Arimaa Forum (http://arimaa.com/arimaa/forum/cgi/YaBB.cgi)
Team Games >> 2007 One vs TheMob >> Move 35
(Message started by: Janzert on Jun 20th, 2008, 3:05pm)

Title: Move 35
Post by Janzert on Jun 20th, 2008, 3:05pm
And chessandgo responds quickly:

35w De3n rd4n De4w Re2w

Janzert

Title: Re: Move 35
Post by Soter on Jun 20th, 2008, 3:30pm
First glance idea:  35b ed3e ee3s dc4n xxxx

?

Title: Re: Move 35
Post by Fritzlein on Jun 20th, 2008, 4:47pm
My analysis from before was that chessandgo needed to play his fourth step Da3n, because if he didn't we could play 35s dc4w db4w ce7s xxxx, with the idea of meeting a dog capture with 36s Dd4w ed3n Dc4n ed4w and a return capture the following move.  Now I see that 36g Rc2n Rc3n xxxx xxxx disrupts our plan and leaves us winning only a rabbit (maybe), but if we could get a rabbit for sure that isn't so bad because after a trade we have CRR for D and no longer have the dog hostage to worry about.  Just food for thought.

Title: Re: Move 35
Post by mistre on Jun 20th, 2008, 5:33pm
Fritz, I came across the same dilemma in trying to engineer a dog trade.  The problem is we need to be able to fill the c3 trap with a piece and stop his A3 dog from advancing and we can't do both in 4 steps.  Whatever I tried led to one of his dogs getting free on our side of the board and we have nothing to stop it.

I think Soter's move has potential with dc4w, db4w as the last two steps.  This way, it prevents him from taking our Dog right after the Horse and also blocks his a3 dog from advancing.  In this line I see us down a Horse for a Dog.  So essentially we traded C-D for H, so slight advantage to us, except he will have the last remaining H, so I am not sure.

So a potential move:

35b ed3e ee3s dc4w db4w

Any one have anything else better?

Edit: I think there are lines after this move where we only get a R for our H, which is not acceptable.  We need to either find a better 35b or better moves after this one.

Title: Re: Move 35
Post by 99of9 on Jun 21st, 2008, 1:59am
how about
35b ed3e ee3s Hf2s ee2e

This might be our chance to take over the other trap, but it's a sharp line, and we'd need to play out the exchange sequence.

Title: Re: Move 35
Post by Soter on Jun 21st, 2008, 4:01am

Quote:
Now I see that 36g Rc2n Rc3n xxxx xxxx disrupts our plan and leaves us winning only a rabbit (maybe), but if we could get a rabbit for sure that isn't so bad because after a trade we have CRR for D and no longer have the dog hostage to worry about.  Just food for thought.

It's much too early to say something with confidence ( is it  at all possible when playing Arimaa?  :)), but if it was up to me, I'd accept CRR vs D.


Quote:
Edit: I think there are lines after this move where we only get a R for our H, which is not acceptable.  We need to either find a better 35b or better moves after this one.

mistre, could you post some specific lines that are unfavourable to us ?( I simply have  little free time now...)


Quote:
how about
35b ed3e ee3s Hf2s ee2e

Hmmm, I'm afraid that Jean's H will run away and his E will eat our dog without  giving up the horse hostage. But I only looked at it for a while.

P.S. How would you Mobsters continue after this one ( is it at all plausble?)

35b  ed3e ee3s dc4w db4w
36w Dd4s Dd3e Rd2n Da3n
36b Hf2s ee2e Cg3e dg2n
37w De3n Hf1w Rh1n Ch3n

Title: Re: Move 35
Post by Fritzlein on Jun 21st, 2008, 8:35am
Hmmm, this is looking worse for us than I thought two moves ago, as none of our current moves seem good to me on preliminary examination.  I wonder if there is a tactical save here somewhere, or if we will have to cut our losses.  :-(  I'll try to post some lines yet this weekend.

Title: Re: Move 35
Post by warren on Jun 21st, 2008, 11:25am
It looks to me like we're going to lose our dog for minor compensation. It seems Chessandgo will get a choice between losing a rabbit and letting our horse go free. Which is he likely to pick? Can we get better compensation than that?

Update: the best move I see so far is 35b dc4s dc3n Rc2n xxxx. This move:
* makes progress on rescuing our horse
* temporarily neutralizes the c3 trap
* threatens to pull that rabbit to c6

For example
35b dc4s dc3n Rc2n xxxx
36w kill our dog; his horse ends at f2
36b use our dog to move his c3 rabbit to c5
He then either has to give up the rabbit or abandon our horse.

Title: Re: Move 35
Post by mistre on Jun 21st, 2008, 1:06pm

on 06/21/08 at 08:35:14, Fritzlein wrote:
Hmmm, this is looking worse for us than I thought two moves ago, as none of our current moves seem good to me on preliminary examination.  I wonder if there is a tactical save here somewhere, or if we will have to cut our losses.  :-(  I'll try to post some lines yet this weekend.


Just curious, but since you weren't around for the "kill the cat" vote, I was wondering if you would have voted for our selected 32b or did you find a better move?

Title: Re: Move 35
Post by Fritzlein on Jun 21st, 2008, 3:08pm

on 06/21/08 at 13:06:17, mistre wrote:
Just curious, but since you weren't around for the "kill the cat" vote, I was wondering if you would have voted for our selected 32b or did you find a better move?

Without taking time to analyze, I would have killed the cat.  The danger to our dog was going to happen anyway, right?

The last major turning point I see was move 29, without which chessandgo's horse crossing would not have had the time to work.  However, I guess we still had chances after that to give up our rabbit with a slightly better game, so if we aren't winning at least a little bit now, then something else must have gone wrong.

Title: Re: Move 35
Post by warren on Jun 22nd, 2008, 1:32am

on 06/21/08 at 01:59:43, 99of9 wrote:
how about
35b ed3e ee3s Hf2s ee2e

This might be our chance to take over the other trap, but it's a sharp line, and we'd need to play out the exchange sequence.


99's move was counterintuitive to me at first, but now I think it leaves us with a strong position and is my current favorite move. The key is our b5, d5 and g6 rabbits, which are waiting to make serious goal threats when trades occur.

A straightforward line is:
35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w
36b Cg3w Cf3x dg2n rb5e rc5s
37w rc4s rc3x Dd4w Dc4e Rc1e
after which we can advance our g6 rabbit and wreak havoc. He can choose to instead swarm the c6 trap on move 36w, but in the lines I've looked at we retain our material advantage without any obvious positional catastrophes.

Title: Re: Move 35
Post by 99of9 on Jun 22nd, 2008, 2:51am

on 06/22/08 at 01:32:17, warren wrote:
99's move was counterintuitive to me at first, but now I think it leaves us with a strong position and is my current favorite move.

I'm glad I shared it.  In reply to Soter's 36w I tried an even sharper line:
36b Cg3w Cf3x dg2n rg6s rg5s
but I couldn't find a way to force goal if he killed the horse.


Quote:
A straightforward line is:
35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w
36b Cg3w Cf3x dg2n rb5e rc5s
37w rc4s rc3x Dd4w Dc4e Rc1e

I love that you're using a bot tactic here (sending the b5 rabbit to it's death to protect our horse: "lemmings").

Another option for him is
36w Eb3e Ec3e dc4s dc3x Ed3e
36b Cg3n dg2n Rc2n Rc3x hb2e
but I think that still leaves us with quite a good position since we have all the threats.

Title: Re: Move 35
Post by warren on Jun 22nd, 2008, 3:48am

on 06/22/08 at 02:51:45, 99of9 wrote:
I love that you're using a bot tactic here (sending the b5 rabbit to it's death to protect our horse: "lemmings").

I used Bomb extensively in my analysis, so it's not surprising that I would use a bot tactic. In this case it seems to be a legitimate move rather than just delaying the inevitable. If you do 36b Cg3w Cf3x dg2n rg6s rg5s instead, you're one crucial step away from having a goal threat so it doesn't work out.

Title: Re: Move 35
Post by 99of9 on Jun 22nd, 2008, 4:52am

on 06/22/08 at 03:48:48, warren wrote:
I used Bomb extensively in my analysis, so it's not surprising that I would use a bot tactic. In this case it seems to be a legitimate move rather than just delaying the inevitable.

Yes, it's definitely legitimate, and seems to help us quite a bit.


Quote:
If you do 36b Cg3w Cf3x dg2n rg6s rg5s instead, you're one crucial step away from having a goal threat so it doesn't work out.

Well, we don't have a 4 step goal, but I was hoping we would have a 4 step way of forcing goal on a later move, but I couldn't make it work.

