I've posted it a few times buried in other threads, but here's a new thread dedicated to a set of principles I call the "Leela Zero opening gospel" which define an order of priority of early opening moves that I've interpreted from reviewing lots of openings with Leela Zero. Edit: as a video https://youtu.be/vupa_IM1wWY

In general, play the following classes of opening moves in order from top to bottom. There will of course be exceptions, notably when there is an urgent situation involving weak groups or when there is some local sequence with high temperature happening, but once things quieten down these principles will help you find the next biggest move. The negative numbers are common examples of these urgent local situations.

-3) if your 3-4 is being pressed, answer it

Click Here To Show Diagram Code

[go]$$W
$$ . . . . . . . |
$$ . . . . . . . |
$$ . . . . O . . |
$$ . . 1 , . . . |
$$ . . 2 X . . . |
$$ . . . . . . . |
$$ . . . . . . . |
$$ --------------+[/go]

-2) if your 4-4 is 3-3 invaded, answer it

Click Here To Show Diagram Code

[go]$$W
$$ . . . . . . |
$$ . . . . . . |
$$ . . X . . . |
$$ . . 2 1 . . |
$$ . . . . . . |
$$ . . . . . . |
$$ ------------+[/go]

-1) if your 4-4 is double approached, attach out

Click Here To Show Diagram Code

[go]$$W
$$ . . . . . . . |
$$ . . . 4 1 . . |
$$ . . . . . . . |
$$ . . . X . . . |
$$ . 3 . . . . . |
$$ . . . . . . . |
$$ . . . . . . . |
$$ --------------+[/go]

1) play in empty corner (4-4 or 3-4)

Click Here To Show Diagram Code

[go]$$B
$$ . . . . . . |
$$ . . . . . . |
$$ . . 1 . . . |
$$ . . . . . . |
$$ . . . . . . |
$$ . . . . . . |
$$ ------------+[/go]

2) approach (usually low) or enclose (big high or big low) a 3-4

Click Here To Show Diagram Code

[go]$$B
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O , . . . . . , . . . 1 . , X . . |
$$ | . . . . 1 . . . . . . . . 1 . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

3) press if they ignored your approach to 3-4

Click Here To Show Diagram Code

[go]$$B
$$ . . . . . . . |
$$ . . . . . . . |
$$ . . . . X . . |
$$ . . 1 , . . . |
$$ . . . O . . . |
$$ . . . . . . . |
$$ . . . . . . . |
$$ --------------+[/go]

3.5) kosumi/knight answer approach to your 3-4, i.e prevent press #3

Click Here To Show Diagram Code

[go]$$W
$$ . . . . . . . |
$$ . . . . . . . |
$$ . . . . X . . |
$$ . 1 1 , . . . |
$$ . . . O . . . |
$$ . . . . . . . |
$$ . . . . . . . |
$$ --------------+[/go]

4) 3-3 invade or approach a 4-4

Click Here To Show Diagram Code

[go]$$B
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . O . . . |
$$ | . . 1 . . . . . . . . . . 1 . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

5) double approach a 4-4

Click Here To Show Diagram Code

[go]$$B
$$ . . . . . . . |
$$ . . . . 1 . . |
$$ . . . . . . . |
$$ . . . O . . . |
$$ . X . . . . . |
$$ . . . . . . . |
$$ . . . . . . . |
$$ --------------+[/go]

5.5) answer approach to 4-4, i.e prevent double approach #5

Click Here To Show Diagram Code

[go]$$W
$$ . . . . . . . |
$$ . . . . 1 . . |
$$ . . . . . . . |
$$ . . . O . . . |
$$ . X . . . . . |
$$ . . . . . . . |
$$ . . . . . . . |
$$ --------------+[/go]

6) enclose from 4-4

Click Here To Show Diagram Code

[go]$$B
$$ . . . . . . . |
$$ . . . . . . . |
$$ . . . X . . . |
$$ . 1 . . . . . |
$$ . . . . . . . |
$$ . . . . . . . |
$$ --------------+[/go]

Note the lack of moves on the side, "corners before sides" as the old proverb says.

