Reading vs direction of play

[snorri]

Bruce Wilcox's EZGo concepts have allowed kyu players to advance up to 4 stones in a couple of weeks.

That's not an exaggeration, especially if one is starting at plodding bumpkin level. The more time goes by and advice I see, the more I think Wilcox was underestimated. Maybe because of some politics because he was not a pro.

When I say stuff like, "all of modern go theory is worth maybe 2 stones" I'm really thinking post-Wilcox EZGo level. It is as if I am taking it as a given that all players should have Wilcox's tools in their bag whether they learned them from Wilcox or some other equivalent way. Unfortunately, not all do.

Seriously, you can pick that stuff up in a day. It's well worth the time and money. Bill's two weeks is so you can get in the games to practice it...

The only drawback is that it's kind of a 1-time bump. That and things like getting ahead (which Guo Juan claims is worth 5 stones if it's new to you!) are great, but then the easy gains start to dwindle as one faces the spectre of limited reading...

[Kirby]

As your friend and you are both roughly the same strength, have you answered your own question?

[moyoaji]

As your friend and you are both roughly the same strength, have you answered your own question?

Why didn't I think of that?

[snorri]

Reading is the most important skill in the sense that if your reading is strong enough, you don't need other skills. (That's not my phrasing. It comes from Yuan Zhou, but I like it.) But no one's reading is perfect. Even computers can't exhaustively check all variations in anything but the very late endgame. That's why we have proverbs and joseki books and other lore. It's to help us make some kind of choices in situations where reading is not possible.

Most beginners can barely read at all. So for them, big benefits can be gained by using proverbs and understanding some rules of thumb. Direction of play is one of those things.

But as Bill pointed out, there aren't exactly a lot of 7-dans kicking around that struggle with 10-kyu tsumego. (But there are 10-kyus that can answer 7-dan opening problems.) Most of the gap between 10-kyu and pro is reading. But human reading is not like computer reading. We don't consider the following and then read out bazillions of variations to the end of the game:

$$W Read it out...?
$$ ---------------------------------------
$$ | . . . . . . . . 1 . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . X . . . . . , . . . |
$$ | . . . . . X . . . . . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------

Click Here To Show Diagram Code

[go]$$W Read it out...?
$$ ---------------------------------------
$$ | . . . . . . . . 1 . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . X . . . . . , . . . |
$$ | . . . . . X . . . . . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

You never know. This might be the best move.

So reading includes choosing better candidate moves, and so it assumes some knowledge of go theory. For that reason, it's possible to argue that reading is inseparable from knowledge, at least in humans. Studies in chess at least show that stronger players choose better candidate moves when reading. They prune the tree better, so much better that they often don't even have to go deeper and would rather spend time judging the end result of some options.

But they can go deeper if it's needed.

Brute-force reading is helpful or even necessary in some constrained situations. Ladders, enclosed life & death, some critical cuts and connections, for example. Because of the nature of go, it is often the case that mistakes in those situations are especially costly.

So when I say reading is more important, I can only say human reading, which is informed by good knowledge and infected by bad knowledge at the same time.

[NoSkill]

To me I think reading is more important, but fundamentals are the most important. Fundamentals like good shape, attacking and defending, solid play, and the direction of play together are most of go. You combine reading and fundAmentals to play. Fundamentals is what seperates humans from bots, we can effectively play good moves without reading, or reduce our reading to many times less branches. Imagine reading 1000 moves for each move, but having no clue about what lines to start with. Compare that to a pro who reads 100 moves.

[Toge]

How do you separate theory from reading within a person? Is someone who has no direction of play suffering from extreme form of tunnel vision and able to see only 4x4 grid at the time? Is someone who doesn't subscribe to "theory" reading out every variation of invasion against two point jump on third line before placing the second stone? All we know of a person's ability is how he fares in games against his peers. I don't think it's possible to dichotomize skill.

[SmoothOper]

Summing up the thread, the consensus is reading is more important than reading, but better reading is better, and deeper reading is not better reading. Did I miss anything?

[John Fairbairn]

Normally westerners put a static emphasis on concepts that are really dynamic in Japanese (e.g. katachi/suji). This is a rare arsey-versey case. The Japanese term really means 'direction of groups' and the idea is that groups have directionality, or favoured directions in simpler English.

The value of recognising the favoured directions is that it can direct your reading. Humans cannot cope easily with the bushy trees that erupt when you try to read moves out exhaustively. Humans therefore rely on heuristics to prune the trees. The 'direction of play' concept is especially valuable because it allows you to prune the tree at a level much nearer the root than most heuristics.

Actually I find that amateurs have little difficulty in grasping the concept or in applying it in a rudimentary way. Their main difficulty is rather in realising just how far directionality extends. E.g. they tend not to realise that a group on the 4th line can have has a direct and strong impact on a group on the 17th line. Groups can have a slow, creeping, but still unstoppable momentum. If the cross-board movement of a ladder can be likened to an avalanche, other groups tend to behave like a glacier. Spotting the glacier's movement is ultimately what direction of play is all about, and so in some ways 'depth of (direction of) play is the concept to latch onto.

[SmoothOper]

Normally westerners put a static emphasis on concepts that are really dynamic in Japanese (e.g. katachi/suji). This is a rare arsey-versey case. The Japanese term really means 'direction of groups' and the idea is that groups have directionality, or favoured directions in simpler English.

The value of recognising the favoured directions is that it can direct your reading. Humans cannot cope easily with the bushy trees that erupt when you try to read moves out exhaustively. Humans therefore rely on heuristics to prune the trees. The 'direction of play' concept is especially valuable because it allows you to prune the tree at a level much nearer the root than most heuristics.

Actually I find that amateurs have little difficulty in grasping the concept or in applying it in a rudimentary way. Their main difficulty is rather in realising just how far directionality extends. E.g. they tend not to realise that a group on the 4th line can have has a direct and strong impact on a group on the 17th line. Groups can have a slow, creeping, but still unstoppable momentum. If the cross-board movement of a ladder can be likened to an avalanche, other groups tend to behave like a glacier. Spotting the glacier's movement is ultimately what direction of play is all about, and so in some ways 'depth of (direction of) play is the concept to latch onto.

So you are saying the heuristics in "Direction of Play", allow you to read deeply(by pruning out variations), and there are other heuristics, that will also allow you to read deeply?

[Phoenix]

So you are saying the heuristics in "Direction of Play", allow you to read deeply(by pruning out variations), and there are other heuristics, that will also allow you to read deeply?

Of course there are. The (basic) idea here is that instead of reading out dozens of potential moves, the idea of 'direction of play' helps to limit ('prune', as John eloquently puts it) the number of moves to consider. It is much easier to read out four candidate moves than fourteen.

There are quite a few heuristics in Go that have been passed down which help in a similar way. The ideas of thickness, influence, moyo, aji, etc, can help us to think about where to start reading. Also, take proverbs such as 'don't play near thickness', 'there is death in the hane'... There are even proverbs and some shape knowledge which disposes of reading altogether. 'The bent-four in the corner is dead', 'play at the center of the bulky five'...

Speaking of shapes, a part of Go is about pattern recognition. Take tesuji exercises for example. You can read out a sequence, come to a point where you know a tesuji occurs, and it facilitates your reading in that regard. Or you can use tesuji knowledge to steer the flow of stones into positions that would allow you a tesuji. If your opponent relents, an advantage is often gained.

You've been learning all matter of heuristics to gain the upper hand, in many forms. These are just a few (rather scattered) examples.