Life In 19x19
http://www.lifein19x19.com/

Training tests
http://www.lifein19x19.com/viewtopic.php?f=10&t=9434
Page 1 of 1

Author:  ThousandAtaris [ Wed Nov 27, 2013 4:02 am ]
Post subject:  Training tests

How about we actually start testing these different ways of training and try to get some kind of clue what's the fastest way to improve. Suggestion for a standardized test:
Train everyday for 30 minutes and then play 1 game lasting for about 30 minutes for at least a month.
Don't do more.
For more information it might be useful to play lots of games on the last day.

Possible ways of training:
1. Briefly look at pro games with no mind. Ask John Fairbairn for more information. I would like to know myself how fast one should exactly go over one game.
2. Go over pro games and try to guess what they would do next. The basic idea is to think as much as possible.
3. Tsumego

Author:  RBerenguel [ Wed Nov 27, 2013 4:21 am ]
Post subject:  Re: Training tests

There's a problem with this, and it is that no-one is made equal. My raw, untrained forte may be strategy (it is not, I think, but bear with me in the example) whereas yours may be tactics. If I focus on tsumego for training and ignore everything else I may improve very slowly (since I'm playing against my natural inclinations,) but if you focus on tsumego you are playing with your strengths and would rise faster. Of course, as soon as the natural abilities plateau arrives you need to switch to another training focus. The thing is, that a training study like this would need:

    * Relatively large population of certain, baseline players in general
    * Large populations using a specific training regime (to make sure an even distribution of natural abilities)
    * A very long time commitment

Unlikely, overall.

TL;DR: What works for me is not the same that would work for a young Dosaku or what would work for you

Author:  leichtloeslich [ Wed Nov 27, 2013 4:42 am ]
Post subject:  Re: Training tests

That "standardized test" will be next to meaningless. As RBerenguel pointed out, to draw any meaningful conclusions from the data gathered you need at least a reasonably long timeframe and a reasonably large samplesize. (Those are the two major dealbreakers, but there are others.)


Anyway, I wasn't even aware there was any real debate over what the best training method is. I thought pros/strong amateurs unilaterally agreed that it's doing tsumego and playing games.

Author:  RBerenguel [ Wed Nov 27, 2013 4:50 am ]
Post subject:  Re: Training tests

leichtloeslich wrote:
Anyway, I wasn't even aware there was any real debate over what the best training method is. I thought pros/strong amateurs unilaterally agreed that it's doing tsumego and playing games.


I guess there would be debate on many things related to doing tsumego and playing games. For example, many simple problems in a short time, complex problems with long time limits, and so on and so on. But this is just food for thought...

Author:  ThousandAtaris [ Wed Nov 27, 2013 5:04 am ]
Post subject:  Re: Training tests

RBerenguel wrote:
There's a problem with this, and it is that no-one is made equal. My raw, untrained forte may be strategy (it is not, I think, but bear with me in the example) whereas yours may be tactics. If I focus on tsumego for training and ignore everything else I may improve very slowly (since I'm playing against my natural inclinations,) but if you focus on tsumego you are playing with your strengths and would rise faster. Of course, as soon as the natural abilities plateau arrives you need to switch to another training focus. The thing is, that a training study like this would need:

Untrained forte strategy? How could that even be possible. Anyway you should focus on your weaknesses. So if somebody has already solved a lot of tsumego it might not be as useful to do that even more. So maybe participants should also tell us how they have trained previously.

Author:  ThousandAtaris [ Wed Nov 27, 2013 5:07 am ]
Post subject:  Re: Training tests

And also this one month test is better suited for kyu players. Easier to see results.

Author:  RBerenguel [ Wed Nov 27, 2013 5:09 am ]
Post subject:  Re: Training tests

ThousandAtaris wrote:
RBerenguel wrote:
There's a problem with this, and it is that no-one is made equal. My raw, untrained forte may be strategy (it is not, I think, but bear with me in the example) whereas yours may be tactics. If I focus on tsumego for training and ignore everything else I may improve very slowly (since I'm playing against my natural inclinations,) but if you focus on tsumego you are playing with your strengths and would rise faster. Of course, as soon as the natural abilities plateau arrives you need to switch to another training focus. The thing is, that a training study like this would need:

Untrained forte strategy? How could that even be possible. Anyway you should focus on your weaknesses. So if somebody has already solved a lot of tsumego it might not be as useful to do that even more. So maybe participants should also tell us how they have trained previously.


