Getting over your wall.

[Bill Spight]

And so here we are, talking about kyu problems on L19 like we're experts.

Don't worry, be happy.

[jts]

These numbers may not be particularly informative by themselves, since they don't say anything about what I was doing off of KGS (or for that matter, about the quality of the games I was playing on KGS). But if a dozen more people would chip in, I think we would have some interesting data to draw inept inferences from.

Note two things - this lists games played, not time elapsed. I'm not sure why time elapsed is treated as a particularly interesting metric in these discussions. Also, I've listed the games cumulatively, as the total number of games played on KGS up to the point where a certain rank was held securely, and for more than one game.

14k - 8
13k - 9
12k - 14
11k - 50
10k - 61
9k - 71
8k - 86
7k - 100
6k - 127
5k - 158
4k - 231
3k - 409
2k - 504

Now, I'm going to resist the urge to over-interpret this. There are certain anomalies that I could connect to certain changes in how I was playing, or a specific low-odds win, or whatever. But the real utility of this will be if more people contribute and we can figure out something like an average ratio of (cumulative games at 10k):(cumulative games at 5k).

[Boidhre]

Interesting jts, mine is quite silly because I generally got at least 2-3 games a week in offline against much stronger opponents than me when I was a ddk and this contributed a lot to my progress I think. Also I played occasionally on other servers, especially turn based ones. So a large fistful of salt is needed. I was curious about the numbers though. Mostly I am posting this as a cautionary example of interpreting these numbers, according to mind one barely needs to play go to become a sdk.

30k-20k: 20 games
20k: 4 games
19k: 2 games
18k: 6 games
17k: 11 games (here I started my habit of playing games against stronger players at a reduced handicap, good for learning not so good for winning a lot and gaining rank in a small number of games, it's the same at 15k)
16k: 5 games
15k: 11 games
14k: 1 game
13k: 3 games
12k: 6 games
11k: 2 games
10k: 4 games (around here my health deteriorated and I started playing a lot less online)
9k: 1 game
8k (?): 22 games (as in 22 games played on KGS on 19x19 since I had a stable rank of 8k and stable for me is just won a couple of games at that rank given I play so little. My rank has bounced between 5k? and 9k?, is currently 7k? and has been pretty much meaningless for a good while now)

Overall take home lesson for me looking at the above is that is a really miserable total of 19x19 games on KGS for a year's playing. If I restrict it to ranked games only it's a mere 74 games played. I didn't choose this for the numbers above because the free games I played, be they teaching games or friendly games should be counted I think. If I counted only ranked I'd be only having 2 games played at 15k which doesn't represent how much I was playing at all. Overall my numbers of games per rank are low because of KGS's rank inflation and me playing games offline and neglecting online play for periods. So I'd come back after not playing a ranked game for a couple of weeks but having been doing tsumego and playing offline games I had improved so often my rank would have inflated to say 14k from 15k and I'd play one game and progress to 13k, not play for a bit, play one or two games and hit 12k and so on. My KGS strength and my real world strength are not that well correlated most of the time, it both under and over estimates me depending on what I've been doing since I last played. However I think if you took out the KGS inflation and had something like the IGS or wBaduk systems you'd still have weird numbers from people who primarily play offline. Double promotions being more common for them and whatnot.

[wineandgolover]

I can't agree strongly enough with earlier advice. If you hit a wall, have a significantly stronger player review your games. Trust me, you are making 10+ mistakes per game. Identifying and eliminating just a few of them will shatter the wall.

And if that doesn't work, time off works wonders, too.

(Finally, don't be discouraged by those who never ran into a wall before gaining dan. They talk a good game, but you can catch them.)

[Bill Spight]

Trust me, you are making 10+ mistakes per game. Identifying and eliminating just a few of them will shatter the wall.

Good point.

And if that doesn't work, time off works wonders, too.

I have had the experience of getting better while not playing. (Not that it was voluntary, it was through moving to somewhere without go players. Before online go, OC. ) It is interesting to see that others have had similar experiences.

[Phoenix]

I have had the experience of getting better while not playing. (Not that it was voluntary, it was through moving to somewhere without go players. Before online go, OC. ) It is interesting to see that others have had similar experiences.

I've talked about this before on another thread, but I beat (read: obliterated) the vice-captain of my school's Chess team at a point where I was feeling more comfortable with Go.

