Open hex 13x13 problem Hex, Havannah

45 replies. Last post: 2016-09-12

Reply to this topic Return to forum

Open hex 13x13 problem
  • lazyplayer at 2016-09-02

    From the variations of a game that we played with Arek:

    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2e6

    Is white winning? Has black made any losing error after k10?

  • Jos Dekker at 2016-09-06

    Analyzing this game, I only had at winning sequences for white so far :)

    I haven’t looked for any mistake in blacks moves yet. I’ll try :)

  • JKB at 2016-09-06

    I can’t claim this is a mistake, but I don’t have full understandind of the top left corner. I don’t actually see, why black prefers the pattern he played to something more normal to my eyes as this:

    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3d3c5b5 

    But in this variant I still favor W.

  • shalev at 2016-09-06

    Maybe 17. d7 for black?

  • lazyplayer at 2016-09-07

    shalev, 17. d7 seems fine for white:

    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2e6d7g5e7d6b7b6c6c5d5d4f5f6e5g3h4f7

    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2e6d7g5d6c5b7b5f5f6e5g3h4f7g7h5g4d8c9c8a9b9a10b10a11b12b11d11c10f10j9j6l3g9i8f8k6l7l4j8j7i9

    JKB, your variant is a brutal white win:

    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3d3c5b5b6f6f5h4g6f4f7d8

    And i suggest you change your mind about what you consider normal. Arek and maybe some others have played some games of that but i think it was only experimental. If you play d3 c5 b5, you end up “nearly” wasting a turn when you play b5. If you only need a 2nd row escape and you don’t really need b5 then it is ok. If you actually need a 3rd row escape, then probably it’s better to just get it with one move: http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3

    The “main” question is between b4 b3 d3 f2 or “play somewhere else”. And after f2, white again has puzzle between c5, e6 or “play somewhere else”. Also note that I’ve taken e6 from some old Maciej games. And he has taken c3 in his last games from me and Arek. :)

  • lazyplayer at 2016-09-07

    Another way to play the corner is this (on top-left):

    https://littlegolem.net/jsp/game/game.jsp?gid=1803010&nmove=7 <- here i was really unsure between c5 and e4

    Finally still another is the one you see on bottom-right here:

    https://littlegolem.net/jsp/game/game.jsp?gid=1804634&nmove=4

    I think against this last one “play somewhere else” may be the best, as it is shown in my game. Then usually L11 2nd best, and K11 3rd best. Now we’ve only to see if i win or lose that game, and prove my beliefs over the board or not :D

  • lazyplayer at 2016-09-07

    correction: when i said " then probably it’s better to just get it with one move:" i wanted to point out this: http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4d3

    as being “usually” better than this in the initial stages of the game: http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3d3c5b5

    and by the way, the value of “playing somewhere else” (tenuki using Go terminology) depends a lot on what is available somewhere else... naturally!

  • lazyplayer at 2016-09-07

    In games Arek vs Daniel you find many examples of b5 being played by player that will lose:

    https://littlegolem.net/jsp/game/game.jsp?gid=1789616&nmove=12

    https://littlegolem.net/jsp/game/game.jsp?gid=1781170&nmove=10

    https://littlegolem.net/jsp/game/game.jsp?gid=1742130&nmove=10

    Also one example where b5 was played by winner:

    https://littlegolem.net/jsp/game/game.jsp?gid=1640449&nmove=10

    it’s 3-1 to me... but i’ve not looked all their games...

    in summary i really think daniel should stop playing b5 early in the game... :)


  • shalev at 2016-09-07

    I still think 17. d7 has a chance. Maybe this

    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2e6d7g5f7d6b7b6c6c5d5d4f5f6e5g3h4h6g4c7b8c9b11a10c8e8e7f8d9d8g7

  • lazyplayer at 2016-09-07

    well, you’ve misplayed the bottom, it seems white win:

    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2e6d7g5f7d6b7b6c6c5d5d4f5f6e5g3h4h6g4c7b8c8b9c10c9d9d8e8e7f8g7h8

    but i agree d6 is better than it looks... (it looks horrible to me, but the visual is misleading)...

  • JKB at 2016-09-07

    2 lazyplayer. Thanks for such detailised comments! I kinda see that b6 is usually a bit “slow” (using Go terminology again) and this threat isn’t that big usually.
    Correct my thinking, if I’m wrong. This pattern: http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3d3c5b5 is bad, because B doesn’t really need these ladders escape on the top. While this pattern: http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2 allows B to “attack” the left side because of b4 and c4 stones.
    After shalev posted his move (d6) I appreciated that top corner pattern a bit more for black.

