How to Play the Sanrensei

[@marcelocarpinetti9184]

It is great to hear about the sanrensei again, which seems to be a "forgotten" opening nowadays.

[@Rubrickety]

The knight’s move after the wedge was very instructive. I think I might automatically make either the solid connection or the “triangle jump” which is a knight’s move from both black stones.

[@codyw9296]

I can always count on your channel for high quality content. This style of videos are very helpful to a kyu level player like me. Please keep these videos coming!

[@philgan2085]

Even if it's not recommended by AI I still regularly lose to the sanrensei opening. It's strong enough for humans :)

[@Wailn_]

This is such a great video, thanks Michael!

[@georgesmith827]

Excellent video! I just jumped 3 kyu levels just by watching this. 😀

[@vvlaunay]

Thank you ! As a 1-2D on KGS it was very instructive !

[@paulbloemen7256]

If one likes the sanrensei opening, I played it for many years, one probably plays the two 4-4 points as White too. So one may encounter the sanrensei opening from both sides regularly. I wonder what the best strategy for White is to counter it. And that may be useful information for Black too.

[@uw10isplaya]

Hi Michael. If you're looking for a video topic, I'd like to suggest the 3-4 low approach -> pincer -> knights move -> push and cut. This sequence pops up a lot in my games, and aside from the variation after the low pincer when you can just atari into tiger's mouth, I find the sequence too complicated for me. Whenever it comes up I take a peek at it with AI, but while AI is great for me in pointing out specific moves, I find the "logic" of an extended sequence difficult to follow. Managing the corner and the side feels very difficult without playing the exactly correct moves. Thanks for all your videos!

[@BenKyoBaduk]

When I check Sanrensei in KataGo with white playing two star points, it shows black's third move as -0.3 points, which I've been interpreting as totally playable, at least up to strong amateur levels, perhaps even pro level still. I would love to hear your thoughts on where different levels of pro might draw the line. My guess is that top pros would avoid this move, while other pros would play it if they wanted to, given that it's such a miniscule loss and much less than many of their moves will be losing in the middle-game anyway. In Katrain and likely other AI GUI programs it has the same default color label as the moves that score 0.0 or positive point values (top moves). I imagine that any 3rd move for black that's better than, say, -0.5 is still a good move for most Go players, the more important aspect being how comfortable and confident they are with that sort of move and the continuations it encourages. Thanks Michael-sensei!

[@Мопс_001]

I feel like trying to compare opening moves by their numbers in AI is a *very* bad form. I get it if you'd like to check a good move in the middle of the game or a critical point. But openings/frameworks are the absolutely worst stage to judge the position basing on AI's opinion. That puny - 0,1, even -1 or -2 doesn't tell a shit in the opening when players can make things extremely complicated and that point will be lost in the process. Professionals would describe the opening position like boring, simple, comfortable, precise, sloppy. And they would choose them based on this description where points aren't even considered. They can make points and win from any opening, just that in some it would be easier for them (even if the AI tells that this is, oh god, -0,5 points loss but the player knows the opening so well it won't matter)

[@BenKyoBaduk]

@@Мопс_001 I agree. I did try to work that into my message, that a lot of it is about how comfortable the player will handle the formation, how comfortable it makes them in the middle-game / rest of the game, I totally agree that AI score alone misses some really important factors, yup! Anyway, I was just curious if Michael had some insight into where most pros studying with AI draw the line in terms of point scores, because I think it would still be a factor they consider along with the other factor we talked about.

[@lagazettedesfrancais8155]

thank you.

[@marcelocarpinetti9184]

Dear Michael, I would really appreciate it if you could take a few minutes in subsequent videos to talk about how to deal with White's invasion from the white stone in S3 (monkey's jump, turn etc.). How much would Black need to worry about that with regard to the idea that this supposed potential territory might not all turn into actual profit. How much territory can Black actually expect from this area?

[@willikey]

Very interesting video, thank you!

[@dauraktv]

Really nice video

[@xdmamakkopitiam8930]

Yay sanrensei getting some love

[@anuzis]

Appreciate this great video! It would be helpful if future videos could also discuss how/when to handle W jump to S5 around 14:20. On the one hand, it feels inappropriate for B to block on the 2nd line prematurely at S4, on the other hand, allowing W to enter that space can be very difficult to contain, especially if B doesn't have a helping stone at P6 in. It's intuitive to try blocking W's S5 advance with a 1-space jump at S7, but that allows W to continue with a kosumi at R6. It's also intuitive to respond to W S5 with a 3rd line shoulder hit at R6, but that lets W continue to jump to S7 greatly neutralizing the value of the right 3 lines of the board and still being difficult to contain.

[@bastienp7334]

Thanks ! My fav opening with black is star and 3-4 5-4 enclosure but having the opponent playing the sanrensei when I'm white makes me feel quite uncomfortable, even when i manage to win in the end

[@SamuelGrguric]

we all learn from this then in our games w will just invade early like 90% of the time XD

[@Bladavia]

Probably any opening is fine as long as you're not playing Shin Jinseo himself, many openings that have fallen out of fashion haven't been clearly refuted, just forgotten.

[@Turtle1631991]

Case in point. We saw high chinese in EGF pro qualification finals just the other day.