The part that struck me in retrospect is that I have hardly played a dozen games of Chess in my lifetime! I had enough experience with competitive activities though to say that he was definitely a strong player. I remember John Fairbairn (I believe...) reporting a similar experience in another thread.

It's odd how there seems to be a level of synergy between fractionally related activities. I think the way of the future for competitors in mental games (in every sense!) will involve cross-training regimens of the sort. Once we figure out what's really going on backstage of course.

Edit: I had recently worked on my mental calculation around the same time. I could tell you the solution to a quadratic problem in seven seconds flat. Maybe some weird neurological combination? A branching of neural pathways? Who knows!

[Unusedname]

Trust me, you are making 10+ mistakes per game.

This reminds me of my favorite Go quote.

"A dan rank is the number of moves you make between each mistake.
A kyu rank is the number of mistakes you make each move."

I forget the exact words but this was the idea.

[Sumatakyo]

I just skimmed the answers and don't think this was suggested, so:

Play on another Go server.

I got beyond the KGS 5k "wall" by playing on Tygem.

[daal]

I can't agree strongly enough with earlier advice. If you hit a wall, have a significantly stronger player review your games. Trust me, you are making 10+ mistakes per game. Identifying and eliminating just a few of them will shatter the wall.

I'm sure that you are correct in saying that those of us facing a wall are regularly making mistakes (and as Unusedname points out, per move), but I'm not so sure that you're conclusion is always valid. The question for me revolves around eliminating mistakes. I've certainly identified a buttload of them, but aside from playing on the first line in the opening, I doubt that there are many that I would never make, so it comes down to what causes me to make them.

I won't go so far as to say that I've hit my last wall at 5k, but in my world I see a few passersby, but also quite a few regulars who like me have been slugging it out at this level for a year or more. In order to shatter a wall, one needs to start consistently beating players who are a stone or two stronger, and I suspect that the reason stuck people don't is that they don't believe they can.

In other words, the mortar that holds the bricks together can be psychological.

I keep telling myself for example that under pressure, I get easily flustered and my already modest reading abilities are reduced to the point where the probability of one game losing mistake or another is large enough to keep me from advancing.

I dunno, maybe I'm just spouting BS, but then again, it's not through lack of examining my mistakes that I haven't gotten stronger. I just keep making them.

[SmoothOper]

I here from a lot of players at club that you usually hit a wall around 5k. How do you get over said wall, and have you guys experienced this?

On IGS, I am about 5kyu, it does seem that there is a bit of a wall.

I don't necessarily know the answer, but some things I have tossed around are.

A) Need to be smarter, not just have better style.
B) The informational materials on strategy aren't comprehensive enough for higher levels. The books I have say something like this should be good up to about 1 dan, which was a 1 dan like 50 years ago.
C) Learn the standard sequences more thoroughly. At this level people seem to know the correct Joseki better, and also know how to punish the ones they use better. Which is a shame, because I just figured out how to punish the deviations for the ones I know.
D) Practice counting, I wish there were tsumego like problems for counting, because I find this skill doesn't develop in parallel with tsumego.

I have looked around there are number of books "Breaking through to shodan", "Raising to shodan", they all seem to be oriented towards Joseki sequences.

[Twitchy Go]

I here from a lot of players at club that you usually hit a wall around 5k. How do you get over said wall, and have you guys experienced this?

On IGS, I am about 5kyu, it does seem that there is a bit of a wall.

I don't necessarily know the answer, but some things I have tossed around are.

A) Need to be smarter, not just have better style.
B) The informational materials on strategy aren't comprehensive enough for higher levels. The books I have say something like this should be good up to about 1 dan, which was a 1 dan like 50 years ago.
C) Learn the standard sequences more thoroughly. At this level people seem to know the correct Joseki better, and also know how to punish the ones they use better. Which is a shame, because I just figured out how to punish the deviations for the ones I know.
D) Practice counting, I wish there were tsumego like problems for counting, because I find this skill doesn't develop in parallel with tsumego.

I have looked around there are number of books "Breaking through to shodan", "Raising to shodan", they all seem to be oriented towards Joseki sequences.

A) This is a little vague as to what you mean, but I think I disagree. There should always be fundamentals you can improve on. They are just 1d fundamentals instead of 8k fundamentals for example.

C) Just to play devils advocate here. Might the next step be choosing the correct joseki in the global sense? Not just playing the one you know correctly.