  • lazyplayer at 2016-09-07

    JKB; i agree with what you’ve said. Also take note that at the beginning, b5 should be bad because naturally there are “high value” moves somewhere else in the board. Later on, b5 could be bad because that threat is no longer very relevant. So overall, i think it’s rarely useful. But my my experience and thinking is limited, so who knows... :)

  • Arek Kulczycki at 2016-09-09

    I think you have no right to neglect black j6. Anyway I only managed to ALMOST win with black after j6...
    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2e6d7g5f7d6j6k11j11k9l9l3j3k4b7c7b8c8b9c10b12a12c9d9d8e8e7f8g7g11

  • Maciej Celuch at 2016-09-10

    Arek: In Your situation black plays 23.k4 . Black wins.

    In my opinion here White wins and 18.G5 is the loosing one for white because of that Black wins .

    Here are some examples

    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2e6d7g5f7d6j6l3b7c7b8c8b9c10b12a11c9d9d8e8e7f8g7

    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2e6d7g5f7d6j6k4l3k11j9j11j10i11i9

    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2e6d7g5f7d6j6k4l3k11j9c11l4k5l5k6l6k7l7k8l8k9l9j11j10i11i9d12h10g12f12g11f11g10f9


  • Arek Kulczycki at 2016-09-10

    Oh, right Maciej, I haven’t even looked for that. In that case the interesting line could be http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2k11j11k9l9e6

    (I hope this isn’t valid anymore, but can’t check right now http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2k11j6)

  • lazyplayer at 2016-09-10

    Maciej, then what white should play instead of 18 g5? 18 d6?

  • Maciej Celuch at 2016-09-10

    lazy: 18.c9 is the solution for white. 18.g5 do not cooperate with f2 stone, and 18.c9 is the kind of “let's see what does the opponent” move. Cooperating with f2 stone is showned here:

    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2e6d7c9f7f5g5h3

    “let's see what does the opponent” examples:

    19.g6 , 19.h5 , 19.g5 , 19.f6 , 19.f5



  • Maciej Celuch at 2016-09-10

    lazy: 18.d6 is wrong because of blacks h4 threat (J6 and then h4 for black).

  • lazyplayer at 2016-09-10

    Maciej, yes, thanks for pointing out... i knew white had to find a “quick” win, but just didn’t bother searching for it... In one of your examples earlier i think you missed a resource for white, but anyway it remains black win: http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2e6d7g5f7d6j6k3k4l2c7


  • shalev at 2016-09-10

    Maciej, 18. c9 is very nice! I didn’t really think about it yet, but is it obvious that white connects to the right in this position? How is it supposed to go?

    Example of a black attack on the right side: http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2e6d7c9f7f5d5d6b7c7b8c8b11a10h3g5j5k3l3k4l4k5l5k6l6k8l9m7k7i8j8j11k12j10i9f11g10f12h9h11e10

    What did I miss?

  • lazyplayer at 2016-09-10

    shalev, this

  • Maciej Celuch at 2016-09-10

    shalev: http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2e6d7c9f7f5d5d6b7c7b8c8b11a10h3g5j5k3l3k4l4k5l5k7l8m6k6i7j7i9i8f10

  • shalev at 2016-09-10

    lazyplayer, my line already assumed white plays this. I’m worried about the right side, not the left side.

  • shalev at 2016-09-10

    Maciej, fair enough, but that seems really brittle. Like, here’s an improvement for black that wins that line:

    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2e6d7c9f7f5d5d6b7c7b8c8b11a10j5k3l3l2m2k4g6h4l4k5l5k7l8m6k6i7j7i9i8f10h8

    (White is forced to block the top because of this)

  • Maciej Celuch at 2016-09-10

    Shalev, ok you’re right. Now do the same with 20.d6 rather than 20.f5 anyway white wins

  • Maciej Celuch at 2016-09-10

    ohh sorry 20.d6 is bad, but this is good

  • lazyplayer at 2016-09-11

    shalev, shame on me, i had forgot the problem on right... indeed i’m not even sure i had seen it when i played I4... hence i must conclude that the whole open problem is bullshit and white is totally losing! sorry! oops!

    indeed even i5 is suspicious there, maybe i should have played: http://www.trmph.com/hex/board#13,m1i10e9e4h7i6

    anyway, speaking of i5, another “open problem” is this: http://www.trmph.com/hex/board#13,d1d1e9j11i5h4

    find a move that gives white some HOPE to win for real (instead of just creating confusion)!

  • lazyplayer at 2016-09-11

    hum i mean, find a move that gives hope for black!

  • Maciej Celuch at 2016-09-11

    shalev, I take back voting for white.  The i4-i5 group is just too weak.

  • lazyplayer at 2016-09-11

    Even the great Maciej is confused by m1, heheh :))

  • Maciej Celuch at 2016-09-11

    Lazyplayer, to be honest I missed the black’s j5 thread on the right :-) When shalev mentioned it I realized that I was wrong. Then I was just simply defending white player until the end. 

