This is a dedicated game ot discuss the game Geoff vs 7ics in the recent championship. The game can be found here:

http://www.trmph.com/dnb/game/lg-1615321

I apologize for my limited participation in the discussion until now, but the plan is if we get some interesting analysis and alternate lines to produce a video illustrating these, which may be of interest to players rated 1850+ (also possibly to ambitious lower rated players).

*This is a dedication thread to discuss

*dedicated (I miss being a member who could edit my posts).

Thanks Christian.

To put it simple, maybe everyone can go into Geoff shoes reviewing this game and trying to decide where one would have changed something to one's play?

As I said in the other thread, for me it is as soon as 11. where you need to decide if you want to do the preemptive sacrifice or let the 6-loop happen.
e.g.:
http://www.trmph.com/dnb/board#5,h7d3i8i6f5f3f7f9d7c4k8

When there are 6L and 4L pending it is good to have a look of one of the tables of wccanard giving the value of the game when you add a chain:

position    |value value(posn+n) for n=4  5  6  7  8  9 10 11-------------------------------------------------------------   .        | 0                        4  5  6  7  8  9 10 11   4L       | 4                        0  1  2  3  4  5  6  7   6L       | 6                        2  3  4  5  6  7  8  9  4L+6L     | 2                        2  1Q 0Q 1  2  3  4  5

if formating holds you'll see that in the case 6L+4L you seek to get another chain of 6 so that the value if zero if you are not in control.

carroll, i just saw your suggested variation in the “video matches”-thread and i'd say it's a win for p1. however, i dont agree with you regarding the suggested line for move 14. allow me to substitute that line ;)
http://www.trmph.com/dnb/board#5,h7d3i8i6f5f3f7f9d7c4k8i4k6j11

Move 11 is the first move that The Shark knows to be losing.

The Shark doesn't know who is ahead after move 10, but it is of the opinion that player 2 is ahead after both Christian's suggested 11th move (k8) and Geoff's actual move (e4).

Tanks, yes clever move which prevents i10 because now the 6L sacrifice creates 3 real chains.

I was able to extend my analysis:

After move 10, Goeff was ahead and could have maintained the lead with j7, f1, a4, or b7.

Also, after Goeff's 15th move (b7), 7ics could have played i4 and been ahead by 3.

Waoh four quite different moves!
D&B is full of mysteries.

Thanks for the analyses, everyone.
It's nice to know that 5x5 D&B isn't strongly solved quite yet. 10 moves of mystery still remain (if you're p1)

Hey! I've done some analysis to the game as promised and diego44 joined me to assist. The analysis is focussing on the first 10 lines that remain a mistery and is trying to convey Geoff's and 7ics's thoughts as well as showing a bunch of variations for alternative play (the variations have been verified to be correct from move 13 onwards). We hope you'll enjoy reading it and feel free to ask if there's anything unclear, wrong, fishy or whatever. :)

Geoff Cameron (p1) vs 7ics (p2)

“1.h7”:http://www.littlegolem.net/jsp/game/game.jsp?gid=1615321&nmove=1

Geoff's standard opening line makes it really hard to aim for only one chain, since it cuts the board in between. Common responses are 2.g6 and 2.g8, which use the line to create a potential third area.
Alternatively one can play 1.f5 which has a similar effect. I personally prefer 1.f5 over 1.h7, since it seems less possible to create three chains. The positions 1.f5 2.g4 3.h3 and 1.f5 2.g6 seem to make an outcome with three chains less likely than 1.h7 2.g8. However, this fits Geoffs playing style, since he usually doesn't hesitate on sacrificing. Maybe he even wants his opponent to try to create a third chain in a small area unexpensive to blow up.

