https://www.littlegolem.net/jsp/game/game.jsp?gid=2155321

I resigned because I didn't anticipate black's final move.

I calculated two other possible responses which led to wins for white.

spartacus, play mostly 11x11 games for a few weeks, you've to understand the basics first.

As far as I know there are two good ways to learn, one is to simply copy the strong players and the other is to play many games on smaller boards.

Another valid idea is to pay a lot of attention to endings, the games among weaker players are almost always decided near the end. The last person to blunder loses. In that game the last blunder is at 40. 40 L2 seems white win. The lesson in this case is that L2 is generally than M1 on the empty board.

40 K11 is another option worth consideration.

Overall white had the upper hand during most of the game, I can't immediately see any obvious win for black.

Most of all you have to play bad moves on small board so that you quickly see why the bad moves are bad. Best would be 9x9 vs a training bot, but I don't know if any bot is available like that.

You cannot see that on a big board because of 10 moves you make maybe just 1 is bad or maybe all 10 are bad - you don't know.

Then when you already know good moves on small board you have to indeed copy strong players. Or the way I prefer - just trust me in 1 rule. Big board is proportionally the same as the small board ;)

For example your play here

https://www.littlegolem.net/jsp/game/game.jsp?gid=2155321&nmove=6

is proportionally equivalent to this:

http://www.trmph.com/hex/board#11,b1b1i3c9i8h11

Which is more obviously terrible.

HI guys, thanks for your response! I just have a few questions.

Lazy player, could you clarify your point about the final blunder? The analysis of the move 40. L2 doesn't make sense. L2 was a move played by black on turn 43. On turn 40 white (me) played L3. Was that a mistake? Also, you mention white is ahead most the game and that you can't see an obvious win for black. I was white, so are you saying I should not have resigned in this final position?

Arek, can you explain what you mean as proportionally the same?

By the way, I have played many games of Hex on small board. I have studied games by strong players here. I have even studied games by both of you. I don't know how much I learned from this…also I have read material online about hex strategy. I am only looking for concrete analysis of my mistakes in this particular game. Thanks in advance.

I want to make it clear that I played Hex on other platforms , against AI, and also IRL, so the title may be a little misleading. It's not my first game. It was my first intense game on this particular platform, where I felt like I knew what I was doing, thought I was gonna win, but then lost.

spartacus, the hypothetical 40 L2 would have been white win with perfectplay. It hasn't happened because you've blocked him on top with M1.

The idea of copying small boards on larger boards, just scaling things up, it's highly misleading. I used to think like that but then I realize it's not true. Nonetheless small boards are useful to learn the small patterns. When you switch to larger board you'll have to learn new larger patterns. The small patterns reappear in the local battles.

One quick comment: You would have been better off playing 4. j4. You've already got a stone (c1) on that left edge, by playing 4. d10 you let black play 5. j4. After that move, Black now has a corner on both of their edges (north and south), which is balanced, while you have overcommitted to your left edge and have neither corner on your right edge, so you're already at a disadvantage.

Oh yeah, and a stronger approach to the final sequence would be (I think this is what lazy was talking about): http://www.trmph.com/hex/board#13,c1c1j9d10j4i12f10g8i7h9i9h10i11i6k4k5h7h6f7g5f5f6d7e7d8d9e8f9g7i8j7j8l7k7l6k6m4l4m3l2l3j3i4i3h4h2

Some followups to consider:

http://www.trmph.com/hex/board#13,c1c1j9d10j4i12f10g8i7h9i9h10i11i6k4k5h7h6f7g5f5f6d7e7d8d9e8f9g7i8j7j8l7k7l6k6m4l4m3l2l3j3i4i3h4h2h3i2g3g2e3d2f2f3d5e5e6d6b7c6

http://www.trmph.com/hex/board#13,c1c1j9d10j4i12f10g8i7h9i9h10i11i6k4k5h7h6f7g5f5f6d7e7d8d9e8f9g7i8j7j8l7k7l6k6m4l4m3l2l3j3i4i3h4h2h3i2f3g3g4f4d5e5e6d6b7c6b6c5b5d2c4e3

A good rule of thumb is to hold your opponent to the higher row ladder, if possible. Letting Black get a second row ladder on the north edge gave them more control.

