Pseudo sente, superficial shape, reading and other mistakes

Knotwilg · #1

I've recently lost 5 games in a row against 2-3k on OGS and I was unsettled by it but then read that OGS deflated their ranks. Nevertheless I have reviewed my games for mistakes and have learned a lot. My mistakes fall into these categories:

1. Pseudo sente

I play moves which intend to be sente but aren't. Sente is a big concept in what we try to learn from AI. It will often discard moves because they end in gote and there are still large moves on the board, so either it will recommend a variation that keeps sente or the big move in the first place.

2. Superficial shape

I remember from comments on a page I once compiled on SL that I treated shape too much as a static visual thing and not enough as a dynamic thing. At the time I didn't know if that was true nor how to correct that but thanks to AI reviews I'm now convinced of this bad tendency. Underneath shape lies the quest for efficiency. Good shape reaches a certain goal with a minimum amount of stones, or multiple goals with a certain amount of stones. That may result in shapes like the table shape. However, there are situations where a table shape can be made and looks like the right thing to do but it's only superficially good shape and actually inefficient.

3. Bad reading

A smaller number of mistakes were simply bad reading, where I could land a devastating blow by capturing or killing but simply missed it.

4. Other

Sometimes the move Katago suggested was something I couldn't fathom. The move I chose was much worse but not in a way I would have been able to assess in a game or independent review. They were not mistakes at my level of comprehension, rather they represent an opportunity to learn new moves.

I will post examples as exercises.

Knotwilg · #2

Example 1

White to play

Click Here To Show Diagram Code: [go]$$c Position at move 33 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . O O X . . . . . . . X O . | $$ | . X O O . . O X . . X . . . . X O . . | $$ | . . X , O . X X . , . . . X . , O . . | $$ | . X . . . . . . . . . . . . . O . . . | $$ | . . . . O . . . . . . . . . . . . . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . X . . . . . . . . . . . . X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X X . . . . . . . . . . . . O . . | $$ | . . O . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . O . . . | $$ | . . . . . O . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

My mistake:

What KataGo recommends instead:

Kirby · #3

Knotwilg wrote:

What KataGo recommends instead:

ez4u · #4

Knotwilg wrote:

Example 1

White to play

Click Here To Show Diagram Code: [go]$$ Position at move 33 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . O O X . . . . . . . X O . | $$ | . X O O . . O X . . X . . . . X O . . | $$ | . . X , O . X X . , . . . X . , O . . | $$ | . X . . . . . . . . . . . . . O . . . | $$ | . . . . O . . . . . . . . . . . . . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . X . . . . . . . . . . . . X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X X . . . . . . . . . . . . O . . | $$ | . . O . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . O . . . | $$ | . . . . . O . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

My mistake:

Why does KataGo think this is a major mistake?

What KataGo recommends instead...

Note that "What Katago recommends instead" does not answer the question "Why does KataGo think this is a major mistake?"

Knotwilg · #5

Indeed. I left it for exercise

Uberdude · #6

My eye is drawn more to the top left and the lack of a White stone at f16 so the troubles black f17 will bring in reducing eyespace and activating E15 wedge, than top right. Top right has miai for safety: if black connects q18 White is not unhappy to answer at r14 or s13 as these moves have similar territorial value and also weaken Black's 2 space extension group on right. So I'm thinking either strengthen the top left or take a big point on lower side like l4 or o3. Pushing at e15 makes sense because it's unrealistic to think white d11 invasion will be a productive move with the weak group above, so better to give black the left side and try to ovetconcentate and get b6 too.

Knotwilg · #7

Thanks for adding coordinates Uberdude.

Here's the missing part of the analysis:

Classification of my mistake:

Knotwilg · #8

From the same game, some 20 moves later. How should White deal with the crosscut?

Click Here To Show Diagram Code: [go]$$Wc $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . O O X . . . . . O O X O . | $$ | . X O O . . O X . . X . . X . X O . . | $$ | . . X , O . X X . , . . . X . , O . . | $$ | . X . . . . . . . . . . . . . O . . . | $$ | . . X O O O . . . . . . . . . . . . . | $$ | . . . X X O . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . X . . | $$ | . . . . X . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . X . . . . . . . . . . . . X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . X . . . . . . . . . . . . . . . . . | $$ | . O X X . . . . . . . . . . . . O . . | $$ | . O O . . . . . . . . . . . . . . . . | $$ | . . . O . X O . . , . . . . . O . . . | $$ | . . . . . O X . O . X . X . . O . . . | $$ | . . . . . . . . . . . . . . X X O . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

My mistake

What I should have done:

Classification of the mistake:

John Fairbairn · #9

Quote:

1. Pseudo sente

I play moves which intend to be sente but aren't. Sente is a big concept in what we try to learn from AI. It will often discard moves because they end in gote and there are still large moves on the board, so either it will recommend a variation that keeps sente or the big move in the first place.

We have nothing to learn about sente from AI. We can learn more from squirrels.

In AI play sente is a trivial concept and occurs in ways that we all understand perfectly well already: In this locality, I play A, he has to play B.