D) Have you considered looking into endgame problems? They require counting to solve, and if you can settle boarders quickly(by having practiced) you can count the whole board score better. Alternatively you could make your own problems. Take a pro game that goes to scoring and go X moves in and try to count the current score. Go Y moves a head and repeat. Finally you might try treating the last 10 moves like an endgame problem and try to work out the move order that results in the game score.

[Pippen]

If you hit a wall you gotta change something:

1. Check if your playing style suits you and/or if it has some fundamental defects, otherwise try new things.
2. Memeorize pro games (about 100-120) games.
3. Do some go-problems.
4. Stop worrying about your rank. If your brain is made for 2k then this is what it is. Trying and cursing to overcome this wall'd be like a cat trying to fly like a bird...some things are just not meant for you, learn to live with it or you will have a lot to grief about that is not to change anyway (which is stupid and a waste).

Point No. 4 is important, because it keeps the fun alive and you can keep perspective. Basically everybody should be able to make it to 4-5D if one does Go 5h/day with a pro as a teacher. But is this wise?

[SmoothOper]

A) This is a little vague as to what you mean, but I think I disagree. There should always be fundamentals you can improve on. They are just 1d fundamentals instead of 8k fundamentals for example.

Sure, a slam dunk might be considered a fundamental at some basketball level, but if you don't have the hops, you might need to hit the weight room.

C) Just to play devils advocate here. Might the next step be choosing the correct joseki in the global sense? Not just playing the one you know correctly.

Yes, but this sort relates to the dearth of literature in this area, most of the comprehensive strategy books state they aren't suited for 1D+ training. There are quite a few that have solved all the problems related to some strategy including joseki choice, and you don't always get to choose.

D) Have you considered looking into endgame problems? They require counting to solve, and if you can settle boarders quickly(by having practiced) you can count the whole board score better. Alternatively you could make your own problems. Take a pro game that goes to scoring and go X moves in and try to count the current score. Go Y moves a head and repeat. Finally you might try treating the last 10 moves like an endgame problem and try to work out the move order that results in the game score.

Yes, I have looked at endgame problems, that helps, I think the problem is I need to be able to count territory faster so that I can do it more often.

[Twitchy Go]

Sure, a slam dunk might be considered a fundamental at some basketball level, but if you don't have the hops, you might need to hit the weight room.

True, so do you mean that it is the execution of the concept not the concept itself that is(might be) the problem when facing a wall. For example, someone understands the idea of leaning attacks, but doesn't have the reading power to take advantage of the influence gained in the coming fight. If so I agree with you.

Yes, but this sort relates to the dearth of literature in this area, most of the comprehensive strategy books state they aren't suited for 1D+ training. There are quite a few that have solved all the problems related to some strategy including joseki choice, and you don't always get to choose.

While you don't always get to choose which joseki is played you can still influence what line it moves down and resist an unfavorable result. And I haven't looked at theory books in quite some time and can't comment on their effectiveness near/above dan level. However I have found Yuan Zhou's Single Digit Kyu Game Commentaries useful. As well as lectures and lessons from high dans/pros.

Yes, I have looked at endgame problems, that helps, I think the problem is I need to be able to count territory faster so that I can do it more often.

Here is an idea I've heard but haven't tried. Begin by picking arbitrary move numbers to count at. Estimate the score at move 30, 50, 100 and so on. In theory this won't take too much of your game time and you'll get better at counting as you continue doing it. And I think us SDK players really only need a rough idea of the point balance for the most part. Just enough to know if it is wiser to reduce or invade, contain or kill. So just taking a few soundings over the game seems good enough while you're getting the feel of counting I think.

[SmoothOper]

Sure, a slam dunk might be considered a fundamental at some basketball level, but if you don't have the hops, you might need to hit the weight room.

True, so do you mean that it is the execution of the concept not the concept itself that is(might be) the problem when facing a wall. For example, someone understands the idea of leaning attacks, but doesn't have the reading power to take advantage of the influence gained in the coming fight. If so I agree with you.

No, smarter is demonstrable in reading sometimes, but reading isn't the only payoff, back to the basketball analogy, if someone has some hops, they are also going to be able get extra rebounds, block more shots, and be more open for passes, be able to shoot over people etc. as well as be more efficient in practice in addition to dunking. In this way I mean smarter, not better at reading.