Am I an Authority?
- Rarely, but sometimes. I know some things about Go well, but not many. I know a lot of things at a poor level of rigor. It is an open question what my rank is. Perhaps at my strongest I am (was) AGA 2 dan. I've never made it above KGS 1 kyu.
- I have AGA dan-level friends who are KGS 4k-6k. This suggests that KGS ranks are stronger (is it true?).
Thoughts on Types of Go Problems
- It has been argued (e.g., by James Davies) that "status" problems are closest to real game situations because in a game nobody is telling you "that group can be killed" or "this one can live in ko" and so on. I agree with this and think status problems are good practice, to a point...
- However, the problem I find with status problems is that they do not always motivate me to push myself to the limit. That is, if I find a ko and am reasonably convinced that is the best that can be done I might not push myself to keep reading to find the kill (or life, as the case may be). So...
- I also like the more traditional "Black to live" and "White to kill" type problems. If I know there is a solution I will keep reading until I find it or until I run out of time (I usually only give myself about 10 minutes; anything that takes longer is too hard for my current level).
- I very much detest problems of the type "Black to capture something" or "White to do the best she can." I am left thinking, "what does that mean?" Or, I find something for Black to capture, only to look at the solutions and see that something bigger was possible. I admit the artificiality (in the sense of comparison to a real game) of telling the solver what the problem is but I like the motivation it provides, as described above.
Thoughts on How to Work on Go Problems
- I haven't studied or played much in the last year. I recently became inspired to return to studying Go, with a focus entirely on tsumego. I was inspired by the words found at BenjaminTeuber/Guide to Become Strong. BenjaminTeuber suggests giving oneself a time limit to work a problem and not looking at the solution. To do this one has to keep reading until one is confident one has the answer, so confident that looking at the solution is not necessary.
- While at first this sounded hard, I soon realized that for some level of problem we can all do this. It is just a matter of finding the right level of problem. So, I began with the simplest I could find and indeed I did not need to look at the solutions utill...
- After working on harder and harder problems, at some point they became too difficult. My mind had trouble keeping track of what I'd tried. After 10 minutes or so (the time limit I give myself) I do look at the solution regardless of how far I've gotten. If I haven't convinced myself that I've got it right by then I conclude the problem is too hard, I look at the solution, try to visualize it and see why it is correct, and then move on.
- Incidentally, this suggests a solution to the problem posed by unkx80/RandomStuff. There unkx80 laments that he can't tell when the problems he creates for beginners are too hard for for them. I say, it doesn't matter. If one just works a problem for 10 minutes and doesn't worry about if/when it is solved and then uses the solution to learn when necessary then it doesn't matter if the problem is too hard or not. In fact, the meaning of "too hard" goes away. No problem is "too hard" to do this. What is gained is 10 minutes of reading practice. That is enough. Eventually, with this practice, problems that could not be solved in this time become solvable. And that is the point afterall.
- Note that when I find that a collection of problems are all unsovable in 10 minutes (or become so after a certain point) I stop and find some simpler ones. This is mostly so that I reserve these harder ones for later when I will need the challenge.
- And finally, isn't the tsumego equivalent to a game situation the following: "Status? Do the best you can in a few minutes." That is, in a game one is not told the status (see previous section) and, moreover, one takes finite time to find one's move (usually a few minutes). Granted, it is not uncommon to spend some time of your opponent's thinking time to work things out and one may also use bits of time throughout the game on one group--working and working it until it is solved. In total how much time does this typically amount to? I say something of the order of 10 minutes. So the problem: "Status? Do the best you can in 10 minutes" is very good practice indeed.
Ultimately, to get better at Go, you just need to play more...
Notes on SL's Problems and Exercises
- Spot the Atari -- Completed 10/18/05. Of course these are very simple, 30k level problems.
- Beginner move function problems -- Completed 10/18/05. These are nifty problems for beginners as they bring up a lot of whole-board issues. A lot can be accomplished through process of elimination though. No need to revisit.
- fusojigen / problems -- Completed 10/18/05. Quite simple problems, though a few aren't so well constructed. No solutions given. No need to revisit.
- BeginnerExercises -- All solved as of 9/1/05, no need to revisit. Problems vary widely in difficulty. Not all beginner level. Solutions not of uniform quality either.
- KanazawaTesujiSeries -- In general, seem to get too hard for me around problem 47. Quit after problem 55 on 9/14/05. Should revisit in time. Problems seem to grow in difficulty but not uniformally. Solutions are exceptionally well written.
- OneDay1Problem -- These seem even harder than problems 47+ in KanazawaTesujiSeries. I failed to solve the first two and quit on 9/14/05. Should revisit after KanazawaTesujiSeries seem easy.
- KyuExercises -- Completed on 9/20/05. Some of these problems are very good for my level and would be worth reviewing in time. The solutions are not in good shape. Many are missing or not fully formed. Some present failed approches first, which I find annoying. I would like to point to the solutions in KanazawaTesujiSeries as models to follow.