It's easily possible, having natural aptitude for something is the rule, not the exception. I am a very slow runner, but have very good muscle endurance. So I won't run sprints, I may run a 10k. I may train sprints to improve my speed, but devoting all my time to becoming a sprinter would ruin my real ability of long running.

Author:  ThousandAtaris [ Wed Nov 27, 2013 5:27 am ]
Post subject:  Re: Training tests

RBerenguel wrote:
ThousandAtaris wrote:
RBerenguel wrote:
There's a problem with this, and it is that no-one is made equal. My raw, untrained forte may be strategy (it is not, I think, but bear with me in the example) whereas yours may be tactics. If I focus on tsumego for training and ignore everything else I may improve very slowly (since I'm playing against my natural inclinations,) but if you focus on tsumego you are playing with your strengths and would rise faster. Of course, as soon as the natural abilities plateau arrives you need to switch to another training focus. The thing is, that a training study like this would need:

Untrained forte strategy? How could that even be possible. Anyway you should focus on your weaknesses. So if somebody has already solved a lot of tsumego it might not be as useful to do that even more. So maybe participants should also tell us how they have trained previously.


It's easily possible, having natural aptitude for something is the rule, not the exception. I am a very slow runner, but have very good muscle endurance. So I won't run sprints, I may run a 10k. I may train sprints to improve my speed, but devoting all my time to becoming a sprinter would ruin my real ability of long running.

Running is a different thing. Also training for month in a different way doesn't hurt any amateur. It could actually be totally opposite and prevent a possible injury. And hey nobody is forced to participate in any of these test anyway.

Author:  tchan001 [ Wed Nov 27, 2013 5:39 am ]
Post subject:  Re: Training tests

ThousandAtaris wrote:
Possible ways of training:
1. Briefly look at pro games with no mind. Ask John Fairbairn for more information. I would like to know myself how fast one should exactly go over one game.

Not sure how successful you'll be looking for JF to answer all your questions. I doubt he has the time or interest in helping you train this way. Now if you were to hire someone like RJ as a teacher, I'm sure he'd be able to offer you a lot of detailed information :)

Author:  RBerenguel [ Wed Nov 27, 2013 5:45 am ]
Post subject:  Re: Training tests

ThousandAtaris wrote:
Running is a different thing. Also training for month in a different way doesn't hurt any amateur. It could actually be totally opposite and prevent a possible injury. And hey nobody is forced to participate in any of these test anyway.

Running is not a different thing, since we are basically talking about training regimes. I could have used chess, soccer or acing exams. All apply as well as anything else.

I agree that changing regimes for a month should not be a problem, but I actually use all three methods, daily if I can. While also playing some serious games in the NGA. What is helping me? Everything. I know tactics is my biggest blind spot (as well as deciding where to push my reading further and when to refrain,) and my shapes are sometimes too flimsy. But I don't want to focus on just tsumego (even if previous experience showed I get an edge when doing tons) or just pro games (for shape.) I try to fill all holes slowly, while improving also where my more or less natural good points are (it used to be the opening directions, but looks like that would only take me to 5k and now I need to work on it a little better.)

I guess standardised tests for a lot of people are pointless, and would rather suggest each person try their own regime and if possible, A/B test against himself (kind of.)

Author:  ThousandAtaris [ Wed Nov 27, 2013 6:23 am ]
Post subject:  Re: Training tests

RBerenguel wrote:
ThousandAtaris wrote:
Running is a different thing. Also training for month in a different way doesn't hurt any amateur. It could actually be totally opposite and prevent a possible injury. And hey nobody is forced to participate in any of these test anyway.

Running is not a different thing, since we are basically talking about training regimes. I could have used chess, soccer or acing exams. All apply as well as anything else.

I agree that changing regimes for a month should not be a problem, but I actually use all three methods, daily if I can. While also playing some serious games in the NGA. What is helping me? Everything. I know tactics is my biggest blind spot (as well as deciding where to push my reading further and when to refrain,) and my shapes are sometimes too flimsy. But I don't want to focus on just tsumego (even if previous experience showed I get an edge when doing tons) or just pro games (for shape.) I try to fill all holes slowly, while improving also where my more or less natural good points are (it used to be the opening directions, but looks like that would only take me to 5k and now I need to work on it a little better.)

I guess standardised tests for a lot of people are pointless, and would rather suggest each person try their own regime and if possible, A/B test against himself (kind of.)

I think it's best to gather all data. So post all the relevant information. Still I like standardised tests because humans are really pretty similar.