Title: Re: Move 35
Post by jdb on Jun 22nd, 2008, 5:25am

on 06/21/08 at 01:59:43, 99of9 wrote:
how about
35b ed3e ee3s Hf2s ee2e

This might be our chance to take over the other trap, but it's a sharp line, and we'd need to play out the exchange sequence.


This move looks like it has potential.

What happens if gold plays a move to protect the f3 trap? The threat to silver's hostage horse is not going to go away, so if gold can slow down silver's goal attack enough, he would have a good position. I don't think defending the f3 trap will work for gold, but it needs to be looked at.

35b ed3e ee3s Hf2s ee2e
36w Rc1->e3

Title: Re: Move 35
Post by 99of9 on Jun 22nd, 2008, 5:51am

on 06/22/08 at 05:25:50, jdb wrote:
35b ed3e ee3s Hf2s ee2e
36w Rc1->e3

Good question, that does slow us down quite a bit.

Title: Re: Move 35
Post by warren on Jun 22nd, 2008, 6:48am

on 06/22/08 at 05:25:50, jdb wrote:
This move looks like it has potential.

What happens if gold plays a move to protect the f3 trap? The threat to silver's hostage horse is not going to go away, so if gold can slow down silver's goal attack enough, he would have a good position. I don't think defending the f3 trap will work for gold, but it needs to be looked at.

35b ed3e ee3s Hf2s ee2e
36w Rc1->e3

That 36w seems to quite effectively defuse our attack. I'm returning to thinking that we'll lose our dog for his choice of a rabbit or letting our horse go free.

Title: Re: Move 35
Post by Fritzlein on Jun 22nd, 2008, 8:04am

on 06/22/08 at 02:51:45, 99of9 wrote:
Another option for him is
36w Eb3e Ec3e dc4s dc3x Ed3e
36b Cg3n dg2n Rc2n Rc3x hb2e
but I think that still leaves us with quite a good position since we have all the threats.

This is an important point I had missed.   We are very willing to give up our dog in exchange for a rabbit and a freed horse.  CRR for D looks great on this board.  That means that we don't have to retreat our dog from c4 if our elephant leaves.  We can play 35s Ed3e Ee3s xxxx xxxx.  Not pushing the horse right away gives us more flexibility, and maybe those two extra steps can be of some use.  My first contribution will be to see if I can make something of that.

Title: Re: Move 35
Post by Fritzlein on Jun 22nd, 2008, 8:20am

on 06/22/08 at 06:48:58, warren wrote:
That 36w seems to quite effectively defuse our attack. I'm returning to thinking that we'll lose our dog for his choice of a rabbit or letting our horse go free.

I'm not sure the move 35b dc4s dc3n Rc2n xxxx guarantees us any compensation for our dog.  Can't he just take our dog, and then if we flip his rabbit, move Da3->b6?

Title: Re: Move 35
Post by Soter on Jun 22nd, 2008, 8:58am
With regard to any 35b which contains ed3e ee3s without pushing the horse to f1 - what if Jean answers with something like Rg1w Rd2n Rd3e Da3n? In this case his rabbit stands on f1 and f3 tap gets a third guardian; isn't it dangerous for Silver? Would flipping his H to d2 be a good idea?

I think we should take time with this move and not decide too quickly - this may be one of the pivotal moves of the game.

Title: Re: Move 35
Post by Fritzlein on Jun 22nd, 2008, 9:48am

on 06/22/08 at 08:58:10, Soter wrote:
With regard to any 35b which contains ed3e ee3s without pushing the horse to f1 - what if Jean answers with something like Rg1w Rd2n Rd3e Da3n?

You are quite right.  The rabbit on f1 prevents us from pushing the horse back later, and I can't make any lines work for us without that horse push.  I consider my suggestion refuted.  However,


on 06/22/08 at 05:25:50, jdb wrote:
35b ed3e ee3s Hf2s ee2e
36w Rc1->e3

I'm not sure that JDB's defense actually refutes 99of9's attack.  Recall that we are quite happy to win R for D with our horse being freed, and every bit as happy to win HR for HD, since with horses off our CRR for D will weigh even more heavily.  Consider

35b ed3e ee3s Hf2s ee2e
36w Rc1e Rd1e Re1n Re2n
36b Cg3e dg2n rg6s rg5s
37w Eb3e Ec3w dc4s dc3x Dd4e
37b ef2n ef3s Re3e Rf3x rg4w

And now I can't find a move for Gold that is better than a horse trade.  For example

38w Hf1w Eb3n hb2n Rd2n
38b ce7s rf7s rf6s rd5s
39w hb3e hc3x Eb4s Rg1w Rd3e
39b ef2w Re3e Rf3x ee2n ce6s

with a dominating position for us.  I consider 99of9's move rehabilitated, and possibly even winning.  If we can't find a re-refutation, then it may turn out that chessandgo's fourth step on 35w (allowing this attack) was a fatal blunder!

(My apologies that my evaluation of the position keeps swinging wildly.  That tends to happen in sharp positions.)

(This time I'm glad I analyzed without Bomb.  Our strategic attack takes too long to pay out for Bomb to understand it.  Another example of a position that computers can't play even though it is wild and tactical.  Not that I won't use Bomb to double-check my analysis!)


Title: Re: Move 35
Post by warren on Jun 22nd, 2008, 10:04am

on 06/22/08 at 08:20:33, Fritzlein wrote:
I'm not sure the move 35b dc4s dc3n Rc2n xxxx guarantees us any compensation for our dog.  Can't he just take our dog, and then  if we flip his rabbit, move Da3->b6?

You're right, we hopefully have a better 35b then that. Does 35b Rd2e ed3s Rc2n ed2w do a better job of saving our horse?

Title: Re: Move 35
Post by warren on Jun 22nd, 2008, 10:25am

on 06/22/08 at 09:48:47, Fritzlein wrote:
I'm not sure that JDB's defense actually refutes 99of9's attack.

You seem to be right. I therefore currently tentatively support 99's 35b.

I'm having trouble remembering all the lines, so I think I'll make another analysis tree now.

Title: Re: Move 35
Post by Fritzlein on Jun 22nd, 2008, 10:31am

on 06/22/08 at 10:04:56, warren wrote:
You're right, we hopefully have a better 35b then that. Does 35b Rd2e ed3s Rc2n ed2w do a better job of saving our horse?

We might save our horse, but it looks like an awfully slow plan, assuming we're trying to get the horse all the way to freedom and not just one square away where it can be retaken hostage as soon as our elephant leaves.  Partially freeing the horse only buys us time, and the amount of time it buys us is less than our investment of time.  Without actually calculating it out, I'm confident chessandgo could hurt us before our horse was truly freed.  

Title: Re: Move 35
Post by Fritzlein on Jun 22nd, 2008, 10:33am

on 06/22/08 at 10:25:35, warren wrote:
I think I'll make another analysis tree now.

Thank you!  Especially for a sharp move that can get refuted and rehabilitated many times, a tree is very helpful.

Title: Tree
Post by warren on Jun 22nd, 2008, 10:38am
Different moves for us (35b) are separated by =.

===============================================
Various replies to:
35b ed3e ee3s Hf2s ee2e (99).

36w Rc1e Rd1e Re1n Re2n (jdb)
36b Cg3e dg2n rg6s rg5s (F)
.37w Dd4e De4e Hf1w Rg1n (99)
.37w Eb3e Ec3w dc4s dc3x Dd4e (F)
. 37b ef2n ef3s Re3e Rf3x rg4w
. .38w Hf1w Eb3n hb2n Rd2n
. . 38b ce7s rf7s rf6s rd5s
. . .39w hb3e hc3x Eb4s Rg1w Rd3e
. . . 39b ef2w Re3e Rf3x ee2n ce6s


36w Eb3e Ec3w dc4s dc3x Dd4w (99/gnobot)
36b Cg3w Cf3x dg2n rg6s rg5s

36w Da3n Da4n Da5n Da6e (soter)
36b Cg3w Cf3x dg2n dc4n rd5e (warren)
.37w Eb3e hb2n Ec3n hb3e hc3x
. 37b ef2n Hf1n ef3n Hf2n Hf3x
. .38w Ec4s dc5s Ec3e dc4s dc3x

