Re: the best way to study
Posted: Tue Jul 26, 2011 8:08 pm
Take a tip from one who has already tried the "do hundreds" of problems route...and not got beyond 1k so far.
I agree with Numsgil's comment just above, and what's more, there are good reasons why.
I've recently been reading a lot of cognitive psychology, and it was interested to learn about the way memory works. If you're studying go, or anything else, for hours on end, it's likely you're overloading your working memory and, on top of that, depleting vital chemicals in your brain that make learning possible (serotonin, dopamine and others). Judy Willis suggests taking "syn-naps" to allow the brain to refill its supply of these neurotransmitters.
Have you ever studied for a long time and then asked yourself exactly what you have learned, only to find you're not really sure? I have
You need resting time to allow memories to transfer from working memory to long term memory. Further, you need time to consolidate and make sense of new information - this process is called "relational memory", if I have understood correctly. That's why you dream, and that's why often the answer to a puzzle often comes easily after you have taken a break. See Walker and Ellenbogen (2007) for more.
It seems better to study little and often, taking frequent breaks to allow information to be transferred to the LTM and to allow the brain to refresh itself (apparently, neurotransmitters can be replenished quickly if they are not worn down, but only slowly once you have become exhausted). And when you learn a lot, you cannot expect the benefit to be felt immediately - there is no substitute for the time that enables RM to perform its magic.
I am studying three things now: guitar, Japanese language, and go.
With guitar, I noticed significant improvements after I replaced long practice sessions with short, ten-minute ones focussed on one skill (e.g., alternate picking or tapping) or one section of one song (e.g., the first bar of the solo in "Purple Haze").
With Japanese, I find that by learning one new grammar or word at a time, I can learn and apply in real life more easily.
With go, I have only been playing for the last month, after a two-year lay off. I started at 4k (briefly 5k) on KGS, and am back to 1k there. I feel as though the bulk of the shallowly learned junk has cleared from my mind, and that what remains are the few bits and pieces of good advice and shape that I managed to pick up in spite of alternately cramming and skimming.
Further, there is no benefit eventually, in going back over things again and again. Rohrer and Pashler (2007) showed that "overlearning" is no better, over the long term, than reviewing things from time to time. In other words, there is no point in continuing to study your book of go problems once you have learned how to read them all - you'd be better off getting a new book, with more difficult problems in it. After all, who among us continues to recite the alphabet or practices tying their shoelaces? Getting better at anything means learning new things, not endlessly grinding through the old.
It's not psychology, but I think it's relevant: in the late 19th century Frederick Taylor demonstrated that a man could carry a great deal more pig iron in an hour if he rested between actions. If I recall correctly, his subject actually rested 33 minutes in the hour, and worked for 27 minutes in total. In other words, his study showed that one who rests more than they work can actually accomplish more work than one who works flat out. Could it also be that a similar approach to studying go, in which you work hard and rest frequently, might be the best?
I agree with Numsgil's comment just above, and what's more, there are good reasons why.
I've recently been reading a lot of cognitive psychology, and it was interested to learn about the way memory works. If you're studying go, or anything else, for hours on end, it's likely you're overloading your working memory and, on top of that, depleting vital chemicals in your brain that make learning possible (serotonin, dopamine and others). Judy Willis suggests taking "syn-naps" to allow the brain to refill its supply of these neurotransmitters.
Have you ever studied for a long time and then asked yourself exactly what you have learned, only to find you're not really sure? I have
You need resting time to allow memories to transfer from working memory to long term memory. Further, you need time to consolidate and make sense of new information - this process is called "relational memory", if I have understood correctly. That's why you dream, and that's why often the answer to a puzzle often comes easily after you have taken a break. See Walker and Ellenbogen (2007) for more.
It seems better to study little and often, taking frequent breaks to allow information to be transferred to the LTM and to allow the brain to refresh itself (apparently, neurotransmitters can be replenished quickly if they are not worn down, but only slowly once you have become exhausted). And when you learn a lot, you cannot expect the benefit to be felt immediately - there is no substitute for the time that enables RM to perform its magic.
I am studying three things now: guitar, Japanese language, and go.
With guitar, I noticed significant improvements after I replaced long practice sessions with short, ten-minute ones focussed on one skill (e.g., alternate picking or tapping) or one section of one song (e.g., the first bar of the solo in "Purple Haze").
With Japanese, I find that by learning one new grammar or word at a time, I can learn and apply in real life more easily.
With go, I have only been playing for the last month, after a two-year lay off. I started at 4k (briefly 5k) on KGS, and am back to 1k there. I feel as though the bulk of the shallowly learned junk has cleared from my mind, and that what remains are the few bits and pieces of good advice and shape that I managed to pick up in spite of alternately cramming and skimming.
Further, there is no benefit eventually, in going back over things again and again. Rohrer and Pashler (2007) showed that "overlearning" is no better, over the long term, than reviewing things from time to time. In other words, there is no point in continuing to study your book of go problems once you have learned how to read them all - you'd be better off getting a new book, with more difficult problems in it. After all, who among us continues to recite the alphabet or practices tying their shoelaces? Getting better at anything means learning new things, not endlessly grinding through the old.
It's not psychology, but I think it's relevant: in the late 19th century Frederick Taylor demonstrated that a man could carry a great deal more pig iron in an hour if he rested between actions. If I recall correctly, his subject actually rested 33 minutes in the hour, and worked for 27 minutes in total. In other words, his study showed that one who rests more than they work can actually accomplish more work than one who works flat out. Could it also be that a similar approach to studying go, in which you work hard and rest frequently, might be the best?