Author:  Bill Spight [ Wed Nov 27, 2013 7:16 am ]
Post subject:  Re: Training tests

leichtloeslich wrote:
That "standardized test" will be next to meaningless. As RBerenguel pointed out, to draw any meaningful conclusions from the data gathered you need at least a reasonably long timeframe and a reasonably large samplesize. (Those are the two major dealbreakers, but there are others.)


Anyway, I wasn't even aware there was any real debate over what the best training method is. I thought pros/strong amateurs unilaterally agreed that it's doing tsumego and playing games.


Honinbo Shusai said it was to play over 1,000 professional games. :)

Several years ago for its New Year's issue Kido magazine had all (or almost all) of the Nihon Kiin pros give a word of advice to amateurs on how to get stronger. As I recall, the main advice was to play go. Some pros just said to play a lot. Some said to get a rival. Some said to play against stronger players. (Taking three stones was the usual advice of this kind.) Some said to play thickly. (I guess that belongs in this category.) Some advised doing problems. Some specified tsumego problems. That's all I recall.

A classic series of books aimed at kyu players is Sakata's Killer of Go series, which is still in print, last I checked. It has five volumes. The first is "The Killer of Go", then Go no Tesuji to Zokusuji (Tesuji and Anti-suji), then a volume on the fuseki, then one on joseki, then one on yose and tsumego. Tsumego does not even get a whole volume.

There are many skills in go. I find the current emphasis on tsumego interesting, but curious. :)

Author:  xed_over [ Wed Nov 27, 2013 7:54 am ]
Post subject:  Re: Training tests

tchan001 wrote:
ThousandAtaris wrote:
Possible ways of training:
1. Briefly look at pro games with no mind. Ask John Fairbairn for more information. I would like to know myself how fast one should exactly go over one game.

Not sure how successful you'll be looking for JF to answer all your questions. I doubt he has the time or interest in helping you train this way. Now if you were to hire someone like RJ as a teacher, I'm sure he'd be able to offer you a lot of detailed information :)

I read that to mean asking John about the "no mind" part. Not about training per se.

Author:  SoDesuNe [ Wed Nov 27, 2013 8:15 am ]
Post subject:  Re: Training tests

Bill Spight wrote:
A classic series of books aimed at kyu players is Sakata's Killer of Go series, which is still in print, last I checked. It has five volumes. The first is "The Killer of Go", then Go no Tesuji to Zokusuji (Tesuji and Anti-suji), then a volume on the fuseki, then one on joseki, then one on yose and tsumego. Tsumego does not even get a whole volume.


"Killer fo Go" and "Tesuji and Anti-Suji of Go" are available through SmartGoBooks (http://gobooks.com/). Other Go books from Sakata translated to English are the two volume "Modern Joseki and Fuseki" (maybe a rearrangement of the japanese books on Fuseki and Joseki?) and "The Middle Game of Go", which is a follow-up on "Modern Joseki and Fuseki". The last three books can still be obtained through amazon.com for instance.

Author:  leichtloeslich [ Wed Nov 27, 2013 11:09 am ]
Post subject:  Re: Training tests

Bill Spight wrote:
Several years ago for its New Year's issue Kido magazine had all (or almost all) of the Nihon Kiin pros give a word of advice to amateurs on how to get stronger. As I recall, the main advice was to play go. Some pros just said to play a lot. Some said to get a rival. Some said to play against stronger players. (Taking three stones was the usual advice of this kind.) Some said to play thickly. (I guess that belongs in this category.) Some advised doing problems. Some specified tsumego problems.

I equate tsumego with general reading-problems, so that's a language issue on my side.
(A "reading-problem" is one where the answer can actually be read out and does not rely on experience. This would exclude most fuseki- and a lot of joseki-problems, but include tesuji- and yose-problems.)

Bill Spight wrote:
There are many skills in go. I find the current emphasis on tsumego interesting, but curious.

Well, Go is a fighting game. It seems natural to emphasize fighting in training. (Which is what tsumego is about.)
The success of aggressive Korean/Chinese fighting styles seems to validate this emphasis.

And lastly, a quote from one of the books you mention, "Killer of Go":
Sakata Eio wrote:
It is necessary to cultivate one's skill at solving questions at a glance, in other words, life and death problems (tsumego). It is recommended that amateurs assimilate the "Gokyo Shumyo" (...). It was published more than 150 years ago, but it contains basic life and death shapes, and is invaluable in game situations. Unless one first fixes the complete Shumyo collection in mind, one will not attain the qualifications to make one feared as a "killer". However, even a fairly strong amateur will have trouble solving all of the problems in the Shumyo correctly on the spot. If not, that's quite all right. After some days pass, one must take up the challenge again, and then yet again later, until in the end one masters the whole anthology.

from here.