[@decidrophob]

Composing a response made me reach the conclusion that sanrensei is probably not forgotten at all. Quite the contrary, it is too famous that black thinks white is too well-prepared. This probably applies to Shin Jinseo himself as well. As a top pro, you have to invest resources in some specific openings to study broad and deep, and hopefully edge opponents by the openings with which only you, but not the opponent, are familiar. I think Shin and other top pros consider that there is less chance that you can "lead" by choosing a too-well-known opening. Also, at the experimental stage, sanrensei is probably not forgotten at all even in terms of game records. It may thus be worth exploring what pros have been trying with respect to sanrensei on Fox or Tygem. I am pretty sure that many pros, if not the top most, have tried sanrensei many times even today, and probably they have not done too well. I imagine that sanrensei may be more difficult for the humans to play than Katago that only deducts 0.3 points, which is completely negligible.

[@burpsan]

Love it!

[@michalpolacek8227]

Super video for me (7kju). Thanks you.

[@alejandrofernandez4529]

I want more!!! ; )

[@decidrophob]

Thank you for the video as always. I could clearly sense that you have tried to show "speedy" play by black, the same idea that has existed since Takemiya's books. The choice of joseki "by black" on the right bottom was sente for black for instance. You showed the subsequent variation starting from around 10:00 allowing tenuki to the follow up play by white -- two sentes in a row! And indeed, it does look critical to take sente flexibly especially when we play direct 33 josekis. I am a bit puzzled by the narrative to characterize white strategy. For instance, if the shoulder hit at G16 is such a good move and white responds ordinarily not like the fancy AlphaGo tenuki style, then why doesn't the white apply the same narrative and hit the shoulder before black at N16? (My shallow understanding is that Katago seems to approve of shoulder hits much less than AlphaGo. Quite often when I try shoulder hits, Katago deducts something like 1.5 point or more, and also suggests patient responds like the ones in this video. It may be simply that my timing of shoulder hits are worst, but then I am curious what is bad about N16 in this video for instance.) Also, if I were white, I would be tempted to play something like G16 rather than F17 keima, if I had to play the left top at that timing, to counter the black moyo. (I personally do not want to allow black's shoulder hit as white!!!) Do you say that this sort of extension is too weak against 33 invasion by black, even when the global fuseki is sanrensei?

[@MichaelRedmondsGoTV]

My understanding of this opening is that the shoulder hit needs to be supported by a corner enclosure (White's move in the upper left) to be optimal.

[@decidrophob]

@@MichaelRedmondsGoTV Thank you very much for the response. What is your opinion about white's two space high jump (G16) from the star point? That is again a pretty modern shape stimulated by AIs, but do you say that this is only applicable in relation to one space pincer?

[@KilgoreTroutAsf]

After that intro I thought he would title the video "sanrensei is OK"

[@marcelocarpinetti9184]

Another question: is it ok to use the sanrensei in a "territorial" manner?

[@64standardtrickyness]

at 11:50 why is black 16 necessary? what if black cuts M4 directly?

[@MichaelRedmondsGoTV]

M4, M3, L4, K5. This only works with the context of all those forcing moves Black had in the displayed variation pushing along in the center. Black 16 at P6 was a necessary reinforcement to the right while waiting for White to move out in the center.

[@64standardtrickyness]

@@MichaelRedmondsGoTV What if I play 16 at M8 threatening O9 and M4?

[@vincentlui4505]

@@64standardtrickyness white playing 16 is annoying for black

[@JazzMessengerToGo]

There is an opening on the lower right where W can play S5. Why doesn't B play S4 to block it and how does B respond to W S5?

[@tante8074]

I swear I was literally ordering patterns of the senrensei this week and now this pops up, can you read my mind?

[@trevorjrooney]

It should be really interesting to see how close humans can come to AI. A good human opening, but AI do not favor it. Why? Which makes me wonfer about Lee Sedol's decision to quit thanks to AI.

[@C-shadow]

My instance of katago says 0.3 points lost to play sanrensei. That's nothing in a human game, and on one move it's also almost nothing in an AI game, but AI trains against AI, and when you're training to face an opponent that has near-perfect positional judgement and doesn't make mistakes, there's less value in inducing them to invade. It's riskier to make more groups as a human, so inducing your human opponent to do that is a great strategy. When there are four approaches, four enclosures, and two invasions to choose from, that are all 0.3 points better, the AI just doesn't have a reason to consider sanrensei as an option. Humans, on the other hand, might use strategy rather than raw intuition+reading, and also might do things just because they are interesting.

[@MichaelRedmondsGoTV]

0.3 is within the margin of error as I see it, in that the weights of programs will change as they are developed, and that could easily change to 0 or a positive number, or in reverse change to a larger negative.

[@brandonrogers7471]

What do you think about black enclosing at N16 instead of O17? I have noticed Takemiya play this enclosure in some of his sanrensei games so I am curious about it.

[@mauricioachigar]

is there any reason related with politeness for white to invade at R3 instead of R17? i assume R3 is black's lower right corner, so white is invading right in the face of his opponent. (of course both moves are symmetric). thank you for your great videos michael

[@MichaelRedmondsGoTV]

While there seems to be a tendency to choose the UR above the LR corner there is no convention about politeness after the first move, even in a symmetrical position. For the purpose of a commentary, I like to do explanations towards the top of the board as it gives better visibility when using a real demo board.

[@kaceysaff9353]

Were any follow-up videos ever recorded for this?

[@maxbrass8725]

43 comments! Does Mr. Redmond have a certain appreciation for ladders?

[@zhifengwang8568]

三

[@수분있는건과일꽈자양]

왠아이산렌세이