“2.d3”:http://www.littlegolem.net/jsp/game/game.jsp?gid=1615321&nmove=2

I've never seen this response on h7 before, but I certainly think very positive about it. 7ics usually doesn't try to forcefully create three chains but is instead keeping both the possibility for one and the possibility for three chains alive (quite popular playing-style among top-players), which is pretty hard to fight against, since a move that prevents three chains might support the one-chain possibilty and the other way around.
2.d3 builds space at the bottom where a third chain could possibly exist and is at the same time part of the 6-Loop that is later on threatening to be formed and thus supports the one-chain possibility as well, since the 4x4 field in the top-left (a11-i3) might very well turn into only one chain. 7ics will adapt his play depending on Geoff's moves, since he doesn't have a clear number of chains as his aim.
Other lines following the same idea are h3, f3, c4, c6 and c8, but I think that d3 is well chosen and I don't see any need to vary at this point.

“3.i8”:http://www.littlegolem.net/jsp/game/game.jsp?gid=1615321&nmove=3

This line is played in order to prevent g8, trying to turn the top part of the board into one large area. There is quite a bunch of variations possible at this point. The lines e2, a4, f11, h9, h11, k8, k6 and probably a bunch of other lines all try to prevent a third chain, which is what 3.i8 does. 3.i8 is in my opinion the most obvious of these lines and it certainly serves its purpose. That being said, there is the before mentioned downside that a move that decreases the probability for three chains also increases the likeliness for one chain, so this actually kinda supports 7ics's playing style in a certain way. Alternatively he could have kept on preventing the one-chain option with moves like e4 or g6 sticking to the plan of blowing up a threatening third chain for small costs, though I won't pretend this is more effective than what he does with i8. In fact, I'm pretty confident that this is still winning for Geoff.

“4.i6”:http://www.littlegolem.net/jsp/game/game.jsp?gid=1615321&nmove=4

4.i6 has basically the same effect as 2.d3: builds space on the right edge of the board and is partaking in forming the shape of the 6-Loop. Note that responses at one edge might be transferred to the other, meaning that both edges together (bottom and right edge) contain either 0 or 2 chains adding up with the one in the 4x4 field — however this is of course easier said than done. i6 also prevents the use of g6, which seems to be a good line for player 1.
The shape in the top-right is well known from the 1.f3 2.h3 3.g4 opening and favors player 1, which might be a reason to avoid 4.i6 here. Alternative lines for move 4 working alike are f3 and c4, but I'm not quite sure they are better than i6.
Another interesting response on i8 seems to be e8 which cuts the top in between and just messes with the structure maybe causing wrong play.

“5.f5”:http://www.littlegolem.net/jsp/game/game.jsp?gid=1615321&nmove=5

The reasoning behind this move is very unclear to me, but I think it's just supposed to seperate top left from bottom-right (draw a line between all moves that are currently done and you'll see what I mean). I did consider this to be the mistake until The_Shark verified that the sacrifice in move 11 was wrong - oh well, maybe it is a mistake and 7ics just made a re-mistake later on. After all I just don't really get what advantage results from playing f5 here, as it seems to support the possibility for the 6-Loop and is in my opinion advantageous for 7ics rather than Geoff.
Instead I would play g4, variations being:
“Variation with g4 no.1”:http://www.trmph.com/dnb/board#5,h7d3i8i6g4c6j7e8d9d5f7f5g2g10a6b7b5a8b11h3k2j3k4i2j1h1c10d1i10e4f3c4a2e2f1c2h11h9f9g8j11e10f11d11j9k10k8b3a4b1a10b9c8d7e6g6h5i4j5k6
“Variation with g4 no.2”:http://www.trmph.com/dnb/board#5,h7d3i8i6g4c6j7c4b5c8b9c2d7f5k4f11j3c10e10j5k6i4f7d9d11e8h11j1i10k10g10g8h9f9a2g2h3h5g6e6d5e4f3b11a10j9j11k8b7a8a6b3a4b1e2d1f1i2h1k2
“Variation with g4 no.3”:http://www.trmph.com/dnb/board#5,h7d3i8i6g4c6j7e8d9d5f7e10g10f5a6e2e4c4f3f11f9h9g8j11i10h11h5g6e6d7c8b5b7b11c10d11b1c2d1k2j3k4h1i2j1b3a4a2b9a10a8j9k10k8k6j5i4h3g2f1
“Variation with g4 no.4”:http://www.trmph.com/dnb/board#5,h7d3i8i6g4i4h1k8k4g10h11e4i2k6e6d7b5b7f7h9i10g8c8e10j9j7j3j5f11f9e8d9b11c10d11c4d5a6c6a2c2f3f5g6h5h3g2j11k10k2j1b9a10a8b3a4b1e2d1f1