36w Eb3e Ec3w dc4s dc3x Hf1w (warren)
36b Cg3w Cf3x dg2n rb5e rc5s
.37w rc4s rc3x Dd4w Eb3e hb2n (99)
. 37b dg3s Rg1w dg2s rd5s (soter, jdb)
. .38w rd4n Dc4e Dd4e De4e (F)
. . 38b rd5s rd4s rh7s rh6s
. . .39w hb3s Ec3w Df4e Dg4e
. . . 39b rg6s rg5s rg8s rg7s
. . . .40w Eb3e Ec3w rd3w rc3x Rd2e
. . . . 40b rg6s rg4s rg5s rg3s
. . .39w hb3s Ec3w Df4e Da3n (F)
. . . 39b rh5s rg6e rh6s rh4s
. . . .40w Eb3e Ec3w rd3w rc3x Dg4e
. . . . 40b ef2e eg2n rh3s rh2w
. 37b dg3s Rg1w dg2s rg6s (Soter)
. .38w Dc4e Dd4e De4e Df4e (99)
. . 38b rd5w rc5w rb5s hb3s (RW)
. 37b e->d3, ???? (warren)
. 37b rd5e re5s re4s re3s (99)
. .38w Ec3e hb3ex Ed3e Rg1w (bomb/warren)
. . 38b rh7s rh6s rh5s rh4s (99)
. . .39w Rf1e He1e Rc1e Rd1e (bomb/warren)
. . . 39b dg3s dg2n Rg1n rh3s
. . . .40w Ee3w re2n Rd2e Hf1e
. . . . 40b Re2w ef2w re3e rf3s
. . . . .41w Rb1e Rc1e Re1e Rd1e
. . . . . 41b ee2n Re1n rc8w cc7w
. . . . . .42w Rf1w Hg1w rf2n Hf1n
. . . . 40b ef2s Rg2w Rf2n Rf3x ef1n (99)
. . . .40w Dc4e Dd4e De4e Ee3e (99/bomb)
. . . . 40b rg6s rg5s dg3e Rg2n (jdb)
.37w rc4s rc3x Dd4w Dc4e Rc1e (warren)
. 37b rg6s rg5s dg3e rg4s
. .38w He1e Rg1n Rd1e Hf1e
. 37b rh7s rh6s rh5s rh4s (F)
. .38w Rg1n He1e Rd1e Hf1e (F)
. . 38b ef2n Rg2w ef3w Rf2n Rf3x (F)
 37w Eb3e Ec3w rc4s rc3x Rc1e (equivalent to above 37w)
 37w Da3n Da4e rc4s rc3x Db4e (F)
. 37b rg6s rg5s dg3e rg4s (bomb/warren)
. .38w Eb3e Ec3e Ed3e Hf1e (bomb/warren)
. 37b rh7->h3 (jdb)
. .38w PASS
. . 38b dg3s Rg1w dg2s rh3s
. 37b ef2w ee2n ee3w dg3n (F)
. 37b rf7s rf6s rf5s rf4s
. .38w He1e Rd2e Dc4n Dd4e
. . 38b dg3n dg4w rf3e rg6s
. . .39w De4s Eb3e hb2n Ra2n
. . . 39b ef2n ef3s De3e Df3x rg5s
. . . .40w Ec3e hb3e hc3x Ed3e Ra3n
36b Cg3w Cf3x dg2n rg6s rg5s (warren, bad move)

36w Eb3e Ec3e dc4s dc3x Ed3e (99)
36b Cg3n dg2n Rc2n Rc3x hb2e

36w H->e3 Rd2n (99)
36b rf7s rf6s rf5s rf4s (mistre)
.37w rf3n He3e rf4e Hf3n
.37w Dd4e De4e Df4w rf3n
36b ef2w ee2w Rd3w ed2n  (mistre)
.37w Rg1w Rf1n Da3n Da4n
. 37b Cg3e dg2n rg6s rg5s (w)
.37w Rg1w Rf1n Rc1e Rd1n (w)
. 37b Cg3n dg2n dg3e dh3n
. .38w Da3n Da4e rb5w Db4n
. . 38b rh7s rh6s ce7w cd7s
. . .39w ra5s Db5w Rd2e Rh1n
. . . 39b Cg4n dh4w Cg5w dg4n
. . . .40w He3n He4e Cf5w Hf4e
. . . . 40b dg5w rg6s df5n df6w
. . . . .41w rg5w Hg4n Rh2n Rh3n
. . . . . 41b de6e Ce5n df6e Ce6e Cf6x

36w Hf1w He1n He2n Dd4e (JDB)
36b rd5s rd4s Cg3n dg2n (W)
37w Da3n rd3w He3w De4w
37b Cg4w dg3n Cf4s Cf3x dg4w
38w Eb3n Eb4s dc4w rc3x Dd4w
38b rg6s rg5s rg4s ef2w

36w Rc2n Rc3e Rd3e Dd4s (warren)
36b Cg3e dg2n rg6s rg5s

====================

35b dc4s dc3n Rc2n xxxx (warren/bomb)
36w take dog
.36b flip rabbit to c5
. 37w  Da3->b6 (F)

====================

35b ed3e ee3s dc4n rd5e (RW)
36w Rg1w Dd4e De4s Da3n (RW + mistre's correction)
.36b Hf2n ee2w Cg3w dg2n
36w Rg1w Re2n Re3w xxxx
.36b ee2s Hf2e He2e ee1n
36w dg2s Hf2w xxxx xxxx
36w Rg1w dg2e Hf2e xxxx (mistre)
36w Rd2n Rd3e Dd4e Rg1w (F)
.36b ee2w Hf2w ed2n dg2s (F; but was "dg2x")
.36b ee2s Hf2w He2w ee1n
. 37w Eb3e hb2n Ec3e hb3e hc3x
. .37b ee2e Re3s Cg3w Cf3x dg2n
. . 38w De4e Ed3e Ee3e Da3e

==========================

35b ed3e ee3s dc4w db4w (mistre)
36w Dd4s Dd3e Rd2n Da3n (soter)
.36b Hf2s ee2e Cg3e dg2n
. 37w De3n Hf1w Rh1n Ch3n

I will update this post as more moves are discussed.

Title: Re: Move 35
Post by Soter on Jun 22nd, 2008, 10:50am

Quote:
(My apologies that my evaluation of the position keeps swinging wildly.  That tends to happen in sharp positions.)

No problem at all  :) ; I think such phenomenon is typical for deep games like Arimaa.
I'm gonna have less work very soon, and I'll try to contribute more. For now I favour 99's proposal.



Title: Re: Tree for 99's 35b
Post by Fritzlein on Jun 22nd, 2008, 11:28am

on 06/22/08 at 10:38:55, warren wrote:
35b ed3e ee3s Hf2s ee2e (99).
36w Eb3e Ec3w dc4s dc3x Hf1w (warren)
36b Cg3w Cf3x dg2n rb5e rc5s

Hmmm, I had only looked at JDB's defense, and not this line at all.  Things don't look as good for us here.  After 37w Da3n Da4e rc4s rc3x Db4e we have only a slight material lead of CC for D, which is counterbalanced by our horse still being hostage.  Given chessandgo's pair of active dogs, if we cover up with 37b ef2w ee2n ee3w dg3n I think in the long run we would have to be happy getting a dog for our hostage horse.  If we play for attack rather than control, it seems to work out the same:

37b rf7s rf6s rf5s rf4s
38w He1e Rd2e Dc4n Dd4e
38b dg3n dg4w rf3e rg6s
39w De4s Eb3e hb2n Ra2n
39b ef2n ef3s De3e Df3x rg5s
40w Ec3e hb3e hc3x Ed3e Ra3n

Again we have CC for H, and it isn't clear who is ahead.  If we can't do better in this line (e.g. on 36s) then 99of9's move no longer clearly outpaces the alternatives in my mind.

Title: Re: Tree for 99's 35b
Post by jdb on Jun 22nd, 2008, 12:29pm

on 06/22/08 at 11:28:23, Fritzlein wrote:
Hmmm, I had only looked at JDB's defense, and not this line at all.  Things don't look as good for us here.  After 37w Da3n Da4e rc4s rc3x Db4e we have only a slight material lead of CC for D, which is counterbalanced by our horse still being hostage.  Given chessandgo's pair of active dogs, if we cover up with 37b ef2w ee2n ee3w dg3n I think in the long run we would have to be happy getting a dog for our hostage horse.  If we play for attack rather than control, it seems to work out the same:

37b rf7s rf6s rf5s rf4s
38w He1e Rd2e Dc4n Dd4e
38b dg3n dg4w rf3e rg6s
39w De4s Eb3e hb2n Ra2n
39b ef2n ef3s De3e Df3x rg5s
40w Ec3e hb3e hc3x Ed3e Ra3n

Again we have CC for H, and it isn't clear who is ahead.  If we can't do better in this line (e.g. on 36s) then 99of9's move no longer clearly outpaces the alternatives in my mind.