Author:  Bill Spight [ Wed Nov 27, 2013 12:42 pm ]
Post subject:  Re: Training tests

SoDesuNe wrote:
Bill Spight wrote:
A classic series of books aimed at kyu players is Sakata's Killer of Go series, which is still in print, last I checked. It has five volumes. The first is "The Killer of Go", then Go no Tesuji to Zokusuji (Tesuji and Anti-suji), then a volume on the fuseki, then one on joseki, then one on yose and tsumego. Tsumego does not even get a whole volume.


"Killer fo Go" and "Tesuji and Anti-Suji of Go" are available through SmartGoBooks (http://gobooks.com/). Other Go books from Sakata translated to English are the two volume "Modern Joseki and Fuseki" (maybe a rearrangement of the japanese books on Fuseki and Joseki?) and "The Middle Game of Go", which is a follow-up on "Modern Joseki and Fuseki". The last three books can still be obtained through amazon.com for instance.


Modern Joseki and Fuseki and The Middle Game of Go are both harder than the Killer of Go series.

Author:  SmoothOper [ Wed Nov 27, 2013 4:21 pm ]
Post subject:  Re: Training tests

RBerenguel wrote:
There's a problem with this, and it is that no-one is made equal. My raw, untrained forte may be strategy (it is not, I think, but bear with me in the example) whereas yours may be tactics.


Not the strategy vs. tactics straw man again. Fundamentally, choosing what tactics to train on is a strategic decision, since if you are studying one set of tactics, you can't very well simultaneously study another set of tactics.

Now take your pick, tsumego, joseki, Fuseki, semiae,

Recently, I have been doing the at a glance series, I don't think it helps my raw reading as much as my recall.

Author:  Monadology [ Wed Nov 27, 2013 5:05 pm ]
Post subject:  Re: Training tests

SmoothOper wrote:
RBerenguel wrote:
There's a problem with this, and it is that no-one is made equal. My raw, untrained forte may be strategy (it is not, I think, but bear with me in the example) whereas yours may be tactics.


Not the strategy vs. tactics straw man again. Fundamentally, choosing what tactics to train on is a strategic decision, since if you are studying one set of tactics, you can't very well simultaneously study another set of tactics.


Why are you equating strategy in selecting your training method with strategy in a game of Go? These are obviously not interchangeable. You could be good at one but not the other.

Additionally, the strategy/tactics distinction as I understand it is basically a global/local distinction. Surely, then, it takes tactics to choose which tactics to study over a short period, but strategy to coordinate which tactics you study and in what ways across a great period of time. Also it's unclear whether you are arguing that the two aren't distinct or that strategy is always more fundamental than tactics.

Finally, how is anything in RBerenguel's post a straw-man? Who is he misrepresenting?

Author:  SmoothOper [ Wed Nov 27, 2013 5:37 pm ]
Post subject:  Re: Training tests

Monadology wrote:
SmoothOper wrote:
RBerenguel wrote:
There's a problem with this, and it is that no-one is made equal. My raw, untrained forte may be strategy (it is not, I think, but bear with me in the example) whereas yours may be tactics.


Not the strategy vs. tactics straw man again. Fundamentally, choosing what tactics to train on is a strategic decision, since if you are studying one set of tactics, you can't very well simultaneously study another set of tactics.


Why are you equating strategy in selecting your training method with strategy in a game of Go? These are obviously not interchangeable. You could be good at one but not the other.

Additionally, the strategy/tactics distinction as I understand it is basically a global/local distinction. Surely, then, it takes tactics to choose which tactics to study over a short period, but strategy to coordinate which tactics you study and in what ways across a great period of time. Also it's unclear whether you are arguing that the two aren't distinct or that strategy is always more fundamental than tactics.

Finally, how is anything in RBerenguel's post a straw-man? Who is he misrepresenting?


He is misrepresenting those poor tacticians, who keep fighting with the strategists, who don't exist.

Author:  EdLee [ Wed Nov 27, 2013 5:41 pm ]
Post subject: 

In addition to RBerenguel's, leichtloeslich's and Bill's first replies,
as usual (i.e. as always), nobody mentions age as a factor.

ThousandAtaris, it's a good quest, but extremely difficult to pull off. Good luck. :)

Page 1 of 1 All times are UTC - 8 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/