Barcelona Go Tournament 2015

Written by alejo on February 24th, 2015
No Comments

Last weekend took place the “XXXIII Barcelona Go Seigen Tournament”. Class A, qualifying for the Spanish final. This year only 39 players joined the tournament, a little bit less than last year. However, there were more dan players than DDK players. Though you may regard this as good news, since the level of the players if increasing, some may argue that the amount of newbies has decreased. Personally, I had some trouble trying to settle my initial ranking.

Last year I joined as a 10k and got 3 victories and 2 defeats. I lost against a colleague of mine who is 9k by 1.5 points. During this year I’ve improved very little. Meanwhile, this colleague went to Finland and played an EGF tournament as a 5k and got reasonably balanced results. Recent games against him have been about 60% winning rate for him. So, I’d say I’m around 7k. But, in Spain, rankings between 5k and 12k are full of people with an unstable ranking. Players with very rare attendance to tournaments. Actually, last year’s 10 kyu players were actually people between 12k and 5k. In the end, my ranking remained unchanged.

The first game I had a 11k player as an opponent. As always, my fuseki left me in a bad position, so I made a terrible invasion and the bottom and my opponent did at the top of the board. The two weak groups collided and I won the capturing race thanks to an eye-stealing tesuji.

Second game 

Played against a 10k. Again, terrible fuseki leaves me with a weak central group and black with lots of influence in a quarter of the board. Luckily, I manage to create some tsumego madness and my opponent misreads the black’s 60 move tesuji. Moves around the K10 area are blurry in my memories. Black actually left white with very few liberties and white escaped through K9… somehow. White got a nice moyo in the bottom and right side, but black invaded the corner. Close to the end, black was winning by around 5 points.



However, the D18 stone has lots of aji. More than I saw at first glance. After a few minutes fiddling with the idea of invading the top left corner, I played D18. Below you have the sequence.  


It’s now, that I’m writing this post, that I’ve realised that if black had played 6 instead of 2, white wouldn’t have been able to reduce black so many points. And thus, black would have won. 


Third game  

First game against a 8k. My opponent started very strong at the fuseki but messed it up during a ko fight. It involved a wall of his main moyo and the life and death of the wall. We estimated the ko to be worth around 40-45 points. This was his ko threat. Interesting idea, but this time it didn’t work. The game was pretty much over after this sequence.  



 Fourth game  

Improving through the 8 kyu players. The game started with me starting to create a moyo at the bottom. An early invasion by my opponent resulted in a life-and-death situation.

20150222a1 White just played the marked stone. There is a battle going on between L7 and M7. After a few stones both groups escaped and black tried to start a capturing race between K4 and L4, by playing at G2. White descends and blocks the connection. Later, a white stone at M2 creates both a snapback and steals the eye-shape.


A few move later, black creates a few walls towards the centre. However, white manages to reduce through K10 and makes two small living groups on the edges of the board.  


Fifth game

At last the fifth game, against the 8 kyu with best results. This was actually the nicest game of all since I didn’t start too badly. In fact, my opponent played too close to my Kobayashi opening, so this allowed me to pincer. I created two nice moyos. On the other hand, I didn’t handle properly the 3-3 invasion of my opponent on the bottom left right corner and it allowed him to both reduce my territory and create a nice wall. I tend to neglect the F4 hane. Wrong. Terribly wrong.   Most of the game is available below, but first I’d like to point out this sequence. After some counting, my opponent was ahead till I found this sequence. It was actually white’s mistake to answer at B. But we were just playing the endgame… so this was somewhat unexpected. You can find the results in the SGF file around move 194.




In the end, I got a book as a present for my good performance. Others claim I’m a sandbagger. Next year I’ll be into the single digit kyu league. Though I’m certain I’ll lose some stones…

By the way, the local club, La pedra, recently moved to a new place, more fashionable, with more space, less dusty… so if you ever happen to be in Barcelona, visit them.