35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w
36b Cg3w Cf3x dg2n rb5e rc5s
37w Da3n Da4e rc4s rc3x Db4e

37b rh7->h3
38w PASS
39b dg3s Rg1w dg2s rh3s

If gold waits until 38w to defend, it looks like he loses material to stop the goal. On move 37w, Gold can capture the rabbit in three steps and use the other step to defend. This may be enough to enable a defence on 38w, but I am not sure.


Title: Re: Tree for 99's 35b
Post by warren on Jun 22nd, 2008, 2:00pm

on 06/22/08 at 11:28:23, Fritzlein wrote:
Hmmm, I had only looked at JDB's defense, and not this line at all.  Things don't look as good for us here.  After 37w Da3n Da4e rc4s rc3x Db4e we have only a slight material lead of CC for D, which is counterbalanced by our horse still being hostage.  Given chessandgo's pair of active dogs, if we cover up with 37b ef2w ee2n ee3w dg3n I think in the long run we would have to be happy getting a dog for our hostage horse.  If we play for attack rather than control, it seems to work out the same:

37b rf7s rf6s rf5s rf4s
38w He1e Rd2e Dc4n Dd4e
38b dg3n dg4w rf3e rg6s
39w De4s Eb3e hb2n Ra2n
39b ef2n ef3s De3e Df3x rg5s
40w Ec3e hb3e hc3x Ed3e Ra3n

Again we have CC for H, and it isn't clear who is ahead.  If we can't do better in this line (e.g. on 36s) then 99of9's move no longer clearly outpaces the alternatives in my mind.

After
35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w
36b Cg3w Cf3x dg2n rb5e rc5s
37w Da3n Da4e rc4s rc3x Db4e

how about
37b rg6s rg5s dg3e rg4s (bomb/warren)
which seems to pretty much force
38w Eb3e Ec3e Ed3e Hf1e (bomb/warren)
freeing our horse.

Update: by the way, Bomb suggested this move after a slightly different variation (different way of capturing c4 rabbit), but it works here too.

Title: Re: Tree for 99's 35b
Post by warren on Jun 22nd, 2008, 2:38pm

on 06/22/08 at 12:29:07, jdb wrote:
If gold waits until 38w to defend, it looks like he loses material to stop the goal. On move 37w, Gold can capture the rabbit in three steps and use the other step to defend. This may be enough to enable a defence on 38w, but I am not sure.

Yeah, doing a Rc1e at the end of move 37w allows Gold a stronger defense:
35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w
36b Cg3w Cf3x dg2n rb5e rc5s
37w Eb3e Ec3w rc4s rc3x Rc1e
37b rg6s rg5s dg3e rg4s
38w He1e Rg1n Rd1e Hf1e

Title: Re: Tree for 99's 35b
Post by Fritzlein on Jun 22nd, 2008, 2:57pm

on 06/22/08 at 14:00:24, warren wrote:
After
35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w
36b Cg3w Cf3x dg2n rb5e rc5s
37w Da3n Da4e rc4s rc3x Db4e

how about
37b rg6s rg5s dg3e rg4s (bomb/warren)
which seems to pretty much force
38w Eb3e Ec3e Ed3e Hf1e (bomb/warren)
freeing our horse.

Oh, yeah, that's a much better attacking move than I came up with.  I feel better about that line now.

Title: Re: Tree for 99's 35b
Post by Fritzlein on Jun 22nd, 2008, 3:36pm

on 06/22/08 at 12:29:07, jdb wrote:
37b rh7->h3
38w PASS
39b dg3s Rg1w dg2s rh3s

Very nice attacking idea.  Yes, Gold can take the Rc1e step on 37w and have a goal defense that loses a rabbit instead of a horse, but Gold also loses the activation of his horse and a lot of time.  After

35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w
36b rb5e rc5s Cg3w Cf3x dg2n
37w Eb3e Ec3w rc4s rc3x Rc1e
37b rh7s rh6s rh5s rh4s
38w Rg1n He1e Rd1e Hf1e
38b ef2n Rg2w ef3w Rf2n Rf3x,

Gold can't take our horse, we have material advantage of CRR for D, and most importantly we should be able to get enough of a swarm in the east to hold f3 with little pieces if necessary while our elephant guards c3.  So I am back to liking 99of9's move (again).

Title: Re: Move 35
Post by warren on Jun 23rd, 2008, 12:42am
Here's an excerpt of the tree after
35b ed3e ee3s Hf2s ee2e
that includes only our best moves (in my opinion; say something if you disagree) but all of his moves. Let's try to attack this plan by either:
  • finding a move for gold better than those in this tree or
  • showing that one of the leaves of this tree is bad for us.


36w Rc1e Rd1e Re1n Re2n
36b Cg3e dg2n rg6s rg5s
37w Eb3e Ec3w dc4s dc3x Dd4e
37b ef2n ef3s Re3e Rf3x rg4w
38w Hf1w Eb3n hb2n Rd2n
38b ce7s rf7s rf6s rd5s
39w hb3e hc3x Eb4s Rg1w Rd3e
39b ef2w Re3e Rf3x ee2n ce6s

36w Eb3e Ec3w dc4s dc3x Hf1w
36b Cg3w Cf3x dg2n rb5e rc5s
 37w rc4s rc3x Dd4w Dc4e Rc1e or Eb3e Ec3w rc4s rc3x Rc1e
. 37b rh7s rh6s rh5s rh4s
. .38w Rg1n He1e Rd1e Hf1e
. . 38b ef2n Rg2w ef3w Rf2n Rf3x
 37w Da3n Da4e rc4s rc3x Db4e
. 37b rg6s rg5s dg3e rg4s
. .38w Eb3e Ec3e Ed3e Hf1e

36w Eb3e Ec3e dc4s dc3x Ed3e
36b Cg3n dg2n Rc2n Rc3x hb2e

Title: Re: Move 35
Post by 99of9 on Jun 23rd, 2008, 1:03am

on 06/22/08 at 09:48:47, Fritzlein wrote:
35b ed3e ee3s Hf2s ee2e
36w Rc1e Rd1e Re1n Re2n
36b Cg3e dg2n rg6s rg5s
37w Eb3e Ec3w dc4s dc3x Dd4e
37b ef2n ef3s Re3e Rf3x rg4w


We need to think through what happens if he plays more defensive moves before killing our pieces.  Something like...
37w Dd4e De4e Hf1w Rg1n

Title: Re: Move 35
Post by Fritzlein on Jun 23rd, 2008, 5:41am

on 06/23/08 at 01:03:50, 99of9 wrote:
We need to think through what happens if he plays more defensive moves before killing our pieces.  Something like...
37w Dd4e De4e Hf1w Rg1n

If his dog joins the defense of f3, our d5 rabbit can temporarily secure c3 from d3.  We would then have a threat of our elephant pushing the defensive piece on f4 to f5 for capture in f6, as well as remote goal threats both center and east, which makes it difficult for his horse to know where to defend.

Title: Re: Move 35
Post by warren on Jun 23rd, 2008, 12:28pm
I wonder where RonWeasley is? A fresh pair of eyes would be nice, and also in a few days we'll need him to move...

Title: Re: Move 35
Post by mistre on Jun 23rd, 2008, 5:04pm
All lines that I tried had me liking 35b ed3e ee3s Hf2s ee2e.

The best I could do for Chessandgo is to trade Dogs and Horses and leave us with a strong goal threat.

The reason it is such a good move strategically is that it immediately puts our east Dog in a strong position from a very weak one and it gives us solid control over the c6 trap. Finally it opens us up to advance rabbits on the east side.

I am ready to cast my vote for this move.  I have no doubt that Chessandgo will find some devious move that we haven't seen yet - so I guess we should keep looking.  I can't find one.



Title: Re: Move 35
Post by 99of9 on Jun 23rd, 2008, 7:01pm
Gnobot is worried (at 13 ply after our proposed move) about:

35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Dd4w

Is our subsequent threat sufficient to get either a horse trade or a mobile horse after:
36b Cg3w Cf3x dg2n rg6s rg5s

I think we're fine, but it's probably worth someone double checking.

Title: Re: Move 35
Post by Janzert on Jun 23rd, 2008, 9:44pm

on 06/23/08 at 19:01:51, 99of9 wrote:
Gnobot is worried (at 13 ply after our proposed move) about:

35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Dd4w

Is our subsequent threat sufficient to get either a horse trade or a mobile horse after:
36b Cg3w Cf3x dg2n rg6s rg5s

I think we're fine, but it's probably worth someone double checking.