Application 1: White denies the Kobayashi opening

Click Here To Show Diagram Code

[go]$$B
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 2 . . . . . , . . . . . 1 . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 7 . . . . . . . . . . . . . . . . |
$$ | . . . . 9 . . . . . . . . . . . 6 . . |
$$ | . . . 4 . . . . . , . . . . 8 , . . . |
$$ | . . . . . 5 . . . . . . . . 0 3 . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

Up to move 4 both sides follow the gospel, but 5 departs, with a class 4 move (approach 4-4) when there is a class 2 (shimari from 3-4) available. Also 3-3 invade a 4-4 is usually a little higher class 4 than approach (this is a pretty minor mistake). With move 6 white follows the gospel, approach 3-4 is class 2, above answer 4-4 approach class 5.5. Black displays fighting spirit "Hey you didn't answer my approach, so I will double approach like we were told was good before AI!" with double approach to 4-4 (class 5), but this is again a mistake as playing in the lower right to prevent press is class 3.5. White confidently follows the gospel with press at 8, a class 3. Black continues his kiai-driven run of mistakes with 9 surrounding the 4-4, but white's surround of 10 is bigger, that this trade is nearly always good for white is the reason pressing a 3-4 is higher priority than double approaching a 4-4 (the 4-4 is a lot less dead than the 3-4, and can make a roomy life later in the opening).

I had already dug it up here: https://senseis.xmp.net/?LeelaZerosOpeningGospel

Can I expand it as well?

Application 1a: Denied Kobayashi with less foolish kiai, black accepts gote from press

Click Here To Show Diagram Code

[go]$$Wcm8
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X a . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . O . . |
$$ | . . . O . 5 . . . , . . . 3 1 , . . . |
$$ | . . . . . X . . . . . . 4 . 2 X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

This is a variation of the previous Kobayashi denied opening, where after the press of white 8 black doesn't succumb to the foolish "if you ignored my move I will follow it up out of fighting spirit even though your move was bigger" idea, and answers the press (class -3). White extends at 10, not on the list but an example of high temperature local sequences that should be resolved before returning to big places from the gospel. The jump out to 11 is similar. The lower right is now finished, white happy to press black down in sente (no extension on side needed until all gospel level moves are done), and returns to a class -1 attach out in response to the double approached 4-4. Which attach, 12 or a doesn't really matter, it is rare for one to more than a percent or two worse than the other so unless 1) you are a high dan or 2) you like pondering the subtleties of the opening leaving not enough time for more important middle-game fights (guilty as charged!) then just pick one and move on.

Application 1b: Kobayashi denied, black wants sente from the press

Click Here To Show Diagram Code

[go]$$Wcm8
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X . . . . . . . . . . . . . . . . |
$$ | . . . . 6 . . . . . . . . . . . O 9 . |
$$ | . . . O . . . . . , . . 5 3 1 , . 8 . |
$$ | . . . . . X . . . . . . 7 4 2 X . a . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

In this variation, black wants to get sente from the lower right to surround the 4-4, so pushes an extra time with 11. White still answers with the normal extension of 12, this is an urgent local situation so takes precedence over e.g. class -1 attach out to save the 4-4. Having made his 3-4 not killable in sente, black takes the 5-5, and white turns, sharing the 2 urgent areas. Black's knight move of 15 is urgent to help his weak group (a has also been played by pros, if tenuki then white s3 doesn't kill but is profitable bullying), and white block at 16 is a move which I originally thought might be too slow but AI says is good even though black can tenuki and not die (the gospel can't help with this judgement). There's some nice life-and-death tesujis in this corner for study. Iyama Yuta and Yamashita Keigo played a similar opening to this in the 2018 Honinbo title match, but the double approach was one low one high: http://www.go4go.net/go/games/sgfview/74807.

@Knotwilg: sure

Many thanks, Uberdude!

BTW, I love the Kobayashi Denied Fuseki. Great name.

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

Visualize whirled peas.

Everything with love. Stay safe.

Only Issue I have: why is class -1 classified as -1 when your example clearly shows that white 8 is good move, i.e. in that situation class 3 is preferrable to class -1 (I understand the mutual damage idea, but the fact that 3 threatens big damage and is therefore kinda sente should also be taken into account into the ranking of the move, shouldnt it?)

Application 2: Why preventing your opponent's good move is slightly lower priority than them doing it, a case for kiai

This is an example where kiai or a rush of blood to the head/balls does lead to good play. The addition of the 3.5) answer approach to 3-4 (prevent press) and 5.5) answer approach to 4-4 (prevent double approach) were in response to someone on reddit asking where they would fit in. I put them just below the class of the good thing they prevent your opponent doing, rather than at the same level. That's because if you treat them as the same, your opponent is making small gains in sente, in that they lessen how good your next move in that area is, whilst retaining sente. An example:

