9x9 Hex solutions?

Hello,
are ALL starting positions for 9x9 Hex solved?
If so, where can I find the complete result?

Ingo.

Last time I checked, not all of the 9x9 opening moves had been solved. See http://webdocs.cs.ualberta.ca/~hayward/hex/ . I'm not sure whether that page has the most up-to-date results available.

complete 9x9 table on page 9 of http://webdocs.cs.ualberta.ca/~hayward/papers/pawlhayw.pdf

Woah, I didn't know about this!

It looks like they conclude a2 is winning in 9x9 (but a3 is losing) - awesome!

Shalev, thx for link! I haad spent hours trying to beat A4 on 9x9, and finally conclued it's not possible.. very pleased to see it's true... :)

And yeah even A2/B2 winning is indeed surprising/beautiful! Hex is a true mess! 7x7 and 8x8 they lose, 9x9 they win! Go figure! :)

Actually, thx Mirko! :)

Thx a lot for the answers, especially to Mirko for
giving the link on the Edmonton paper.

Any special comments on the example line for the game
starting with Black on a6? (Page 9 in the paper)

Ingo.

Ingo, i've looked at it. I think against a6 there are more "logical" winining sequences, that don't involve playing all around the board like that. But i'll give a more careful look later on, when i've some time...

Err, *with* a6.. not against :)

I've looked at it too and it's interesting example. I think white should have played 10.b7 instead of 10.c7. I can't find any solution for black to win against 10.b7.
http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4h4g7f7d8c8c2c6b7

Maciek, I think this is a solution
http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4h4g7f7d8c8c2c6b7h5

Oops, the previous one was wrong, but this I'm sure is correct

http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4h4g7f7d8c8c2c6b7e3

I have been searching for a winning alternative to black 3.h4 and incredibly no find so far... Is h4 the strongest choice?

The earliest improvement I believe I found for black is this:
http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4h4g7f7d8h8

Arek, e3 is wrong. Here is the answer: http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4h4g7f7d8c8c2c6b7e3d5g4g2h2f5

Maciej, I think d5 does it (I didn't verify 100% though):
http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4h4g7f7d8c8c2c6b7d5

You're right shalev, my bad ;-)

Arek, i also would be relly surprising if H4 is the only one!

But who knows, it can be! I've not yet analyzed...

Huh, this seems a trivial and instant black win, but probably someone should check after me :)

http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4d5

Arek, I didn't look in detail, but what's the plan for responding to b8?

http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4d5b8

Arek, http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4d5b7 ?

i was looking at the same, and found this.. :)

http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4c5

This maybe?

lazyplayer, what's the next move after c7?

http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4c5c7

shalev, http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4c5c7b8b7d6d5e4d4e5f2e3e2d3d2b3c3a5 :)

@shalev
maybe this, I don't see anything for white
http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4d5b8h2

@lazyplayer
oh that's easy, no?
http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4d5b7c6

Hm perhps it is not so easy... i go sleep, see you tomorrow :)

Arek, http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4d5b7c6e3 easy?

@lazyplayer
sorry, true, maybe the same h2 as against b8 then...

http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4c5c7b8b7d6d4 <- This is refutation of my proposal :]

Lazy, about c7, b8, what about b6?
http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4c5c7b8b6a7b7a8c3b4b3c4d3e4f2g2f3

shalev, http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4c5c7b8b6b5b7d6d5c6d4c3

The problem with my proposal is the line i've shown you...

http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4c5c7f6e6g4g3b8

Maybe this? :D But now i go!

lazyplayer, I see... I'll look at your new line.

Arek, I'm still looking at your suggestion too... we should really get the solver to do these =P

I withdraw my suggestion due to this:
http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4d5b8h2e7

this is my new favourite http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4d5b8f5

however, against b7 my h2 was correct:
http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4d5b7h2d8e7

http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4d5b7h2d6f5

Arek, yes, b7 loses to that h2, nice :)

Arek, my feeling is blue should play http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4d5e3 against D5... :)

Hmm no, stupid idea.. Error :)

Heh, ask Fat Phil to help us by using MoHex

I tried to transcribe this game so I could study it. I get
http://www.trmph.com/hex/board#11,a10j1b10j2c10j3d10j4e10j5f10j6g10j7h10j8i10j9a6f4h4g7f7d8c8c2c6c7a8b7a7g2h2e9d4e3e6d7d2b5a5c3a4b2g8f6f8e7b6d5h3g6c5d3f5. In case others would also like to experiment.

Hi guys,
I see very interesting discussion here. Few lines are proposed. Maybe I can help you with actual solver used for 9x9 solution? Do you want me to analyze some position?
Or, you can build your own solver from the sources. I can share the database with solution too.

How large is your database? It would be great, if you could post it on BoardGameGeek and then announce it there or in your own thread.

Does the solver use a GUI with a graphic board display?

The database has size 2.6 GB. Solver uses HexGui. Actually you can run solver through programs like wolve or mohex. Some patches on fuego, which benzene framework depends on, are needed to apply. I can provide them. Rest is on my github repositories

You might be interested in a slight modification I made to HexGui. There are more display options for the board. The instructions provided, to modify the code, no longer seem to work. If you could describe your patches to me, I could add the info to that page, thanks very much. My email is twixtfanatic atsymbol gmail period com.

I'm not sure how you could transfer a 2.6 GB file. Is it posted on your github repository? I did not see it. Might there be a way to pare down the tree, to show only enough of each variation that a strong player, or Mohex, could figure out the rest?

Are 41.a5 and 43.a4 of any use? I would like to think of this as the Platonic 9x9 game, but those moves are causing me cognitive dissonance.