After 35b ed3e ee3s Hf2s ee2e OpFor has:

depth 15
time 1547
score -69
pv Eb3e Ec3w dc4s dc3x Hf1w b Cg3w Cf3x dg2n rb5e rc5s w rc4s rc3x Dd4w Eb3n hb2n b ef2w ee2n ee3w

Janzert

p.s. A long run against the current position didn't seem to result in anything useful but here it is just for fun.

depth 14
time 36400
score 163
pv ed3e rg6s ee3s dc4w w Eb3e Ec3w Dd4s Dd3e b rd5s rc8w rd4s db4e w Rg1w

Title: Re: Move 35
Post by 99of9 on Jun 23rd, 2008, 11:05pm
At 15 ply (after our proposed move) Gnobby has changed its mind and now favours Soter's original reply (kill dog and move horse sideways).  Since we're analysing that one already, this is not big news.

Title: Re: Move 35
Post by 99of9 on Jun 23rd, 2008, 11:23pm

on 06/23/08 at 00:42:06, warren wrote:
35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w
36b Cg3w Cf3x dg2n rb5e rc5s

37w rc4s rc3x Dd4w Eb3e hb2n
NOW WHAT DO WE DO??
I think we are in grave danger again, because I can't see a way to set up a goal threat fast enough.

[Edit] Perhaps this is good enough:
37b rd5e re5s re4s re3s


Title: Re: Move 35
Post by warren on Jun 24th, 2008, 12:32am

on 06/23/08 at 23:23:01, 99of9 wrote:
37w rc4s rc3x Dd4w Eb3e hb2n
NOW WHAT DO WE DO??
I think we are in grave danger again, because I can't see a way to set up a goal threat fast enough.

[Edit] Perhaps this is good enough:
37b rd5e re5s re4s re3s


After your 37b, bomb suggests
38w Ec3e hb3ex Ed3e Rg1w
Playing it forward a few turns, it seems gold survives and retains an immense material advantage.

How about instead for 37b just moving our elephant to d3 to protect our horse? It's not a wonderful position for us, since we have two cats and a hostage horse against his dog and horse, but it seems playable. This needs more thought, as this is the most likely line I think.

Title: Re: Move 35
Post by 99of9 on Jun 24th, 2008, 1:01am

on 06/24/08 at 00:32:52, warren wrote:
After your 37b, bomb suggests
38w Ec3e hb3ex Ed3e Rg1w

Then I would answer:
38b rh7s rh6s rh5s rh4s
and now we have two strong goal attacks.

Title: Re: Move 35
Post by jdb on Jun 24th, 2008, 1:33am

on 06/24/08 at 01:01:45, 99of9 wrote:
Then I would answer:
38b rh7s rh6s rh5s rh4s
and now we have two strong goal attacks.


This line is critical. Silver is behind in material. If there is no forced goal, this is deep trouble.

Title: Re: Move 35
Post by warren on Jun 24th, 2008, 1:38am

on 06/24/08 at 01:01:45, 99of9 wrote:
Then I would answer:
38b rh7s rh6s rh5s rh4s
and now we have two strong goal attacks.


We have two goal attacks but I don't think that's enough to make up for being lost materially since he appears to be able to prevent us from actually scoring. For example:

35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w
36b Cg3w Cf3x dg2n rb5e rc5s
37w rc4s rc3x Dd4w Eb3e hb2n
37b rd5e re5s re4s re3s
38w Ec3e hb3e hc3x Ed3e Rg1w

38b rh7s rh6s rh5s rh4s
39w Rf1e He1e Rc1e Rd1e
39b dg3s dg2n Rg1n rh3s
40w Ee3w re2n Rd2e Hf1e
40b Re2w ef2w re3e rf3s
41w Rb1e Rc1e Re1e Rd1e
41b ee2n Re1n rc8w cc7w
42w Rf1w Hg1w rf2n Hf1n

Title: Re: Move 35
Post by 99of9 on Jun 24th, 2008, 2:59am

on 06/24/08 at 01:38:34, warren wrote:
We have two goal attacks but I don't think that's enough to make up for being lost materially since he appears to be able to prevent us from actually scoring.

Following your query, I've played it out against bomb (30s = 11-13 ply) a few times, and won them all, but sometimes it's a long sequence of moves, and I can't be sure bomb is playing optimally.

I suggest that a few others play around with these goal threats.  This could be a crucial line.


Quote:
35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w
36b Cg3w Cf3x dg2n rb5e rc5s
37w rc4s rc3x Dd4w Eb3e hb2n
37b rd5e re5s re4s re3s
38w Ec3e hb3e hc3x Ed3e Rg1w
38b rh7s rh6s rh5s rh4s
39w Rf1e He1e Rc1e Rd1e
39b dg3s dg2n Rg1n rh3s
40w Ee3w re2n Rd2e Hf1e


I would reply
40b ef2s Rg2w Rf2n Rf3x ef1n
gaining material and maintaining a goal threat.

[Edit: I can't believe I'm arguing about what might happen 44 steps ahead!! This is incredible (but important).]

Title: Re: Move 35
Post by 99of9 on Jun 24th, 2008, 3:56am

on 06/24/08 at 01:38:34, warren wrote:
35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w
36b Cg3w Cf3x dg2n rb5e rc5s
37w rc4s rc3x Dd4w Eb3e hb2n
37b rd5e re5s re4s re3s
38w Ec3e hb3e hc3x Ed3e Rg1w

38b rh7s rh6s rh5s rh4s
39w Rf1e He1e Rc1e Rd1e
39b dg3s dg2n Rg1n rh3s
40w Ee3w re2n Rd2e Hf1e
...


I played bomb with longer time controls (3-4 min) and lost one.  Bomb deviated from Warren's line here:
40w Dc4e Dd4e De4e Ee3e

can anyone else play from here and win, or do we have to deviate earlier?

Title: Re: Move 35
Post by RonWeasley on Jun 24th, 2008, 4:02am

on 06/23/08 at 12:28:36, warren wrote:
I wonder where RonWeasley is? A fresh pair of eyes would be nice, and also in a few days we'll need him to move...


I'm here.  It's just that I didn't log in this weekend and I'm way behind on this discussion.

Title: Re: Move 35
Post by Soter on Jun 24th, 2008, 4:36am
I can't play it against Bomb now and I think even these two potential goal threats may eventually get blocked, leaving us with a weakened position. There is one thing, however, that may be interesting:

35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w  
36b Cg3w Cf3x dg2n rb5e rc5s
37w rc4s rc3x Dd4w Eb3e hb2n
37b dg3s Rg1w dg2s rg6s

What do you think about this 37b? Isn't it a slightly tougher nut to crack for Jean? IMO it's rather hard to kill our H and then rush to save the f3 quadrant...

I swear I'd help you guys if I wasn't busy. After the coming Saturday this game will be one of my priorities.

Title: Re: Move 35
Post by jdb on Jun 24th, 2008, 4:43am

on 06/24/08 at 03:56:06, 99of9 wrote:
I played bomb with longer time controls (3-4 min) and lost one.  Bomb deviated from Warren's line here:
40w Dc4e Dd4e De4e Ee3e

can anyone else play from here and win, or do we have to deviate earlier?


40b rg6s rg5s dg3e Rg2n  

Title: Re: Move 35
Post by 99of9 on Jun 24th, 2008, 4:48am

on 06/24/08 at 04:36:33, Soter wrote:
35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w  
36b Cg3w Cf3x dg2n rb5e rc5s
37w rc4s rc3x Dd4w Eb3e hb2n
37b dg3s Rg1w dg2s rg6s

This is a good option to look at.  I think it would continue:
38w Dc4e Dd4e De4e Df4e


Quote:
I swear I'd help you guys if I wasn't busy. After the coming Saturday this game will be one of my priorities.

You've already helped a fair bit with this move, and this post is more good help.  I suspect we will still be on this move this weekend.

Title: Re: Move 35
Post by Soter on Jun 24th, 2008, 4:56am

Quote:
This is a good option to look at.  I think it would continue:
38w Dc4e Dd4e De4e Df4e


Yet another fresh idea for 37b ( but this one might be too risky )

37b dg3s Rg1w dg2s rd5s

Now our rabbit serves as a kind of obstacle; gold dog needs an additional step when moving east and won't be able to reach g4. Haven't checked what happens later tho.



Title: Re: Move 35
Post by RonWeasley on Jun 24th, 2008, 8:59am

on 06/24/08 at 04:48:09, 99of9 wrote:
This is a good option to look at.  I think it would continue:
38w Dc4e Dd4e De4e Df4e


I tried 38b rd5w rc5w rb5s hb3s and lost both h and r but caputed the D (on g4) and had goal pressure on the east side.  Silver had dcc for gold's HDR.  For a depleted board, that might be about even, although silver has a 4-3 advantage in rabbit pushers.