Tags: , , , ,

From pictures to sgf

Written by alejo on February 18th, 2015
No Comments

As I already mentioned in my previous post, I’ve been working on the issue of recording my games while playing. Even if everyone claims that learning to memorise them during the game would help me improve, I’m not able to go all over a game by myself. So, ever since I found out about Photokifu, I’ve been trying to get it to work as automatically as possible. Did I already mention that Photokifu is free?

Today I replayed the “ear reddening” game and Photokifu detected most of the moves. I had to actually tune about 5 of them, but everything else was fine.  252 correct guesses out of 257 is good enough for me since the program will ask you whenever it encounters any abnormal situation and you can manually fix it. And yes, the game is longer than 257 moves, but by the end there is a long ko and I was tired so I decided to finish the test.


In order to accomplish this you’ll need a couple of devices:
1. Standing stick: this allows the phone to be stabilized on a upper position, with a good view of the board.
2. Remote shutter: this Bluetooth device allows you to remotely control the timing of the pictures.

My first attempt excluded the remote shutter and was made through a timelapse app for android. Handling hundreds of pictures and deleting the duplicate images was quite cumbersome. The results were not optimal.  The remote shutter allows me to shoot after a stone has already been placed and there is no hand over the board.

Unfortunately, I couldn’t find an app capable of taking pictures while the screen was off and remotely controlled by Bluetooth. So, mind you, the screen must be on during the entire game. My numbers on a One plus one: 257 pictures, one every 10 seconds (i.e. 42 minutes), on airplane mode, with wireless connection off, with Bluetooth on and screen brightness around 20%, the battery drain was around 10%.

Not that bad, I expected more battery drain. However, the remote shutter is somewhat noisy, with the classical click-click of cheap plastic buttons. And the stick is bulky…




Automatic sgf creation from a game

Written by alejo on February 12th, 2015

softwareDuring the last few years players have been struggling in order to get their games recorded in tournaments. Though dan players memorise them in-game, kyus may not be able to do so, yet. I’m joining a tournament in a couple of weeks. Instead of studying go, I’ve done some research about how to record my games. Well, actually recording them by myself on the phone isn’t a good idea. Saving every move on my phone during the game actually drains my concentration, so it’s discarded. Let’s look at the actual solutions right now:

1.Photokifu:  a windows software which allows you to create sgf files from pictures. Though you may manage to get an sgf file created, it’s not as easy at is seems. Board edges should be parallel to the photo margins. You have to remove the duplicated pictures before submitting it into the program. However, there is the neat function of discovering more than one move at once. All in all, it’ll take you some patience in order to get good results.


2.Kifu snap for android: take a picture from the board and it’ll calculate the score, it won’t generate any sgf files. This program will force you to join its subscription system and then, if you don’t like it, cancel out within the first 10 days. It won’t open my phone’s camera. Unfortunately, there is no way I can give it less than 1 star on Google Play.

3.Imago: a thesis on go image recognition from Tomas Musil. It’ll work on unix systems and you may need some configuration.

4. Kifu: a project back from 2009. Abandoned. Files not publicly available

5.Photokifu for iphone: I don’t have an iphone. Discarded.

For this tournament, I’m going to try Photokifu. I’ll have a tripod by the table with my cell phone on it. Taking a picture every X seconds (still not decided) with time lapse apps. I’ve checked that a picture every 4 seconds during an hour on airplane mode drains 10% of the battery. Nice. This will generate hundreds of  pictures. With Photoshop I’ll batch crop and resize the picture to show only the board. Then run a script to remove duplicate images. Finally, use Photokifu with the resulting images.

Yes, I know, this looks far away from automatic. I’d better learn to memorise them during the game. I just hope I don’t end up just manually checking every 10 images and creating the SGF by myself.

Tags: , , ,

Rabbity shape is alive!!