“6.f3”:http://www.littlegolem.net/jsp/game/game.jsp?gid=1615321&nmove=6

This line is basically why I dislike the use of f5 in the previous move. I think that f3 profits from f5, because along with d3 there's now this shape that is oftenly used by 7ics in general and is hard to play against (see how even flipster, god of dots and boxes, failed to beat that shape once -> “flipser vs 7ics”:http://littlegolem.net/jsp/game/game.jsp?gid=1584096&nmove=4 ). Now this is where the earlier mentioned 6-Loop really starts to matter as it strongly looks like a 6-Loop is gonna be created there. Also the size of the space at the bottom is now the same as the space on the right edge, so this might support transferring lines from one edge to the other.

“7.f7”:http://www.littlegolem.net/jsp/game/game.jsp?gid=1615321&nmove=7

Most obvious line ever - short and sweet: if this line isn't played, then 7ics will just turn the top-left into one really big chain using e6 and can sacrifice everything else in the worst case and probably still wins, because the chain is freaking long. Oh well, after e6 one can sacrifice a box to prevent it from turning into one big chain, but early sacrifices are rarely the best way to go.
Here's two variations that don't seem to work out for player 1, supporting the idea of f7:
“Variation with g2”:http://www.trmph.com/dnb/board#5,h7d3i8i6f5f3g2e6g4e4b3c2a4k8j11i10c6k6e8j3d9g6f7h5c10h11a6i2k4f11b9a8e10d11b11a10a2b1k2j1e2d1f1g10f9h9g8c8b7d7c4b5d5h1h3i4j5j7j9k10
“Variation with f1”:http://www.trmph.com/dnb/board#5,h7d3i8i6f5f3f1e6d5k8k6h3h1j3k10b5j11k2h11e8a6c8e4g4c4d7d9b7c6f7g6h5i4e2g2j1i2b1c2d1i10j9j7j5k4b11c10f9g8h9g10e10f11d11b9a10a8b3a4a2

“8.f9”:http://www.littlegolem.net/jsp/game/game.jsp?gid=1615321&nmove=8

The purpose of f9 is to hold the top part and the 6-Loop together which is done by preventing e8. It's a pretty logical line that follows the original idea of turning the 4x4 field (a11-i3) into one chain. Since this is probably a position winning for Geoff (not for sure though) one could as well try to cause more chaos for a higher chance of causing a mistake. This could be done for example by playing a variation like 8.c6 9.b7 10.e8 which seems to be an even more complex position.

“9.d7"http://www.littlegolem.net/jsp/game/game.jsp?gid=1615321&nmove=9

d7 is cutting that 4x4 field into halves and can basically be played any time, since this is something Geoff should do in any case otherwise 7ics would just create a super long chain. After having tried some variations we came to the conclusion that this is a good time to play it. The reason is that the variations we tried did not look too good for player 1, so one might as well play safely here.
We tried some variations with 9.j7 and avoided the usage of d7 as long as possible which didn't work out:
“Variation with j7 no.1”:http://www.trmph.com/dnb/board#5,h7d3i8i6f5f3f7f9j7e2j1j3i4c4h1b1a4c6a6d7a8c10f11e10g10i10a2c8h3f1g2k2i2g6e6h5e4d5g4b9b7b5b3d1c2j11b11a10j5k6k4j9k10k8h11h9g8e8d9d11
“Variation with j7 no.2”:http://www.trmph.com/dnb/board#5,h7d3i8i6f5f3f7f9j7e2e4g4c6d7g8e8h9f11e10g10j11i10h11b11c8d9a10c10b9d11a6i4k4a4c4d5e6g6h5h3g2f1k2j5k6j3a2b1j9k10k8i2j1h1a8b7b5b3c2d1