Title: Re: Move 35
Post by jdb on Jun 24th, 2008, 9:20am

on 06/24/08 at 04:36:33, Soter wrote:
35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w  
36b Cg3w Cf3x dg2n rb5e rc5s
37w rc4s rc3x Dd4w Eb3e hb2n
37b dg3s Rg1w dg2s rg6s


The idea of blocking golds access to the g and h files with the dog is great. With a slight modification, I think we get a winning attack.
37b dg3s Rg1w dg2s rd5s

Now the dog on c5 can't stop the h7 rabbit from running up the board. With the wall set up by the dog and elephant, that should be enough for a goal.

Title: Re: Move 35
Post by RonWeasley on Jun 24th, 2008, 9:32am

on 06/24/08 at 04:56:48, Soter wrote:
Yet another fresh idea for 37b ( but this one might be too risky )

37b dg3s Rg1w dg2s rd5s

Now our rabbit serves as a kind of obstacle; gold dog needs an additional step when moving east and won't be able to reach g4. Haven't checked what happens later tho.



The immediate threat is 38b rg6s rg5s rg4s rg3s with goal next turn.

Gold must put the D on g4.  This frees the r on d4 to cover c3 if gold pulls up the h.  Otherwise, our e can stick to the D and silver can swarm rabbits on the east side.  Probably trade h for D, but that give silver a clean east side to make threats on.

From our original position, I'm OK trading hd for DC and even losing a rabbit to end up with dcc for HD.  Being one ahead in the piece count is helpful in a depleted endgame.

Title: Re: Move 35
Post by Soter on Jun 24th, 2008, 9:56am
Yeah, I liked it this 37b instantly when I saw it. But we need to do some heavy analysis of possible deviations; e.g. what if it the elephant eats the rabbit and then dog moves east one step?

Now a different kettle of fish: assuming that the mob chooses 35b ed3e ee3s Hf2s ee2e, is something like 36w Da3n Da4n Da5n Da6e even remotely playable for chessandgo? Yes, I know this one looks rather wacky, but don't forget about Jean's penchant for surprises...

Title: Re: Move 35
Post by mistre on Jun 24th, 2008, 10:12am
It appears that the mob is venturing off into over-analyzing the apparent favorite move into the future.  How sure can we be that Chessandgo plays our expected 36w?  Jean surprises us more often than not.  If not, this seems like a waste of time as we all know the tree can change drastically based on one move.

What we should be asking is if there is a better 35b than ed3e ee3s Hf2s ee2e?, and at this point no one is providing alternatives.


Title: Re: Move 35
Post by warren on Jun 24th, 2008, 11:21am

on 06/24/08 at 10:12:28, mistre wrote:
It appears that the mob is venturing off into over-analyzing the apparent favorite move into the future.  How sure can we be that Chessandgo plays our expected 36w?  Jean surprises us more often than not.  If not, this seems like a waste of time as we all know the tree can change drastically based on one move.

What we should be asking is if there is a better 35b than ed3e ee3s Hf2s ee2e?, and at this point no one is providing alternatives.

I agree. After
35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w
36b Cg3w Cf3x dg2n rb5e rc5s
37w rc4s rc3x Dd4w Eb3e hb2n
37b elephant to d3
we are in a roughly even position. There's no point in analyzing the rabbit-pushing horse-sacrifice 37b further until someone points out another 35b that looks like it might be better.

Analyzing lines can be helpful as a way of exploring the strategic landscape, but I think mistre has a point that we're overdoing it a bit and more attention to earlier moves is warranted.

Finding a better 35b is hard. To get things started, let's try to figure out what the second-best 35b is. Are there any other 35bs that do better than losing the dog in return for freeing our horse?

Update: I just updated my game tree on page 2. It's getting kind of complicated so don't shoot me if I put a move in the wrong place.

Title: Re: Move 35
Post by Soter on Jun 24th, 2008, 11:37am
Hmm... My original ed3e ee2s dc4n dc5n is already refuted IIRC, mainly because  Rg1w in the next move destroys the possibility of pushing gold H to f1. Anything else with e->e2 ( ed3e ee2s xxxx xxxx ) looks either slightly worse or utterly bad. For now I support 99's 35b, although I must express my worries that the repertoire of good moves at our disposal seems to shrink recently :-/.

Warren, what about your 35b dc4s dc3n Rc2n xxxx?

Title: Re: Move 35
Post by warren on Jun 24th, 2008, 12:39pm

on 06/24/08 at 09:56:41, Soter wrote:
Now a different kettle of fish: assuming that the mob chooses 35b ed3e ee3s Hf2s ee2e, is something like 36w Da3n Da4n Da5n Da6e even remotely playable for chessandgo? Yes, I know this one looks rather wacky, but don't forget about Jean's penchant for surprises...


I think we take his cat and say thank you. Bomb says:
35b ed3e ee3s Hf2s ee2e
36w Da3n Da4n Da5n Da6e
36b Cg3w Cf3x dg2n dc4n rd5e
37w Eb3e hb2n Ec3n hb3e hc3x
37b ef2n Hf1n ef3n Hf2n Hf3x
38w Ec4s dc5s Ec3e dc4s dc3x

The final position is dccr vs DD. I don't think we have to worry about this 36w.

Title: Re: Move 35
Post by mistre on Jun 24th, 2008, 12:45pm

on 06/24/08 at 11:37:11, Soter wrote:
Warren, what about your 35b dc4s dc3n Rc2n xxxx?


This move was refuted already by Fritzlein:

"I'm not sure the move 35b dc4s dc3n Rc2n xxxx guarantees us any compensation for our dog.  Can't he just take our dog, and then if we flip his rabbit, move Da3->b6?"

Title: Re: Move 35
Post by warren on Jun 24th, 2008, 12:51pm

on 06/24/08 at 11:37:11, Soter wrote:
Warren, what about your 35b dc4s dc3n Rc2n xxxx?

Resulting lines seem to generally result in us threatening to trap a rabbit or dog at c6, loss of our dog, freedom of our horse and less-sharp positions, with material unclear at HDDC vs. hdccr (dog for cat and rabbit). I don't have time right now to post a representative line.

Update: looking briefly at it again, I no longer believe that we necessarily free our horse.

Title: Re: Move 35
Post by RonWeasley on Jun 25th, 2008, 5:41am

on 06/24/08 at 10:12:28, mistre wrote:
What we should be asking is if there is a better 35b than ed3e ee3s Hf2s ee2e?, and at this point no one is providing alternatives.


Just to be different, I'm looking at Soter's original suggestion.  I didn't see the refutation. (?)

35b ed3e ee3s dc4n rd5e

and continuing:

36w Rg1w Dd4w De4s Da3n
36b Hf2n ee2w Cg3w dg2n
and silver tries for domination around f3 including goal threats while gold takes the h.  This looks risky but worth considering.

or

36w Rg1w Re2n Re3w xxxx
36b ee2s Hf2e He2e ee1n
and silver has an easier time dominating f3 and getting a goal threat.  Less risky but risky still.

or

36w dg2s Hf2w xxxx xxxx
and the double threat is temporarily a single threat and silver can try to cover the threats while swarming on both sides.  Risky again, maybe the most risky of the three since it's harder to threaten the H around f3 here.  Still, at this point, I'm not sure trading h for C is unthinkable.

Title: Re: Move 35
Post by mistre on Jun 25th, 2008, 9:07am
Thanks at looking at an alternative.   On your 36b examples, I think you got the east and west directions backwards - I couldn't get them to work the way you described until I switched east with west and vice versa.

Also, your third 36w is not physically possible.  Did you mean 36w Rg1w dg2e Hf2e xxxx?

Just based on your initial analysis, this move looks a bit weaker than taking the E to f2.

Title: Re: Move 35
Post by warren on Jun 25th, 2008, 12:14pm

on 06/25/08 at 05:41:16, RonWeasley wrote:
Just to be different, I'm looking at Soter's original suggestion.  I didn't see the refutation. (?)

35b ed3e ee3s dc4n rd5e

Looking at that 35b, I don't see how white can force a material win if silver keeps swarming. If in fact he can't, this 35b is probably better than 99's. What am I missing? Can someone post a plausible line where silver has no choice but to give up material?

Title: Re: Move 35
Post by Fritzlein on Jun 25th, 2008, 1:38pm

on 06/25/08 at 12:14:14, warren wrote:
Looking at that 35b, I don't see how white can force a material win if silver keeps swarming. If in fact he can't, this 35b is probably better than 99's. What am I missing? Can someone post a plausible line where silver has no choice but to give up material?

