I've recently been reading about Memory Palaces http://en.wikipedia.org/wiki/Method_of_loci and after some experimentation I've been quite impressed by how well the work.

Recent discussions in the Tami's Way topic and the Intuition topic have prompted me to wonder whether there is any mileage to be gained by trying to apply this memory technique to Go.

If indeed one can store hundreds, or even thousands of facts, in a reliable and retrievable way, what should one attempt to store?

Just a checklist of things to remember? A pretty lengthy and detailed checklist could be kept.

A list of principles and techniques? Which ones?

Joseki sequences? An encoding method is not clear to me.

A 'shape' database? What does this even mean?

Has anyone tried this? If not, do you have any opinions or suggestions?

I'm going to butcher the joke, but I remember Keith Arnold joking about a "Japanese Name Dropping Diletante" several years ago. It was a fellow who couldn't find a tesuji to save himself over the board, but once he was shown it after the game he had memorized the correct Japanese name for it.

I have a similar vision every time someone talks about memory work in go. The game has so much variability and depth that it's going to swamp any memorization methodology. Better to learn to read those tsume-go out from scratch.

It really works best for collections of things where each item is pretty easy to remember and it's the number of things you need to remember or their order that is difficult. It also works best for things that are themselves easy to visualize, like faces - if you need to create a symbol to remember something, then how do you remember what it was a symbol for?

As I've been memorizing games I've been giving each game a location and a time. I seriously doubt that it helps to memorize them initially. I'll tell you in ten years if it helps them stick.

IMO, there is a difference between memorising, and knowing and understanding.

If you memorise joseki and tesuji shapes, without considering what function they perform or meaning they contain then, if my own experience is any yardstick, you will not only find it hard to remember but also be just as confused as ever.

If you study joseki, strategic principles and everything else, slowly endeavouring to understand how they are applied, and if you practice them in your own games, then doubtless you will come to know them and to understand them at a better level. Again, if my experience is of any service, I find it much easier to remember things I have studied with careful attention. Memorisation comes naturally when you`re enthralled by something!

For example, let`s take these "five priorities" to get you started in fuseki:

1) First, occupy corners
2) Then, Enclosures/Approaches
3) Extend in front of an enclosure/deny the extension
4) Play any checking extensions (mutually big points)
5) Finally, move into the centre or begin activities against the opponent

There is vastly, vastly more to fuseki than these five principles, but if you first study them in examples, and try them out for yourself, then they will give you something to work with, and you will be able to apply them with reasonable confidence. But if you simply memorise this list, without doing the study, then I suspect you will find yourself feeling pretty confused when you sit down to play a game.

After all, what do you do if the opponent interferes with your order of play? What if you cannot decide whether to make an approach or an enclosure? What if something completely unexpected appears?

I`ve already tried using checklists, "compasses", mind maps and all kinds of complicated memory systems to remember principles and moves and shapes, and found them to be a big burden. The quickest way to master principles is to study examples, to look out for how they are applied in all the games you see, to see what happens when somebody "breaks" a principle, and to attempt to apply them for yourself. If you do that, you should find you`ll come to remember them when needed.

For one more example, I`ll summarise the contents of one of Takemiya`s books for you:

1) Play where the opponent wants to play
2) Don't surround territory
3) Know the state of play (in Japanese, 自分をわきまえる, "Know yourself")
4) Don`t make bad shape carefree

To see what Takemiya means, you`d have to read the book and think about the examples. You`d have to try to apply those ideas to your own games. Without such a foundation, I am pretty sure those "principles" won`t mean anything much to you.

I believe the psychological reason is because a skill is basically created by forming very strong and robust neural connections. You can learn the principle, but without the extensive practice you just cannot obtain the powerfully and broadly wired support for it in the brain.

I`ve used this example before more than once, but it`s a good one: take alternate picking on the guitar. It`s extremely simple, just hold your pick (I won`t go into the details of how you do that) and move it up and down in time with the beat. Strike the strings on the downstroke or upstroke according to the rhythm you wish to produce. That`s it. You probably grasped that within a few seconds (provided you didn`t know it already), but I can assure you that it would take many, many hours of concentrated practice before your alternate picking could compare with, say, Steve Vai.