“10.c4”:http://www.littlegolem.net/jsp/game/game.jsp?gid=1615321&nmove=10

Finally threatening with creating the 6-Loop which can only be prevented by sacrificing - Geoffs mistake. Creating the 6-Loop would finally turn the 4x4 field (a11-j3) into just one chain and all that's left for 7ics's strategy would be to balance the two edges, creating either 0 or 2 chains along the edges. This line has been proven to lose and is in case that Geoff hasn't been winning until here the last possibility for the re-mistake that I've talked about in the discription to the fifth move. In fact, diego44 and I found that a very strong variation for player 2 occurs with the usage of 10.h3:
“Variation with h3”:http://www.trmph.com/dnb/board#5,h7d3i8i6f5f3f7f9d7h3j7b3c4g2d1k4a2i4d9k2j5k6j3b1c2e2f1c8e8g8h9b11g10d11b9i2h1j1c6j11a6k10a4b5a8b7h11i10k8j9d5e6e4g6g4h5a10c10e10f11

“Mid- and Endgame”:http://www.littlegolem.net/jsp/game/game.jsp?gid=1615321&nmove=11
I won't write a detailed explanation of the moves after this, but briefly said:
The sacrifice in move 11 enables 7ics the possibility of blowing up 3-chains and what he does after that is to increase his advantage in gained boxes even further by threatening with a third chain. After move 17 7ics changes his aim from 3 chains to one chain and uses his advantage in boxes to blow up the 3-chain at the top (he could even blow up a 5-chain). Geoff has to doublecross in order to win the chain-war, however after the doublecross 7ics could gain another 5 boxes through sacrifices, so instead Geoff takes the boxes and enforces the sacrifices from 7ics. Geoff is now 3 boxes ahead but loses the chain-war and 7ics wins 12-13 by taking the long-chain.

After h3, I have determined that Geoff is ahead and can win with c8 (not necessarily the winning only move).

Dryad, that game you mentioned that I played against 7ics was not a good game from my side, so I don't think it means much. He (7ics) was a lot lower rated then, so I felt I could manage to beat him somehow after move 7 even though the position didn't look so good – serves me right that he beat me. My move 3 was weird – normally I would play g6 – and on move 5 I think something like j1 looks better than what I played.

As for the game in this thread, I thought Geoff's move 10 was surprising. After the sacrifice the bottom chain has to extend either to the bottom or to the right, and then in the other region player 2 can force some sacrifices. In a game like this I would just play out a few variations and then it's easy to see that p2 is winning – which is why I wouldn't expect such a mistake from a player like Geoff. Perhaps a case of underestimating his opponent as well.

Thanks a lot for these very interesting opening analysis.

I hope to see a great video retracing these lines.

I will do what I can. It has been some busy periods and sadly I have not been able to participate in the discussion. I am looking forward to presenting your foundings though :)

Ok Bill, I guess the sacrifice in move 11 really is the first mistake then, since I don't see what's wrong with 10.h3 and the moves before… but that was to be expected. Does 11.b7 lose after 10.h3? 11.c8 seems rather hard for a human to find and it doesn't really make things easy, so if that's the only winning line after 10.h3 then playing 10.h3 is definitely stronger than 10.c4. Would be interesting to know if there's a more obvious winning-move after 10.h3 than 11.c8.

Yeah flipster, I know you can do much better than in the shown game, but that absolutely doesn't change my opinion on the f3+f5+d3-shape. It's clear that you made a mistake somewhere in the opening-phase otherwise you wouldn't have lost, but that is exactly why I considered 5.f5 to be the mistake in the game between Geoff and 7ics… that setup is just very similar to the one in your game with 7ics. Let me ask you this: do you think 5.f5 is a well chosen move here?
Also: I didn't have the chance to ask you for another date on iggamecenter since you disappeared… when are you available? :O

Okay, I pushed the analysis back another move. The only winning move other than c8 is k4.

And now I'm trying alternatives to h3:

Nope. Geoff was already ahead by 1.

Impressive. I will no be able to do the video before 1st of april, so I hope no one is too impatient.

I apologize to anyone waiting for the video. I am not able to do it at this time and probably also not in the near future.