Thomsen – Andreas Skytte Hagen
Danish Chess Team Championship (8) 2019 (Brønshøj-Skolerne)
Played on March 16 2019.
Jakob is an IM at about the same rating level as me, but my score against him in classical chess before this game was not good: One draw and three losses.
One of Jakob’s preferred openings with Black is also the King’s Indian, so I
thought about if it was really a smart idea to play an opening in which he is
an expert. But as I had solved all the exercises in Opening Simulator – King’s Indian Defence, I thought I would use every opportunity to see my skills in action in this opening.
1. c4 Nf6 2. Nc3 g6 3. d4 Bg7 4. e4 d6 5. Nf3 O-O 6.Be3
I had seen in the database before the game that Jakob plays this
variation. Previously, Jakob had never been exposed to 6…e5. Since I had
expected 1.e4 from Jakob my preparation in the King’s Indian was not thorough, and I ended my preparation here by just concluding “play 6…e5 and see what happens”.
6…e5 7. dxe5 dxe5 8. Qxd8 Rxd8
Here, I was out of book. Jakob is a player I would definitely describe as tactically inclined. I have no idea why he thinks this variation would suit his style. Whenever White takes on e5 and we get the Exchange variation structure all I think about is how to exploit the d4-square. White’s three pawn moves in the opening (d4, c4 and e4) claim a lot of space, but they create a wound on d4. It is normally a wound that can be conceiled by grabbing space with d4-d5 at some point whenever Black contests the White center by e7-e5 (but even here there can be ways to exploit that square through f7-f5 exf5 Nxf5 placing this knight on d4, or
f7-f5 followed by f5xe4 Nc3xe4 Ne7-f5 again placing this knight on d4). In the Exchange variation structure the d4-square is an open wound, giving Black a fixed target of attack. It is not that Black is better here, but having a
fixed target of attack makes it easier to play chess. When I get this pawn
structure I always think of Beliavsky-Kozul, and if you don’t know that game, please play it through to see a model game for Black. I have heard that some White players use the Exchange variation structure to scare Black players away from employing the KID, because the queens can be exhanged “and that does not fit with the style of the KID player because the KID player needs the queen to attack with”. I think the KID player should be able to play without the queen.
Look at the position: This is not a boring endgame. It is a queenless
middlegame position where almost all pieces are on the board. I have played
the Berlin of the Ruy Lopez quite a lot, and because of that I have the
highest respect for queenless middlegame positions. If these kinds of
positions are treated superficially by saying “the queens are exhanged, it
will probably be a draw”, then you are not going to find the right moves.
This is the more human move. It took me a while to play it, because I
had to figure out what to play against 10.Bg5 followed by 11.Bxf6 and 12.b4
threatening 13.c5, but I recalled a tactical theme from Opening Simulator that I could use to verify that Black had an okay position. When I got home I
checked up on what Kotronias recommended for Black, and he recommends 9… Rd7! It seemed so unnatural to me that I did not seriously consider it during the game.
10. Bg5 Rd6 11. Bxf6 Bxf6 12. b4 c6 13. c5 Rxd5-+ (13… Re6 14. Nxf6+ Rxf6 15. Bxa6±) 14. exd5 e4-+ was the tactic I recalled from Opening Simulator.
How to exploit d4 is the key. First, I found it beneficial to exchange the knight on f3, because it guards d4.
When White chooses to move twice in the opening with the same piece I
think Black can often look forward to a good game. It is not without reason
that Nimzowitsch wrote “Each piece should be developed with only one move” (My System). Of course, these statements are not considering this exact position, but I think it is worth thinking about.
11… Nxe4 is a move the computer suggests, and to be honest I did not consider this during the game.
12.Bxf6 Bxf6 13. Be2 Bxf3 14. Bxf3 Bg5+
This check is made in order to drive the white knight away with c7-c6.
15. Kb1 c6 16. h4 Bd8 17. Ne3 Rxd1+ 18. Bxd1
18. Rxd1 to try to exploit the open file drops the h-pawn followed by
the pawn on f2, and the rook in the d-file seems premature because Black’s
knight can protect b7 from c5.
Guess where this knight is going?
19.Bc2 Ne6 20. Ng4 Bc7
To protect e5.
Jakob is what I would describe as a “doer”, always looking for ways to counterattack and bring the position out of balance. There is nothing wrong with his moves here, but we see in some moves from now that White is opening the h-file, and I will exploit that later. So, I would not call his plan coherent. That is what can happen when the position starts to become uncomfortable.
21…Rd8 22. Kc1 Kg7
I wanted to play f7-f6 to make sure no counterplay would happen with a White h5-h6+ followed by Ng4-f6. If I play f7-f6 I would be able to answer h5-h6+ with Kg7-f7. It is probably seeing ghosts, because the White counterplay is non-existing. Maybe my previous losses against Jakob was a factor here, making me very cautious. Jakob is a difficult guy to control.
23. g3 f6 24. hxg6 hxg6 25. Rd1 Bd6!
A good move, because in a little while I will place a piece on d4, and then
White’s control of the d-file is not worth anything. I am pressing here, so
exchanging the rooks would facilitate the defence.
26. a3 a5
It made sense to contain White.
27. Rd3 Nd4 28. Bd1 Rh8
Now it seems that opening the h-file was not something White should have done.
29. f4 Rh1 30. fxe5 Bxe5 31. Nxe5 fxe5 32. Kd2 c5 33. Kc3 Re1 34. b4 b6
As I was pressing and White had no counterplay it made sense to keep pawns on the board and not take the b4-pawn with my a-pawn.
35. bxc5 bxc5 36. Kd2 Rxe4=
When I got home and saw that this position is a draw I had a hard time believing it, but it seems to be a draw. I have to say that I am not sure where I should have improved earlier on, and it shows that the drawing margin in chess is sometimes huge. I thought this was a slow and sure process towards the win, and it can make me very humble when I realize this is probably a draw.
37. Re3 Rxe3 38. Kxe3 Kf6 39. Ke4 Ke6 40. Bg4+
40. g4! Kd6 41. Ba4= My evaluation in the game was that White would never be able to block the position, because he would be hit by Ne6-g5+, but 41…Ne6 42. Be8= is annoying. If I play the pawn to g5 I can’t give a check on g5, and White can block the position. If I give a check on g5 the g6-pawn is lost. Playing the knight to f4 to protect g6 does not drive the White king away.
40… Nf5! because it would prevent White’s plan with g4 and the blockade.
41. Bd1 Nf5 42. Kf3
Now the position seems very lost, so I will give the rest of the game with no comments.
42. g4!, and it will be difficult to break White.
42… Ke6 43. Kf2 Nd6 44.Be2 e4 45. g4 Kf6 46. Bf1 Kg5 47. Kg3 Nf7 48. Bg2 e3 49. Kf3 Ne5+ 50. Kxe3 Nxc4+ 51. Ke4 Kxg4 52. Bf3+ Kg3 53. Be2 Nxa3 54. Kd3 Kf4 55. Kc3 Ke3 56. Ba6 c4 57. Kb2 Kd2 58. Kxa3 c3 0-1