It seems to me that after 35b dc4n rd5e ed3e ee3s chessandgo can force a material gain with 36w Rd2n Rd3e Dd4e Rg1w.  Since we can't defend c3 without our elephant, one way to avoid immediate material loss is 36b ee2w Hf2w ed2n dg2x, but that just delays the inevitable a couple of moves while Gold's position improves.  Ron's idea of flipping the horse with 36b ee2s Hf2w He2w ee1n loses a horse for a cat and his elephant centralizes in time to defend after
37w Eb3e hb2n Ec3e hb3e hc3x
37b ee2e Re3s Cg3w Cf3x dg2n
38w De4e Ed3e Ee3e Da3e

Now one can argue that we are happy to have CCR for H in this position, but nevertheless we lose material from our current situation of CR for nothing.  Maybe there is a better attacking move on 36b than flipping the horse, but in a similar position it seemed that pushing the horse into f3 made for a pretty slow attack, and material loss as well.

I don't think we have a move that doesn't lose material, and the question is just how good an attack we can get for it.

Title: Re: Move 35
Post by Fritzlein on Jun 25th, 2008, 3:03pm

on 06/24/08 at 09:32:12, RonWeasley wrote:
Gold must put the D on g4.  This frees the r on d4 to cover c3 if gold pulls up the h.  Otherwise, our e can stick to the D and silver can swarm rabbits on the east side.  Probably trade h for D, but that give silver a clean east side to make threats on.

I looked a bit at this position to see whether we would necessarily lose further material, because if superficially it looks like we might be able to hold onto a CC for D material advantage.  Our horse can just run away unless he re-hostages it, and if he re-hostages we get more time for our goal attack.  For example:

35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w
36b rb5e rc5s Cg3w Cf3x dg2n
37w rc4s rc3x Dd4w Eb3e hb2n
37b dg3s Rg1w dg2s rd5s
38w rd4n Dc4e Dd4e De4e
38b rd5s rd4s rh7s rh6s
39w hb3s Ec3w Df4e Dg4e
39b rg6s rg5s rg8s rg7s
40w Eb3e Ec3w rd3w rc3x Rd2e
40b rg6s rg4s rg5s rg3s

and we win
or

39w hb3s Ec3w Df4e Da3n
39b rh5s rg6e rh6s rh4s
40w Eb3e Ec3w rd3w rc3x Dg4e
40b ef2e eg2n rh3s rh2w

and we win

Tracing back up the tree, I think chessandgo can't play a greedy 39w, which means we stay ahead by CC for D, which means we are winning in this line, which means that we are winning after 99of9's move against all suggested replies for chessandgo.  (38w in this line was forced, and 37w was the move we were afraid of but don't need to be any more.)

Or have I missed some analysis?  What do we think is chessandgo's best reply to 99of9's move?  How much are we ahead then?

Title: Re: Move 35
Post by 99of9 on Jun 25th, 2008, 3:24pm

on 06/25/08 at 15:03:14, Fritzlein wrote:
Tracing back up the tree, I think chessandgo can't play a greedy 39w, which means we stay ahead by CC for D, which means we are winning in this line, which means that we are winning after 99of9's move against all suggested replies for chessandgo.  (38w in this line was forced, and 37w was the move we were afraid of but don't need to be any more.)

Or have I missed some analysis?  What do we think is chessandgo's best reply to 99of9's move?  How much are we ahead then?

I agree that is the current state, so I am happy to vote for this now.  But I understand if anyone wants a delay, because it does seem like a crucial move.  We are offering free material, and currently believe we can forcibly win a strong attack or some material in return.

I think we were right to analyze a sacrifice move more deeply than normal, even though we didn't have any particular alternatives.  The key was to correctly value this move, because it oscillated between very positive and very negative.  If it had turned out to be losing, then (in the absence of alternative possible positive moves) we could then turn to finding another move that lost us only a bit of ground.

Title: Re: Move 35
Post by RonWeasley on Jun 25th, 2008, 4:35pm
This looks like a vote for our horse push move by 99 and Fritz.  Add my vote.  If we get 5 more, I will make the move.

Notice I'm changing the majority to 8 since warren is new and has been very active.

Title: Re: Move 35
Post by mistre on Jun 25th, 2008, 4:36pm

on 06/25/08 at 15:24:40, 99of9 wrote:
I think we were right to analyze a sacrifice move more deeply than normal, even though we didn't have any particular alternatives.  The key was to correctly value this move, because it oscillated between very positive and very negative.  If it had turned out to be losing, then (in the absence of alternative possible positive moves) we could then turn to finding another move that lost us only a bit of ground.


Thanks for clarifying why we analyzed this move so far into the future.  It makes more sense to me now.  

I'll vote for horse push as well.

Title: Re: Move 35
Post by jdb on Jun 25th, 2008, 6:09pm
I would like to use some more time before voting. We have 16 days in the reserve, and if we are right about this move we will not need the time anyway.

I am comfortable enough that if gold plays the following 36w, silver will have a very good position.

35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w

However, this 36w is not forced. Maybe gold could even try,
36w H->e3 xxxx    (maybe Dd4e)

It is very likely chessandgo will see the danger in capturing the dog on 36w and take enough time to come up with an alternative move. I think it would be wise to spend a few days looking at some of his alternatives on 36w.


Title: Re: Move 35
Post by warren on Jun 25th, 2008, 11:34pm

on 06/25/08 at 18:09:39, jdb wrote:
I would like to use some more time before voting. We have 16 days in the reserve, and if we are right about this move we will not need the time anyway.

I am comfortable enough that if gold plays the following 36w, silver will have a very good position.

35b ed3e ee3s Hf2s ee2e
36w Eb3e Ec3w dc4s dc3x Hf1w

However, this 36w is not forced. Maybe gold could even try,
36w H->e3 xxxx    (maybe Dd4e)

It is very likely chessandgo will see the danger in capturing the dog on 36w and take enough time to come up with an alternative move. I think it would be wise to spend a few days looking at some of his alternatives on 36w.

Without the threat against our eastern dog his position seems to fall apart. Here's one line:

35b ed3e ee3s Hf2s ee2e
36w Hf1w He1n He2n Dd4e
36b rd5s rd4s Cg3n dg2n
37w Da3n rd3w He3w De4w
37b Cg4w dg3n Cf4s Cf3x dg4w
38w Eb3n Eb4s dc4w rc3x Dd4w
38b rg6s rg5s rg4s ef2w
The above 38w is obviously not his best one, as it loses the game, but I don't see any obvious way for him to gain material.

I haven't looked at this very hard at all, but it doesn't look like a good 36w at first glance. I do like the idea of finding better 36w though.

Title: Re: Move 35
Post by Soter on Jun 26th, 2008, 12:52am
I like 99's proposal ( horse push -> f1 ) but I also agree with JDB on considering other 36w possibilities and delaying our vote - just a bit. IMHO two days will suffice.

Title: Re: Move 35
Post by 99of9 on Jun 26th, 2008, 2:04am

on 06/25/08 at 18:09:39, jdb wrote:
However, this 36w is not forced. Maybe gold could even try,
36w H->e3 xxxx    (maybe Dd4e)

I think it would be wise to spend a few days looking at some of his alternatives on 36w.

I agree, I'm happy for people to keep working on this, there are plenty of branches to analyse.

Warren's reply to your example looks quite good.  A different 4th step we should consider is:
36w H->e3 Rd2n

Title: Re: Move 35
Post by mistre on Jun 26th, 2008, 6:12am

on 06/26/08 at 02:04:40, 99of9 wrote:
Warren's reply to your example looks quite good.  A different 4th step we should consider is:
36w H->e3 Rd2n


I was going to post that I couldn't find a good answer to this move, but then I found this 36b which looks like our best alternative:

36b rf7s rf6s rf5s rf4s

He can adequately defend with:
37w rf3n He3e rf4e Hf3n
or
37w Dd4e De4e Df4w rf3n

Does anyone want to analyze from here?

Another option is
36b E -> d3, push R into trap.
37w Rg1w Rf1n Da3n Da4n

Then what is our strategy?  Retreat our east dog to safety and retreat our west dog to defend c6 trap?

Edited to reflect move 36b instead of 37b.  Sorry about that.

Title: Re: Move 35
Post by warren on Jun 26th, 2008, 12:38pm

on 06/26/08 at 06:12:47, mistre wrote:
I was going to post that I couldn't find a good answer to this move, but then I found this 37b which looks like our best alternative:

37b rf7s rf6s rf5s rf4s

He can adequately defend with:
37w rf3n He3e rf4e Hf3n
or
37w Dd4e De4e Df4w rf3n

Does anyone want to analyze from here?

Another option is
37b E -> d3, push R into trap.
37w Rg1w Rf1n Da3n Da4n

Then what is our strategy?  Retreat our east dog to safety and retreat our west dog to defend c6 trap?