So, to conclude, I`d encourage you to use memory palaces to help with certain things, but to use honest-to-goodness study as your principal method.

_________________
Learn the "tea-stealing" tesuji! Cho Chikun demonstrates here:

There is a good chance to understand correct and good principles simply by reading them. Examples and application in games can be another way to understand principles. IMX, it depends on the nature, quality and presentation of a principle whether the fastest understanding comes from simply reading it, studying applying examples or other means towards an understanding.



Although you use this list as an example of principles a learner might be confronted with, repetitive statement of this or related lists are suggestive for beginners. They deserve a much better list of first opening principles, a list related to their by far most frequent opening mistakes:

- Avoid premature endgame.
- Choose the bigger space, direction or group.
- Do not play on neutral intersections.
- Take and deny the opponent from taking the valuable shape points.

While some learners of the first million digits of pi use symbolic encoding such as a city name or animal for every block of, say, four digits, the best "encoding" for joseki sequences is the understanding of their contained meaning: move meanings, Strategic Lines, group meanings, strategic choices, tactical choices etc.

I absolutely agree that I need to understand the principles or techniques or strategies or tactics involved, but...

In my analysis of my games, either by myself, or in reviews from stronger players, certain failures keep coming up. This list of key things to remember gets long, and I would very much like to remember them all, if only to remember to apply them in actual games to see how they turn out, which will help reinforce them.

To illustrate, here's a small subset of my personal list

- thickness is good, and avoiding cut points is important in the opening, but don't over do it
- when cut, think about which group is weaker - often this means the group with less liberties
- blocking a group in is often pointless, especially if playing on line 2 (unless there's a chance to kill)
- When pressured, try to get away without strengthening opponent
- When playing long extensions from corner, and opponent play a one space approach, think about stepping back (2 space extension) to the corner, rather than jumping out
- Don't overattack - take some territory when the opportunity arises
- Don't forget splitting attack - push opponent groups together, and then split them
- Don't be passive - look for active move.
- Don't miss the chance to make bad shape for opponent
- 8 point shape is already alive

All obvious things, when written down, particularly to stronger players. But I don't remember all these things and frequently one of these items (or items from my longer list) will occur, which may not cause losses directly, but will cause less than ideal results.

So I'm tempted to try and specifically remember this list, although as pointed out by jts it's quite hard to remember abstract things rather than facts, in the Memory Palace technique. The list above seems pretty abstract...

Actually the record for pi is some way short of a million Currently its about 68,000 digits. Although I'm sure they do something along the lines of what you're talking about.

In terms of joseki, if we assume its useful to memorize them (a questionable assumption, but lets ignore that), there's at least two problems: 1 - remembering all the moves, and presumably, all the variations (although these could perhaps be called different josekis), and 2 - remembering the justification/result of the joseki, i.e. why you would pick it.

Are you suggesting that somehow the 2nd problem could solve the 1st problem?

Note, btw, in the memory palace technique, you're supposed to heavily stylize the thing you're to remember, perhaps by dressing it in outlandish clothing, or by thinking of a ribald form of the thing. In general, this is easy to do with nouns, like, say, in a shopping list. When it comes to more abstract matters, I don't know enough to say what works. I've heard, for example, that one can use the syllables of the word involved, and have memorable entities associated with all the standard syllables. Whether that's viable I don't know.

This guy has some YouTube videos about memorising joseki using mnemonics.

http://www.youtube.com/user/BasicGo

_________________
Regards,

Peter

That was bizarre. I've never used these sort of memory techniques so maybe I am too quick to dismiss them, but aren't they for remembering random facts? Joseki aren't random facts, they make sense.

I don't think so. Have you tried to provide structure for the abstract things to be remembered? Thereby, you need to recall fewer things at first and need to recall details only whenever they are relevant.

Yes!

I watched a bit of this, but to be honest I think it would be easier (for me at any rate) to remember the purpose and meaning of each move rather than some quirky story about mothers, knights and teenagers. (And actually I did try something similar a long time ago - I called various sequences after characters from Buffy the Vampire Slayer, but I ended up forgetting them all!)