Click Here To Show Diagram Code

[go]$$Bc
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 2 . . . . . . . . . 9 . . . . . |
$$ | . . . , 8 . . . . , . . . . . , 1 . . |
$$ | . . 7 . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 0 . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 5 . . . . . . . . . . . . . . . . |
$$ | . . . , . 6 . . . , . . . . . , . . . |
$$ | . . . 4 . . . . . . . . . . . 3 . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

Moves 1 to 4 take the empty corners at 3-4, all following the gospel at class 1. The reason I picked 3-4s for this example is, as old opening theory also taught us, they are 'unfinished corners' so the next play there of an approach/enclosure is class 2, above a class 4 approach/invade of a 4-4, the board has 4 hot areas. Black 5 follows the gospel with one of several class 2 choices available. If white answers with the patient 6 (class 3.5) then black can grab another class 2 approach at 7, white answers patiently again, and then black 9 grabs another class 2, a shimari this time (top right is slightly better than lower right, due to preventing white's class 2 approach there which works well with 8 to develop top side, this direction of play decision the gospel can't help you with, but if you pick lower right shimari it's not much worse). White 10 didn't follow the gospel by approaching at r5, but is also a good move and illustrates the point this move results in a position, with tally 3-1 to white, that is not quite as good as the same net surplus of 2 white moves of 2-0 for a shimari if black didn't approach there but directly played 7:

Click Here To Show Diagram Code

[go]$$Bc
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 2 . . . . . . . . . 7 . . . . . |
$$ | . . . , . . . . . , . . . . . , 1 . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 8 . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 5 . . . . . . . . . . . . . . . . |
$$ | . . . , . 6 . . . , . . . . . , . . . |
$$ | . . . 4 . . . . . . . . . . . 3 . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

So if white followed the gospel for 6 there are a few class 2 choices, such as taking the one shimari he can, as go Seigen liked to do on this left side facing 3-4s position, or approaches at a or b. To avoid letting black making small gains in sente, white takes an equally valuable class 2 move and we can end up with the kiai-fest of lots or approaches all around the board.

Click Here To Show Diagram Code

[go]$$Bc
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . . . . . . a . . . . |
$$ | . . . , . . . . . , . . . . . , X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 6 . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 5 . . . . . . . . . . . . . b . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . O . . . . . . . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

One caveat to the above argument is the gospel is colour-agnostic, but in reality white has komi so a more patient style of some class 3.5s before class 2s is not so bad if it gives you the kind of game you can control better.

That's a good point. Originally I had just 1-6, then I added -1 to -3 as an attempt to reflect the "hot local situation interlude" and then added the x.5s. It shouldn't come as a surprise that my attempted simplification of playing a superhuman level opening into a single ordered list ran into some trouble I will have a think how to reconcile this.

When I first formulated this list, my feeling was that pressing a 3-4 following an ignored approach was about the same as approaching / shimari (and when I play trying to be bot style, I often press before playing an approach elsewhere when I get the chance), but playing around with various positions tended to find LZ preferred the approach over the press so demoted the press to class 3. However, it is pretty slim, and I will check a wider variety of positions to see if class 2 and 3 really ought to be merged. But that'd still be after -1 so still needs further work. I suppose one thing with the press is it becomes extra effective when it interacts with an adjacent corner in leading the opponent into overconcentration, like in this Kobayashi denied example.

Probably I am missing something. Do you mean that LZ picks as the top choice? If not, what is the point? The double approach on the left has never been the most popular reply to and has had quite poor results in human play. I am guessing that the diagonal play at P4 would be "gospel". It did not really exist in human play up to 2016 but is the most popular reply since then. That said, the reply with the best results in human play has been the one-space high pincer at Q7. It also happens to be blue for the kagago 40-block (125K and slowly counting ). Where does that fit into the gospel?

As an acolyte of the divine Katago , I can't check on the formerly great but now false leela myself, you understand. But if we are going to speak of gospels, I am all in!

_________________
Dave Sigaty
"Short-lived are both the praiser and the praised, and rememberer and the remembered..."
- Marcus Aurelius; Meditations, VIII 21

BTW, if we are going to talk "Gospel" (I like the feeling of this idea), where should we get it from? Will it be the Gospel according to So-and-so?

Meanwhile, I think that any Gospel should at least explain, or be in line with, the bot's policy. Here is the katago 40-block policy for the Kobayashi denied example.