Warren and bomb produce another line of unknown plausibility:
35b ed3e ee3s Hf2s ee2e
36w Hf1w He1n He2n Rd2n
36b ef2w ee2w Rd3w ed2n
37w Rg1w Rf1n Rc1e Rd1n
37b Cg3n dg2n dg3e dh3n
38w Da3n Da4e rb5w Db4n
38b rh7s rh6s ce7w cd7s
39w ra5s Db5w Rd2e Rh1n
39b Cg4n dh4w Cg5w dg4n
40w He3n He4e Cf5w Hf4e
40b dg5w rg6s df5n df6w
41w rg5w Hg4n Rh2n Rh3n
41b de6e Ce5n df6e Ce6e Cf6x

If instead he does
37w Rg1w Rf1n Da3n Da4n
our reply of
37b Cg3e dg2n rg6s rg5s
seems quite strong for us.

Title: Re: Move 35
Post by warren on Jun 27th, 2008, 12:44pm

on 06/25/08 at 16:35:38, RonWeasley wrote:
This looks like a vote for our horse push move by 99 and Fritz.  Add my vote.  If we get 5 more, I will make the move.

Notice I'm changing the majority to 8 since warren is new and has been very active.


Add my vote as well.

Title: Re: Move 35
Post by Soter on Jun 28th, 2008, 3:39am
Yes, let's go with the horse-pushing move.

Title: Re: Move 35
Post by arimaa_master on Jun 28th, 2008, 11:11am

on 06/25/08 at 16:35:38, RonWeasley wrote:
This looks like a vote for our horse push move by 99 and Fritz.  Add my vote.  If we get 5 more, I will make the move.

Notice I'm changing the majority to 8 since warren is new and has been very active.


Add my vote too.

Title: Re: Move 35
Post by warren on Jun 28th, 2008, 11:34am

on 06/28/08 at 11:11:58, arimaa_master wrote:
Add my vote too.

I see seven votes for the horse-move so far (99, Fritz, RW, mistre, Warren, Soter, AM), though Fritz's wasn't an explicit vote. We are therefore waiting either for more analysis (I'll withdraw my vote if any important new analysis is posted), a vote from jdb or a vote from a less active poster such as Janzert.

Title: Re: Move 35
Post by The_Jeh on Jun 28th, 2008, 12:08pm
I'll add my vote, too.

Title: Re: Move 35
Post by jdb on Jun 28th, 2008, 12:21pm

on 06/28/08 at 12:08:39, The_Jeh wrote:
I'll add my vote, too.


Me too.

Title: Re: Move 35
Post by Janzert on Jun 28th, 2008, 7:03pm
I'll go for it too.

Btw, here is OpFor's analyses after making this move, i.e. from 36w.

depth 16
time 9424
nodes 733089475
qnodes 194132715
score -94
pv Eb3e Ec3w dc4s dc3x Hf1w b Cg3w Cf3x dg2n rb5e rc5s w rc4s rc3x Dd4w Eb3n Eb4s b ef2w ee2n ee3w rg6s

Janzert

Title: Re: Move 35
Post by 99of9 on Jun 28th, 2008, 8:07pm

on 06/28/08 at 19:03:50, Janzert wrote:
pv Eb3e Ec3w dc4s dc3x Hf1w b Cg3w Cf3x dg2n rb5e rc5s w rc4s rc3x Dd4w Eb3n Eb4s b ef2w ee2n ee3w rg6s

Interesting that Opfor can't see anything productive for Jean to do on ply 11 and 12.  Thanks for posting, it's good to see that Opfor didn't find any tricky saves for Jean.

Title: Re: Move 35
Post by Janzert on Jun 28th, 2008, 9:04pm

on 06/28/08 at 20:07:19, 99of9 wrote:
Interesting that Opfor can't see anything productive for Jean to do on ply 11 and 12.  Thanks for posting, it's good to see that Opfor didn't find any tricky saves for Jean.


At a depth of 12 opfor puts c&g up by almost a rabbit and sends the dog at d4 west and north for steps 11 and 12. But after searching to 13 steps deep the score drops back in our favor and by 14 steps OpFor seems to think the d4 dog advance is bad (maybe leaves the south east too weak?).

Janzert

Title: Re: Move 35
Post by warren on Jun 29th, 2008, 5:14am
I left bomb analyzing for a while and it suggested

36w Rc2n Rc3e Rd3e Dd4s

What's a good response for us to that?

Update: how about
36b Cg3e dg2n rg6s rg5s

Update 2: this doesn't seem much different than the other moves we've seen. Nothing to worry about further unless he plays it I think.

Title: Re: Move 35
Post by Soter on Jun 30th, 2008, 2:39am
I hope Ron will turn up soon...

Title: Re: Move 35
Post by RonWeasley on Jun 30th, 2008, 5:17am

on 06/30/08 at 02:39:56, Soter wrote:
I hope Ron will turn up soon...


I spent the weekend with the family.  Hope that didn't make TheMob ... angry!  I have made the horse push move.

Keep in mind that if I go on extended vacation, as I will in August, I can give the password to a substitute moderator.

Chessandgo volunteers that he did not expect our last two moves in his analysis.  He didn't say if that was good or bad.

Title: Re: Move 35
Post by warren on Jun 30th, 2008, 9:50am

on 06/30/08 at 05:17:17, RonWeasley wrote:
I spent the weekend with the family.  Hope that didn't make TheMob ... angry!  I have made the horse push move.

Keep in mind that if I go on extended vacation, as I will in August, I can give the password to a substitute moderator.

Chessandgo volunteers that he did not expect our last two moves in his analysis.  He didn't say if that was good or bad.

We expected his 34w but not his 32w, 33w or 35w. I wonder if we can tell him that without giving too much away.

Title: Re: Move 35
Post by mistre on Jun 30th, 2008, 9:55am

on 06/30/08 at 05:17:17, RonWeasley wrote:
Chessandgo volunteers that he did not expect our last two moves in his analysis.  He didn't say if that was good or bad.


Does that mean he expected all of the rest of our moves?!  ;)

It would actually be an interesting exercise to go back and see how many of Chessandgo's moves we saw coming and how many of them we were completely surprised on.


Title: Re: Move 35
Post by RonWeasley on Jun 30th, 2008, 10:06am
He volunteered this even after I told him I didn't intend to share his comments with the rest of TheMob.  He said he didn't mind.  After the game is over, I intend to release a transcript of the in-game comments.  It will be pretty boring.  Even the stuff about his mother.

Title: Re: Move 35
Post by warren on Jun 30th, 2008, 10:09am
It amazes me that Chessandgo responded to our 34b in only  12 hours even though he says he wasn't expecting it! This is especially remarkable given that his reserve overflowed that move, so he could have waited a bit longer for free.

Title: Re: Move 35
Post by warren on Jun 30th, 2008, 10:13am

on 06/30/08 at 10:06:56, RonWeasley wrote:
He volunteered this even after I told him I didn't intend to share his comments with the rest of TheMob.  He said he didn't mind.  After the game is over, I intend to release a transcript of the in-game comments.  It will be pretty boring.  Even the stuff about his mother.

Did you mean to say "... I told him I intended to share"?

Title: Re: Move 35
Post by Fritzlein on Jun 30th, 2008, 10:22am

on 06/30/08 at 09:50:50, warren wrote:
We expected his 34w but not his 32w, 33w or 35w. I wonder if we can tell him that without giving too much away.

Sure we can tell him what we were thinking, because our thinking has completely changed since then.  Also we can tell him whether we think we are winning or losing at the present (as long as we don't mention specific moves or strategies for either of us) because our opinion of whether we are winning or losing keeps changing.  If we are perfectly frank, we have a good chance of misleading him.  :)

Title: Re: Move 35
Post by Fritzlein on Jun 30th, 2008, 10:29am
The Mob game has to be one of the highest quality and most thoroughly analyzed games of Arimaa ever, and even so the two sides regularly surprise each other.  I'm trying to get my mind around how far we are from anyone feeling that Arimaa is "played out" and needs rule changes, as World Champion Capablanca felt about chess in the 1920's.

Title: Re: Move 35
Post by RonWeasley on Jun 30th, 2008, 12:10pm

on 06/30/08 at 10:13:45, warren wrote:
Did you mean to say "... I told him I intended to share"?


As written.  I didn't intend to share until he said he didn't mind.  After that, I did the opposite of what I originally intended.  Clear?  I think Jean wants this to be as fun as possible for everyone.  I really like his attitude.  That and his playing ability.



Arimaa Forum » Powered by YaBB 1 Gold - SP 1.3.1!
YaBB © 2000-2003. All Rights Reserved.