Very cute avatar, btw, PeterPeter!

_________________
Learn the "tea-stealing" tesuji! Cho Chikun demonstrates here:

So, I spent a fair amount of effort on this some months ago and learned a lot, but this thread is hijacked by people whose opinions may differ from what I found, so I'm reluctant to comment.

Doesn't seem hijacked to me yet? Anyway, I'd like to hear anything you found in relation to my original question.

When I was a kyu player, I avoided memorization. (Not that I took it up as a dan player. ) That was perhaps a mistake. I did not memorize joseki, because I did not study joseki. The only life and death problems I studied were in the Maeda series. Because there were not very many at my level, it was easy for me, when reviewing, to remember the answers without reading. Therefore I waited months between reviews, to give me time to forget. Perhaps it would have been better for me to drill them into memory, I don't know.

I think that the method of loci would be good for checklists. Acronyms also help. I still remember a checklist for the first play at contract bridge, devised by the great player, Oswald Jacoby. The acronym is ARCH. Analyze the opening lead. Review the bidding. Count tricks. How can the contract be made (or defeated)?

You may be interested in a book by Frances Yates, The Art of Memory. She discusses memory techniques from the early modern era and earlier, including the method of loci. It is a history book, not a how-to book.

One thing you can do if order is not important is to create an image in which different aspects of the image have meaning. For instance, armor might mean thickness, or a glove might mean honte. A severed limb might mean sacrifice.

Not that I use any of this in go or in life. The only memory technique I use is acronyms. But everybody is different. If you think that memory techniques might help you, try them out.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

I could not make it through the story. I left them on the dance floor.

Not that the story might not be a help. (I suspect that they could have made a better presentation.)

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

I'm reading Moonwalking with Einstein by journalist Joshua Foer, a possibly similar book - also about memory techniques and how memory works, and covers, among other things, his efforts to train for and compete at the USA competitive memory championship. It isn't very specific about methods but it's pretty interesting and well worth a read, assuming you like your books light and entertaining rather than very heavy going. He references the Yates book you mentioned quite a few times.

I have wanted to do this for a while, and I think it is a worthy experiment, though not a guaranteed success.

I've always liked remembering things, in the sense of having them memorized, particularly phone numbers, for their obvious utility, and poetry, for it's beauty and for something to recite while driving, swimming or doing the dishes. I didn't get at all systematic about it until last year when I read "Moonwalking with Einstein," a heck of a good read. Anyway, inspired by it, I tried a half dozen thing and the two that stuck were my own version of the major system for phone numbers, and a sort of memory palace thing for poetry. The former is easy and has worked well. The latter was a little more labored to start (for instance, the image for "The curfew tolls the knell of parting day" contained among other things a dog, some chocolate chips, a strumpet and the Dey of Algiers heading off in a huff) but worked even better and got much easier. Images that are vivid, funny, sense-driven or raunchy and are wholly unconnected with the meaning of the text being memorized work much better. Oddly, images that relate to the text make things harder. It sounds like a complete waste of effort, but it really makes remembering stuff easy and methodical. Kubla Khan, the last one I memorized probably took me an hour or two to lever into my skull and it will be there forever.

Anyway, I tried to think up something that would work for joseki and it didn't really gel. I have been meaning to get back to it. (Yall might be aware that I've been increasingly preoccupied lately.) I think this fellow's approach with the major system is a good one and will probably work for a hundred or two patterns ... 10,000 would be difficult not merely because it's a lot but because the fixed meanings of points mean you will start to run out of plot lines. There's only so many things mom and that knight can get up to. I may try it, but I will do my own version for a few reasons. One is that I changed the Major system around a bit for my phone number thing and I don't want to have two Major systems in my head. Another is, I think it works better if you do your own creative thing. Lastly, well ... mom and the knight. I think the fellow in the videos made a slight tactical error in putting his mom in for the 3-3 point. Bad idea to put family members in memory systems. Especially mom. True, family members are extraordinarily vivid, but so are sex acts and violence. Folks who put family members in their "person-action-object" systems, for instance, inevitably find themselves trying to memorize some six-digit number with an image of Danny Devito, say, doing something personally gallant with a much-loved aunt and then can never really forget it.