HappyHippo, you had an online book somewhere? You can reference it. :)

@Spartacus, by proportional I mean geometrically proportional.

In simple words (I hope):

Assume you draw with a pencil a Hex position with stones on the board (whatever size). Then you change stones into dots (of width ~0).

Then you erase all the grid so you have only edges left.

Then you can draw a different grid inside that frame and the dots will end up as stones again but with different coordinates.

My argument is that this new position is very close to equivalent to the initial position from a different board size.

Then to be even more precise, but this may not be understandable.

To be closer to what I think we will repeat the above process with 1 additional step at the beginning and 1 at the end.

First draw with a pencil a Hex position with stones etc…

Then imagine your picture is 3D and the middle of the board is pulled backwards in the third dimension so that the hexagons become destorted (they are smaller in center of the board).

Then you change stones to dots, erase all the grid, draw another grid (pulled back in 3D again) and so the dots become new stones.

Then the board is pushed back to being 2D and you get an almost equivalent position!

If any of you understand what I meant I will be honored :)

Lazy: Indeed! It's in my profile if you ever need to find it

Spartacus: You could read my book if you haven't already: http://mseymour.ca/hex_book/hexstrat.html

HappyHippo, you've added some new content? You should use some timestamping/versioning!

And congrats for the typesetting, it's really well done, although I prefer the more standard orientation of the board.

I added a few minor sections shortly after publishing. I've been considering some more substantial changes; if I do that it'll be as a “Second Edition” and I'll keep a link up to the original.

A few people have commented on the typesetting, which pleases me because I put a lot of work into it :)

Thanks all! In fact I have read a large portion of Hippo's book. It was fantastic!

I only briefly looked through HappyHippo's book. Apart from that it's beautiful and well written I've just now found the very convenient way for commenting games!

@HappyHippo, what do you think about creating a public option to comment on our games with the use of your interface?

@Arek, what are you thinking here? If you're thinking of something like twixt-commentator, where anyone can comment a game and share it, it's probably not feasible for me to build a website like that currently.

On the other hand, I would LOVE some guest commented games! If you (or another player!) are interested in commenting a game (or games), I could format it and put it on my site (with all due credit to the commentator of course). Let me know what you think

Yes exactly guest commented games. But it's surely convenient to comment when having such tool with diagrams. Whatever works for you, I also might help. I didn't know twixt commentator, something of this kind would be great, although it's way uglier than what you did ;)

In fact someone did make a Hex commentator: http://www.hexmaster.net I thought it would be down but it's still up! Sadly underused.

(This was made by Nate Watson, see this thread: https://www.littlegolem.net/jsp/forum/topic2.jsp?forum=50&topic=572 )

Back to the commentary - we should get in contact by email (you can find my address at the bottom of any page on my site). Do you have a game in mind that you'd like to comment?

I don't have any game in mind. I just would be glad to share thoughts on some top games that happen here and there. And hear others opinion on my owns.

The commentary tool is very nice actually! I would add there some filters and sorting, but it's crazy that there are such tools and I've never even seen or heard about them.

Arek Kulcyzcki, you said that the big board is proportionally the same as the small board, do you mean by this that if 4,4 is a good move on 13x, that 8,8 would be a good move in 26x?

@lguser,

yes I meant that, except that the proportion is distorted

Your question basically is: for board size SxS and move YxY is it equivalent to (k*Y)x(k*Y) on board (k*S)x(k*S).

This is an approximation, but closer approximation to truth would be that for board (k*S)x(k*S) the equivalent is (k^p)*Yx(kp)*Y and 0<p<1

In other words - a9 (1,1) on 9x9 is equivalent to a13 (1,1) on 13x13 and equivalent to a19 (1,1) on 19x19 and so on…. But maybe b8 (2,2) on 9x9 is equivalent to c11 (3,3) on 13x13 and d16 (4,4) on 19x19

In other words again - moves nearby edge change just a little, the further from edge the more they change with board size.