The fact that you invent the new term pseudo-sente shows you are well aware there is a problem there. The problem is that you are using the term sente in the first place. The term you are looking for is the 'initiative'. That is the term AI is demonstrating. But you already know all about the initiative. Take a 5-stone game, move 1. White has sente (first move), of course. But who has the initiative?

In fact, in any game, the player who is ahead is the one who has the initiative. Since knowing who has the initiative is a matter of knowing who is ahead (i.e. who has the better prospects, not the most current territory), it is a matter of evaluating the whole game. That's where AI prevails.

So, sente is local but initiative is global. Does that help? It can.

There is no initiative vaccine yet, and not likely to be one for a long time to come. But recognising that the problem is one of seeing the big picture while working with what we have is a good start. Think social distancing and keep away from bubbles: make local plays only when it are essential. Avoid crowded places: keep groups two yards apart, especially from the strangers' groups. Identify your key workers and protect them. Apply shielding to vulnerable groups. Wear a mask: it may not actually work but if it helps you from touching your face - playing local sente moves - the spin-off is good.

It's pretty obvious that training yourself to concentrate on the big picture is important, yes? But in real life our response to the lurgy is complicated by the need to earn more money so we can have three holidays a year and four cars per family. In amateur go, things are complicated by the need to make a huge moyo, the need to invade, the need to show the opponent who is boss. Yet AI simplifies things: the only need is to win the game and 0.5 point will do.

Assuming, therefore, we can learn from that and accept a lockdown, and play sensibly, we are still faced with an open-ended question: "How do I know who is ahead? I haven't got time to count the whole board every single move." This is where pro techniques such as QARTS or counting only the mistakes or bad shapes rather than territory are so valuable. Try pot noodles instead of making your spaghetti the real Italian way. Just seeing the whole picture, even if the edges are blurred, instead of looking down at your feet all the time, pays off in the long run.

I predict that if you can treat every play as a chance to wash your hands with soap and not just water, and sing Happy Birthday for 20 seconds, giving you a chance to think about the big picture, AI's ability and proclivity to skip around the board will suddenly not seem so mystical.

In go, the initiative is your soap. Sente is mere water.

How did we get into this mess? People are alike the world over, of course, and amateur go players the world over tend to make the same mistakes. But the sente virus started in Asia and caught us off guard. We didn't have our own word for that and it was euphonic. It spread like wildfire. It trampled all over poor little initiative - harder to say and much harder to spell....

Good intentions in importing novel species are not enough. Look at cane frogs in Australia. Or here in the UK, cute little red squirrels have been almost exterminated by the rat-like grey squirrels that hop along my garden fence every day.

There is hope, though. The lockdown here has kept humans from country spaces and has allowed threatened flora and fauna to revive. I've seen some stunning butterflies recently, and plants I hadn't seen since I was a kid.

So, in short, my advice is to train yourself to look at the big picture and to treat sente as an imported feral species and initiative as the beautiful native butterfly of your youth.

Uberdude · **#10**

"Extend from a crosscut" is a good proverb for DDKs taking 9 stones from a madly overplaying White player, but once you get stronger my version of this proverb is

Quote:

When crosscut is a bad move, extend.
When crosscut is a good move, atari.

I've not actually gathered pattern search statistics, but my feeling is in pro games atari is played more than 50% of the time after a crosscut, because pros aren't playing the bad crosscuts you trivially punish by extending. The atari often serves to make one of the cutting stones heavier and/or prevent a ladder. BadukDoctor had a nice lecture on such an example.

Kirby · **#11**

Extension was my first instinct, too. I didn’t think about sente at all!

John Fairbairn · **#12**

Quote:

When crosscut is a bad move, extend.
When crosscut is a good move, atari.

It's rather odd how the proverb "From the crosscut extend" got such distorted traction, especially here in the west. As far as I can recall it did not appear in Go Proverbs Illustrated, but was treated in some detail in Go Monthly but in the context of handicap play. There was also a very brief note about it in Go World 73.

It exists in Japanese, of course, but my impression is that whenever it comes up it is treated in qualified ways that are overlooked here.

For example, Hayashi Yutaka says that when it is used by a player to create confusion (a typical ploy in handicap play of course), extending is the simplest way to avoid that confusion. Hayashi further adds that the extension, though safe, is often a slack move in practice and "that point should be well borne in mind."

Kageyama gave it a long and scathing treatment in an Igo Club supplement. Right at the beginning he says "extending on one side is slack" (he notes that the extender often ends up "too solid" and/or "overconcentrated") and he too points out the different contexts that apply when a strong player uses a crosscut against a weak player. He ends his condemnation by arguing that the proverb should be recast as "separate the cross-cutting opponent", which of course means using atari.

So, the extra context is clearly important, and for those who have learned the proverb without that extra context, it's a classic case of a real-life proverb: "A little learning is a dangerous thing."

And you can have too much as well, of course. Kageyama's book also contains the following superb cartoon by the truly great go cartoonist Fujii Leo (that's his monicker in the lower left). The balloons represent various well-known proverbs, and the advice given is what to do to your overstuffed head when it's too in thrall to hackneyed proverbs. ♫ If I had a hammer...