_________________
Dave Sigaty
"Short-lived are both the praiser and the praised, and rememberer and the remembered..."
- Marcus Aurelius; Meditations, VIII 21

Sorry, I should have been more explicit about the context/purpose of these examples. They show how one player, white here, follows the gospel, while the other doesn't but plays typical human moves and falls behind. When I play the Kobayashi denied approach of 6 against amateur dans, the kiai-driven double approach is probably the most common reply. It was also the move chosen by world-champion 9ps Chang Hao and Tang Weixing when AlphaGo Master played 6 against them in the online series. You are right that in recent professional games post-AlphaGo pros have learnt that answering locally (either kosumi defence or pincer) is better, but there are still examples of top pros like Shi Yue, Cho Chikun, Cho Hyeyeon double approaching in 2017 and 2018. All 3 continuations I posted (full kiai trade, black accepting gote from press with push and jump, black taking sente with extra push and trade) I've encountered in my games and show how, by following the gospel, I take an opening lead (and you can too).

The Shahada of this thread could be "LeelaZero is the one God and Uberdude his his messenger"

As you say, there is a powerful new God on the block, with tempting promises of points and strong handicap play. I am not averse to a conversion, but have not studied with KataGo enough to be the messenger of its gospel (my experience is also with what are now fairly old versions of LeelaZero). I think quite a lot will be the same, but it would be interesting if there are significant differences.

I've documented the illustrated gospel here:

https://senseis.xmp.net/?LeelaZerosOpeningGospel

I still need to digest the Kobayashi discussion, but it inspired me to review two traditional openings with AI

https://senseis.xmp.net/?OrthodoxFuseki%2FreviewWithAI
https://senseis.xmp.net/?SanrenseiFuseki%2FReviewWithAI

All welcome to join in.

In the meantime, the index page

https://senseis.xmp.net/?OpeningsReviewedWithAI

has 6 openings reviewed with AI (LeelaZero and KataGo)

Knotwilg, thanks for your work on SL.

I made a video on the LZ opening gospel: https://youtu.be/vupa_IM1wWY

"My gospel is holier than your gospel", said BlackHole to KataGo: http://www.yss-aya.com/cgos/viewer.cgi? ... 702940.sgf

Interestingly, the several KataGo versions on CGOS all have very negative scores against the BlackHole bot: http://www.yss-aya.com/cgos/19x19/cross ... 5v10k.html

_________________
Sorin - 361points.com

Unless we know black hole is not running on superior hardware / time limits to its opponents, that doesn't say too much: as we recently saw elsewhere KataGo at 200k playouts can beat KataGo at 5k playouts with 3 handicaps.

Uberdude,

I appreciate your work on the Leela Zero Opening Gospel … even I, being a DDK (fell back to 11k again), can learn from your LZOG, thank you!

And now that KaTrain exists (which made it easy even for me to install (including KataGo) on my Mac), can we hope for a “KZOG” from you also?

_________________
“The only difference between me and a madman is that I’m not mad.” — Salvador Dali ★ Play a slooooow correspondence game with me on OGS?

Out of curiosity, where's the following move:

Click Here To Show Diagram Code

[go]$$W
$$ . . . . . . . |
$$ . . . a a . . |
$$ . . . . 1 . . |
$$ . . . , . . . |
$$ . . . O . . . |
$$ . . . . . . . |
$$ . . . . . . . |
$$ --------------+[/go]

Obviously I guess the argument is that it doesn't often arise because approaching it is so high priority. I assume it's effectively 2.5?

EDIT: Ignore this, I've just realised the reason I couldn't find it is because White plays at "a" here, and you'd included it in the 2) diagram.

This is interesting to me - with the other situations (preventing the press, preventing the double approach), playing an additional move to one's own stone was considered slightly lower priority in general than pressing or double approaching your opponent's equivalent shape, and an additional stone to support your solitary 4-4 was the end of the list. Is there a specific reason you've identified that "an additional move to our 3-4" is "as good as approaching the opponent's 3-4", treating it as slightly bigger/more urgent than the other analogous situations?

Not sure that it's saying much, but if I run the AlphaGo teach tool on this position:

Click Here To Show Diagram Code

[go]$$Bc
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . O . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

I get this order of moves:

Click Here To Show Diagram Code

[go]$$Bc
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . . . . . . 2 . . . . |
$$ | . . . , . . . . . , . . . . . , X . . |
$$ | . . 3 . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . 1 . . |
$$ | . . X , . . . . . , . . . . a , . . . |
$$ | . . . . 4 . . . . . . . . . . O . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

On KataGo, I get slightly different results (e.g. press at 'a' instead of making a new approach). But I never see self-enclosure happen before approach. Maybe it has to do with the initial board setup - not sure.

My hunch is that approaching opponent is slightly better than enclosing yourself.

_________________
be immersed