Notwithstanding Phil's and others' point about the quantity of variations, I think there is an argument for a memorization scheme for go patterns that has little or no relation to the meanings of moves, in particular for adults. I took up this game aged 40 and have had, perhaps, more time than other folk my age to devote to it. I've experienced some of the differences between how grown-ups learn this and how kids learn. Quite aside from kids just having all those extra years under their lee, kids can suck up patterns by the boatload and remember them, while adults forget them routinely. And kids can develop a facility for deeper and more precise reading that an adult mind not previously trained has a hard time matching. On the other hand, adults have a capacity for planning, reasoning, strategizing, deliberate repetition and methodical plugging away that kids often do not.

When you ask teachers why all the tsumego, they'll often answer because it makes your reading stronger. Press them on it, and they say it means you recognize the easy or clear patterns quickly and you don't then have to read them out laboriously; you just know and can spend your time on the novel or interesting features of the situation. Surely tsumego increase reading capacity, but they may also enable the perhaps fixed and feeble capacity of someone to read to be put to better use. Part of it is puzzle practice, part is rote, almost ritual, memorization.

In the case of joseki, the argument goes, the pattern memorization is harder and less useful than the reasoning that should go into picking a move. Memorizing a huge volume of joseki is a vast amount of labor and doesn't tell you want to do when play diverges from the text. Learn a few key ones, but mainly learn the reasoning behind moves and then you can make your own decisions about moves. A bunch of problems with this. One is that I and my fellow older folks have studied the meaning of lots of moves, some of them over and over again, and the moves just vanish, and with them, I suspect, much of the meaning. I'd give you an example, except I don't remember one. I study stuff and then, when I play, I am tempted by some suicidal doppelganger of a Taisha that involves me losing both sides and control of a quarter of the board, and I am reduced to trying to read out each variation and the meaning of each move, which does not play to my strengths at all. If I lament this, the answer is, "if you're not clear about it, you can always avoid the Taisha." Well that may be good advice for me, but clearly cannot be universal advice about learning go.

A second is that memorization, if only as a by-product, does seem to have taken place for others. Reviewing games with strong players and pros, they'll instantly come up with stuff about how this move's better than that one due to the ladder etc. They've remembered it; sometimes they analyze on the spot, but a lot of times, they just remember it, having seen it before. Another is that memorization is sometimes deliberate. Go teachers urge us not to memorize joseki; is it clear they don't urge memorization on eight year olds they think have the chance to be strong?

Lastly, and this is why this fellow's efforts intrigue me, I'm not sure we know how vast the amount of labor is for the motivated adult. Simply remembering joseki isn't going to happen easily, and doing it brute force without a method would be huge. But for me it was the same with poetry. Brute force didn't work. I tried 10 times to memorize Gray's Elegy and every time, the next day, a cow from the "lowing herd" in line two would moo at me and I'd forget it all. Then I applied a method and it was easy. And that kind of method plays to the strengths of an adult, like a recipe or an instruction manual or a good map. As to the objection that not working on the meaning means you'll know a lot of useless patterns, I'm not sure that's the case. Certainly with poetry, the palace fades with repetition. When I recite Gray to myself now, there is no rotting ear of corn, no genial flatfish, no Hermione Granger smirking. Now I can contemplate in great detail the poem because I have it at hand all the time. In go, I can learn from the colossal failure of a memorized joseki after playing it at the wrong time or place a couple of times in a tournament game. I can learn a variation or two of it that might solve the problem. I can learn "never play that one." I can't learn any of those things, though, if I can't play the joseki in practice. Lacking the extra years and the brain plasticity, what I can apply to the problem is a method.

Might be worth a try.

Agree with quantumf, it doesn't seem hijacked, although it has more opinions against the technique. I don't see myself as using it sucessfully,but it's interesting to read others that have.

_________________
a1h1 [1d]: You just need to curse the gods and defend.
Good Go = Shape.
Associação Portuguesa de Go