Attachment:

IMG_1653.jpg [ 32.55 KiB | Viewed 10927 times ]

Gomoto · **#13**

Click Here To Show Diagram Code: [go]$$ $$ +---------------------------------------+ $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . X O . . . . . | $$ | . . . . . . . . . . . . O X . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . X O . . . . . . . . . . . . | $$ | . . . . . O X . O . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ +---------------------------------------+[/go]

Kirby · **#14**

Stepping back a bit in that last game, I’m guessing there was a sequence like this:

Click Here To Show Diagram Code: [go]$$Bc $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . O O X . . . . . O O X O . | $$ | . X O O . . O X . . X . . X . X O . . | $$ | . . X , O . X X . , . . . X . , O . . | $$ | . X . . . . . . . . . . . . . O . . . | $$ | . . X O O O . . . . . . . . . . . . . | $$ | . . . X X O . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . X . . | $$ | . . . . X . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . X . . . . . . . . . . . . X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . X . . . . . . . . . . . . . . . . . | $$ | . O X X . . . . . . . . . . . . O . . | $$ | . O O . . . . . . . . . . . . . . . . | $$ | . . . O . 3 2 . . , . . . . . O . . . | $$ | . . . . . O 1 . O . X . X . . O . . . | $$ | . . . . . . . . . . . . . . X X O . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

I think it’s a nice sequence/idea by black. I’m not confident I would have thought of it from this position:

Click Here To Show Diagram Code: [go]$$Bc $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . O O X . . . . . O O X O . | $$ | . X O O . . O X . . X . . X . X O . . | $$ | . . X , O . X X . , . . . X . , O . . | $$ | . X . . . . . . . . . . . . . O . . . | $$ | . . X O O O . . . . . . . . . . . . . | $$ | . . . X X O . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . X . . | $$ | . . . . X . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . X . . . . . . . . . . . . X . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . X . . . . . . . . . . . . . . . . . | $$ | . O X X . . . . . . . . . . . . O . . | $$ | . O O . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . O . . . | $$ | . . . . . O . . O . X . X . . O . . . | $$ | . . . . . . . . . . . . . . X X O . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Looking at it now, it feels like there’s something there. But would I have thought of that during a game? Probably not.

Knotwilg · **#15**

Mistake 3 - same game

Click Here To Show Diagram Code: [go]$$Wc Position at move 101 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . O O X . . . . X O O X O . | $$ | . X O O . . O X . . X . . X . X O . . | $$ | . . X , O . X X . , . . . X . , O . . | $$ | . X . . . . . . . . . . . . . O . . . | $$ | . . X O O O . . . . . . . . . . X O . | $$ | . . . X X O . O . X . . . . X . X O . | $$ | . . . . . . . . O . . . . . . . X . . | $$ | . X . . X . X . . . . . . . O . . . . | $$ | . . X , . . . . . , . . . . . , X . . | $$ | X X O X . . . . . . . . . . . O X . . | $$ | O O O X . . . . . . . . . . . O O X . | $$ | O X X X O . . . . . . . . . . . . X . | $$ | O O X X . . . . . . . . . . . . O O . | $$ | O O O X . O O . . . . . . . . . . . . | $$ | . . . O . X O X X , . O . . . O . . . | $$ | . . O . O O X X O . X X X . . O . . . | $$ | . . . . . . . . . . . . . . X X O O . | $$ | . . . . . . . . . . . . . . . . X . . | $$ ---------------------------------------[/go]

My mistake:

What KataGo recommended instead:

Farodin · **#16**

Knotwilg · **#17**

Farodin wrote:

What exactly is the purpose of all those exchanges before move 7? I have seen it happen a lot in pro games as well, but I find it hard to understand them. My guess for the 1-2 exchange is that a black follow-up at C18 is easier to deal with, but I don't exactly see why. Not to speak of the 3-6 exchange, which is entirely beyond me.

This is a question of a whole other level (pro/AI). I looked at what would happen if White plays 7 directly. The 1-2 exchange remains forcing but the 3-6 exchange would no longer occur in that case, as Black will deal with 3 differently. Analyzing that is beyond me. Let's say that my main mistake is playing my move instead of 7.

Uberdude · **#18**

1 for 2 is not so mysterious, it makes black only have a gote move followup in that area with c18, rather than the choice of b18 or c18. b18 might be preferred if its sente profit against the life of the white group. You sometimes see this hane in endgames to prevent your opponent making sente profit and reduce their option to gote, it can help with tedomari. I remember Vanessa Wong 5d ruing her failure to make a similar hane (on edge not corner) and thus losing in endgame against Osawa Maya (one of the top Japanese amateur women, sister Narumi is pro) in the women's semi-final at the World Mind Sports Games 2012 in Lille.

3-6 aren't a standard idea, but seem to be looking at the the j17 cut, and in the process of defending that white makes some options for eyes at g12.

Pseudo sente, superficial shape, reading and other mistakes

Who is online