- TsumegoFromGames -- Completed 9/27/05. Though I could not solve all of the problems I won't be returning to these because I find the series too poorly constructed. This set of problems vary considerably, both in difficulty and in presentation. The earlier problems are not posed as problems at all but more like illustrations of particular situations that arose. Later in the series, situations are posed as problems and solutions are provided on a separate page. Some problems are quite particular to the whole board situation and so, while they may be good practice, it is hard to take away a specific lesson for future use. Around number 49 or so the series presents problems from high dan and pro games with little or no solutions provided. It is really not a well made series of problems.
- TesujiFromGames -- Completed 9/27/05. A short series of decent problems. No need to revisit.
- SnapbackWorkshop -- Completed 10/5/05. These are relatively easy (though not all trivial) problems all with one variation invovling a snapback. The problems and solutions are quite well made. No need to revisit though.
- RandomTsumeGo -- Completed 10/13/05. A very nice set of problems,well-matched to my current strength. Enough were challenging that revisiting this series is warranted.
- CapturingRaceExercises -- Completed 10/13/05. With the possible exception of CapturingRaceExercise18, these are all fairly easy if one is familiar with the material in CountingLibertiesAndWinningCapturingRaces. No need to revisit unless I forget that material.
- DontPlayOnThePointOfSymmetry, BentFourInTheCornerExercises, BlindSpot -- Completed 10/17/05. No need to revisit. Recommend folding these into KyuExercises or some other series.
- UnderTheStonesProblems -- Completed 10/17/05. Since the title of the series is such a big hint, most of these problems are not too hard. But,the last several are quite difficult to read out and worth revisiting.
- ReadingProblems -- Completed 10/18/05. A very short series of not too hard reading problems, each of which has little or no branching. No need to revisit.
- Ladder problems and exercises -- Completed 10/18/05. These are quite good ladder problems, some not at all trivial. These are worth reviewing again sometime for reading practice.
- Kinokuniya, Completely artificial problem -- Completed 10/18/05. Two challenging problems which could easily be made part of some other series.
- Hikaru no go Problems -- Completed 10/19/05. A mix of simple and challenging problems. Worth looking at again in time.
- Hitachi Number 9, Hitachi Number 314 -- Completed 10/19/05. Not too hard.
- Large Avalanche Turn Outward Problem -- Not so much a problem as a misplaced BQM.
- Small Avalanche Exercise 1, Lucky's Problem, Xuanxuan Qijing Unnumbered Problem 2, RTG Problem 84, Valentine's Day problem -- Completed 10/20/05. All not very hard.
- Very artificial problem -- Hard (impossible?) and no solution given.
- Tim Hunt's problems -- Completed 10/24/05. Clever problems. Some soutions given. (The author explicitly states that he doesn't think it is his job to supply solutions, something I fundamentally disagree with.)
- UnderTheStonesProblem13 -- Completed 10/24/05. This page could be deleted without a significant loss since the problem is covered, more or less, in the UnderTheStonesProblems series. It isn't very hard.
- 3x4 Problems -- Skipping these for now. I don't play much 3x4 go.
- Classical life-and-death problems -- Completed 10/24/05. The only things new found here are FamousSymmetryProblems and even one of these is in another series (can't recall which). Basically, there is not a big contribution here.
- Classical Go Problems -- Very misleading link. It is a list of books!
- Fuseki Exercises -- Completed 10/24/05. In general these problems are very easy, though a few generated some interesting discussion.
- Hard Fuseki Problem -- A misplaced BQM it seems.
- Fuseki Problem 2 -- An interesting analysis to a subtle problem.
- Non-Joseki Exercises -- A few interesting thoughts found here.
- Middle Game Exercises -- I wouldn't call all of these exercises (only the last two perhaps). Typical of a lot of "exercise" collections here, it is chock full of what could better be included in the BQM series.
- Trigantius Archives -- Completed 10/25/05. These full-board problems are quite good and not so easy. I should revisit them.
- Don't help your opponent exercises -- Completed 10/25/05. Very easy problems, considering the title of the series and the fact that they are multiple choice.
- Beginners Endgame Exercises -- Completed 10/25/05. Fairly easy endgame problems.
- Practical Endgame Tests -- Skipped.
- Endgame Problems -- Completed 10/27/05. Some good problems, not too hard.
- Endgame from real games - 2002 Meijin -- Incomplete/nonexistent solution.
- GTL Review 1558 -- Skipped.
- Late Yose Problems -- Completed 10/27/05. These are musings on how to complete an incomplete pro-game record so as to match the reported score. There are a couple of interesting problems (not hard) and the rest are just mistakes by the page author or game record (not interesting).
- Endgame Problem 40 -- Yes, exceedingly picky and of no interest to me at the moment.
- Counting Problem 1 -- Another one of those picky calculating problems with fractional values that some folks enjoy and I do not.