    To “conclude that the whole open problem is bullshit and white is totally losing” try to defend black here (where we add Arek’s interesting line):

    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2k11j11k9l9e6d7c9g6g5f6f5j5i6

    or here (where we add similar line):

    http://www.trmph.com/hex/board#13,m1i10e9e4h7i5k10i4c4c3b4b3d3f2d2k11l10k12j11h12i13e6d7c9g6g5f6f5j5k3l3k4l4k5l5k6l6k8k9l8k7i8j8j9i9d12

  • lazyplayer at 2016-09-12

    Maciej, well, i played white in that game, so i naturally love white position more! Actually in the real game with clock ticking i missed 10 c3, and blundered on the left side. So i did come here to ask about 10 c3, if it really works as it seemed to me or not. But by the time i posted here, i had forgot the problems on right side. So in conclusion 10 c3 seems interesting, but it has to be followed by k11 and then the game will be long lasting and not trivial to analyze.

    I also feel that the idea of white playing somewhere else after f2 may be the best continuation for all corners like that. And in future we’ll see plenty of corners like that! What do you think? Will you try it in your games? I’ll try it in mines!

    Anyway, the k11 lines i’ll analyze when i’ve time, but i expect them to be beyond simple analysis...

    I would like to ask you about the 2nd problem, this one. Here the goal is to find a black move that is not trivially beaten by white. I guess black is losing, but how is it possible that it loses so badly, so that at move 6 the game is already totally over??? is 13x13 so bad?!! i sincerely hope not!

    I’ve tried pretty much everything, including c4, d5, d6, g6, j3, l4, j6, j7, k11, l11...

  • Arek Kulczycki at 2016-09-12

    Definitely agree with lazyplayer on “playing somewhere else”. This is actually the only logical solution in a position where is a choice of two or more seemingly equivalent moves. However it is extremely hard to postpone all the important decisions in a game.

    According to the 2nd problem. I say A4 is just too strong opening to even try to compete. Nevertheless, both j11 and h4 aren’t optimal play in my understanding of the game. I think it isn’t easy win, but black has to be very careful for white direct threats like g6 and j3. Once black has slowed down all the immediate threats he should be actually in a better position. There are things to try for black like k10, f7 and plenty others, but the clue is to order them right.

  • lazyplayer at 2016-09-12

    Arek, you think center will at some point win for black? I think black has to survive for that long... it seems totally impossible...

    I think that maybe this is the best black chance, haha... absurd!

  • lazyplayer at 2016-09-12

    Or maybe something like this... i think black must play very corner-ish to compensate for having played so much center-ish earlier...

  • lazyplayer at 2016-09-12

    And about a4, I’ve got to agree that on 13x13 it seems even more scary than 11x11. It could be like a2 being stronger on 9x9 than on many smaller boards. Maciej indeed did told me in the past that center is weaker on larger boards, and recently I’m getting more and more convinced he is right.

    On the other hand i think center is also bigger on larger board, so you must play into it eventually... :)

  • lazyplayer at 2016-09-12

    Side note, why nobody is playing 19x19? It could be interesting, and with a different “taste” than 13x13...

  • shalev at 2016-09-12

    lazyplayer, maybe also this if we’re going with weird corner moves. It neutralizes a lot of the a4 threats, and gives black a way to connect to the top.

    Anyway, I agree a4 is way too strong, and also black’s other moves are not optimal.

  • HappyHippo at 2016-09-12

    I wish we had an intermediate size between 13x13 and 19x19. 14 or 15 would be great. 15 would probably be better seeing as 14 is pretty close to the 13 we already have here. 

  • lazyplayer at 2016-09-12

    I’ve found found i’m in "hex19.DEFAULT.mc.2009.final" together with strong players...

    what is this? a yearly cup among monthly cup winners? :-)

  • lazyplayer at 2016-09-12

    shalev, well, i know that move, and i agree that usually it’s better than b2, because usually black wants an escape for “ladders” on to. In this case anyway i gave a try to b2 to see if black could do some magic on the left side. It can do some magic, but at the cost of losing totally everywhere else, hehe...

    The best probably is k11, because k11 makes j3 a valid threat again, and so it constrains white reply somewhat. When i played i5 i had naturally the idea of delaying k11 for some time, but “for some time” is 1 move at most, and that’s it.

  • shalev at 2016-09-12

    Note that in the k11 line, white can play k9 instead of k10. This gives black the j10 square instead of the l8 square (plus the option of getting the l7 square, which is not that strong but provides some option value). Sometimes k9 is better than k10. I’m usually tempted to try it in these situations.

  • lazyplayer at 2016-09-12

    shalev, this is the answer to that... ;)

  • shalev at 2016-09-12

    Yeah, fair enough. It’s still “just” the j10 square, but the j10 square is a lot more valuable for black there than the l8 square, because it really solidifies the connection to the bottom.

Return to forum

Reply to this topic




Include game board: [game;id:123456] or [game;id:123456;move:20] or [game;id:123456;move:20;title:some text]