Written by alejo on February 7th, 2015
No Comments




Tags: ,


Written by alejo on January 10th, 2015
1 Comment

gobanDuring the last few weeks I’ve settled as a 4 kyu in KGS. I remember being told the difference among ranking increases the stronger you are: there is more difference between a 3k and a 4k than a 9k and a 10k. Though I was already certain of this, it hadn’t hit me in my face till I reached 4k. Once you get to this level you start to play against 3 kyu players… and the game turns into slaughter.

Though it may be sheer coincidence, 3 out of the 4 games I had against 3 kyu players had something in common: unusual josekis (at least for me). We all have seen unusual josekis played by dan players. However, during the weaker SDK phase, players tend to stick with the most traditional openings. While playing a game, we try to settle traditionally the corners and avoid messing up with the josekis. Strong SDK are quite the opposite. In order to improve, they are trying different opening and josekis and learning from their experiences.

Unfortunately for me, I’ve had to deal with some avalanches and 5-4 corners, even some facing each other. Being used to the most traditional 4-4 or 3-4, the results I get in the corner while facing a 5-4 are anything but a fair result for both players (i.e. joseki). Either I end up surrounded or my groups killed or the result is somehow unequal. Therefore, the term weird-o-seki. It’s not that the sequence doesn’t exist, but it none of the classical ones among weaker SDK players and we get a bad result out of it.

The first time I faced a 5-4, it was a bloodbath. I tend to call it a shit-o-seki. However, it doesn’t make it for the post title.

Tags: , , ,

Round one: FIGHT!!! (or playing in Tygem)

Written by alejo on December 21st, 2014
No Comments

gobanA few weeks ago I read a couple of posts about Korean servers, mainly about Tygem and Wbaduk. Some other players had already advised me to try new servers, specially Tygem and Wbaduk because there are lots of professional games available. Since playing ranked games on both and establishing a rank in both of them would take me a while, I decided to join only Tygem. Dwyrin, on one on this videos, laughs at the fact that “Tygem is the server that you log in when you want to play a traditional and peaceful game”. In the end, the actual text that pushed me to give it a go is below these lines.

Wbaduk Korean players are youths who learn the game from their parents from very young and love to keep playing to win… They played quickly, impatiently and restlessly, fiddling with their stones during their opponent’s turn instead of thinking… But they were also strong fighters. While I was still ~5k at KGS, I played games on wbaduk, but after 20-30 games I found myself hovering around 15k(or weaker, I can’t remember) instead.

It made me think that pre-dan level, wbaduk kKoreansplay like lions: wild, fearless, bestial, powerful. KGS players play like military school graduates: scholarly, full of theories but without the experience of actual hard fighting.

Extract from Siowy in Lifein19x19 forum

As a KGS  player and, more or less, self-teaching western go player, I wanted to know and experience the difference by myself. I consider myself to around 5 kyu KGS, when in good mood. So I joined Tygem as 15 kyu. I lost the first game against a 18 kyu, giving him two stones of handicap. Mainly due to underestimating my opponent and a couple of overplays. Though this first game was lost by some silly mistakes of mine, I still felt amused by the difference of level between tygem and KGS. I could certainly not give 9 stones to this player and expect a victory.

The next game was against a 17k, which felt more like a 10k in KGS. I won, I could have given more handicap too… but still felt uneasy about some moves. After a couple of days, I played a 15 kyu. After my previous overplays, I decided not to underestimate my rival and played my best. He was really around 12-13k on KGS.

The last game I played was against another 15 kyu. Though I was playing my best, the game was close till the endgame. I lost after he invaded my territory and kept sacrificing stones till I made a mistake. This game summarises my experience in Tygem: they are strong fighters, play fast, solve tsumegos in a few seconds and exploit your weaknesses (even if you though you hadn’t). Pretty much like Siowy said: wild, fearless, bestial, powerful. However, I’m still waiting for a “traditional” joseki to show up and be played correctly its full length.

Tags: , , ,