Arimaa Forum
        (http://arimaa.com/arimaa/forum/cgi/YaBB.cgi)
      

        Arimaa >> Game Analysis >> clyring's 2013 Arimaa Challenge practice games
        
(Message started by: clyring on Mar 26th, 2013, 6:18pm)
Title: clyring's 2013 Arimaa Challenge practice games
Post by clyring on Mar 26th, 2013, 6:18pm
With the challenge match starting next week, I have started final preparations by practicing against strong server bots at 2m/move. Anyone who wants to help improve my ability to defend humanity's honor is welcome to review my practice games:



clyring-bot_Marwin2012CC 1-0 (http://arimaa.com/arimaa/gameroom/comments.cgi?gid=261431)

    I think I handled this game mostly well, but feel like I lost my way somewhere between moves 11 and 22. The conversion technique demonstrated after move 46 is beautiful and may be worth studying.

(More to come when I play more practice games. For now I am tired enough from three hours of exertion.)
Title: Re: clyring's 2013 Arimaa Challenge practice games
Post by Fritzlein on Mar 26th, 2013, 7:06pm
In replaying the main thing that jumps out at me is that 18g is a very nice move.  Well done!

On 27g I think it is slightly more accurate to push away the cat and push the camel into the trap, as opposed to retreating your horse.  When taking a camel hostage, you would rather have your horse active, so your game move spends a step against your strategic purpose.  Not a big deal, though.

I would have framed the cat on 30g (your cat unfreezing your dog), just to kill all counterplay.  Of course you handled marwin's counterplay just fine, but with the frame he would have been even more paralyzed.

I approve of your 39g, bringing the camel home to stabilize the defense.  By that point you no longer need the camel on attack to win.  40g, on the other hand, seems like a concession.  Why trade horses when 40g Cd3n Me3w Cd4n Hc5w leaves you with the sheriff in the three quadrants other than the elephant deadlocked one?  Not that you had any difficulty winning from then on, but these minor points are the only things I can see to contribute.  :)
Title: Re: clyring's 2013 Arimaa Challenge practice games
Post by hyperpape on Mar 26th, 2013, 10:02pm
For a beginner: can you explain what you mean by the conversion technique?
Title: Re: clyring's 2013 Arimaa Challenge practice games
Post by clyring on Mar 27th, 2013, 11:08am

on 03/26/13 at 19:06:04, Fritzlein wrote:
In replaying the main thing that jumps out at me is that 18g is a very nice move.  Well done!
Thanks! That means a lot coming from one as strong as yourself.


on 03/26/13 at 19:06:04, Fritzlein wrote:
On 27g I think it is slightly more accurate to push away the cat and push the camel into the trap, as opposed to retreating your horse.  When taking a camel hostage, you would rather have your horse active, so your game move spends a step against your strategic purpose.  Not a big deal, though.
I considered that move first during the game, but decided that moving my horse to block the camel from retreating all the way to a5 would save me a net step even though it is retreating my horse and not pushing the enemy cat back. That said, maybe it would have been more accurate still to move my horse to a4 instead.


on 03/26/13 at 19:06:04, Fritzlein wrote:
I would have framed the cat on 30g (your cat unfreezing your dog), just to kill all counterplay.  Of course you handled marwin's counterplay just fine, but with the frame he would have been even more paralyzed.
Good catch. I didn't see this move at all during the game, but I agree that it would have been cleaner than the played sequence.


on 03/26/13 at 19:06:04, Fritzlein wrote:
I approve of your 39g, bringing the camel home to stabilize the defense.  By that point you no longer need the camel on attack to win.  40g, on the other hand, seems like a concession.  Why trade horses when 40g Cd3n Me3w Cd4n Hc5w leaves you with the sheriff in the three quadrants other than the elephant deadlocked one?  Not that you had any difficulty winning from then on, but these minor points are the only things I can see to contribute.  :)
Ah, I missed the option of declining the exchange of my active horse for his hostaged horse in deciding to capture the dog on move 39g. Yes, that move would have been more accurate.


on 03/26/13 at 22:02:46, hyperpape wrote:
For a beginner: can you explain what you mean by the conversion technique?
Conversion is the process of converting a decisive material advantage into a formally won game. Before move 46g, I have just won two cats and three rabbits, with which it took me a mere nine moves to wrap up the game.

For any bot developers reading this: Was the played 50g a goal in five?
Title: Re: clyring's 2013 Arimaa Challenge practice games
Post by supersamu on Mar 27th, 2013, 12:13pm
Here are my comments about:  clyring-bot_Marwin2012CC 1-0
- until move 20g: not a lot to talk about. You pulled a rabbit while marwin shuffled its pieces, and then captured it while marwin got a very questionable horse frame.

- 22g: why leave g6 with the horse? the whole next sequence with the elephant going after the camel would not have been possible

- 30g: I agree with Fritzlein that framing the cat would have made the game easier for you

- 36g: last step camel to f4,
incidentally threatening false protection at f6

39g: capturing the dog with the camel ending on f4 looks better than wasting moves to capture a hostaged horse at 40g and 41g. That way, you can attack f6 after 39g.

Nice goal attack starting with 50g. While watching the game I wondered why you gave up the position at f6, but your elephant was strong enough in the northwest to generate goalthreats.
Title: Re: clyring's 2013 Arimaa Challenge practice games
Post by browni3141 on Mar 27th, 2013, 3:46pm
Rabbit pulls in a game with opposite wing MH pairs often justifies a attack on the opposite wing. marwin could have attacked c3 while you were pulling the rabbit and sacrifice it for time. It is harder to turn around and race on a wing where you've pulled a rabbit because it provides some space advantage while it is alive. Instead it just let you slowly pull and win the rabbit without trying to do much of anything about it :(. I suppose this bit doesn't apply much to more passive bots like marwin and ziltoid, but it's something to think about. I guess marwin kind of did it later, but in a weird way.

Did you consider
13g Hh6n rg7n Hh7w Cg2n?

13s cb5w mb6s mb5s cd7e is the best response I see, and the game might be sharp, but you've got an EH attack and will probably soon win a rabbit. marwin probably wouldn't have even played an aggressive move here anyway :P
There's my attempt to talk about something before 20g. There wasn't really anything obvious, but I tried!

On 35g was it really necessary to allow marwin's horse to get away?
On 36g it seems the last step could have been better. I like Mg4w best.
41g: Maybe postponing the capture and attacking f6 in some way would have been faster so that marwin couldn't clog things up.
32g: Placing the camel on that square would make me nervous, but it's probably perfectly fine.
I tried to be a little nit-picky since the biggest things have already been said.
Title: Re: clyring's 2013 Arimaa Challenge practice games
Post by Nombril on Mar 28th, 2013, 4:56pm
A couple of things I noticed:

Is it worth framing a rabbit at 32g?

One spot where you could have created multiple threats was at 37g.  When you take a third piece hostage with your elephant, you haven't created new threats.  Even worse, to capture one of the pieces you must let at least one of them go.  What about 37g Mg4w Mf4s Ce3n Mf3w to create a second threat at f3 right away?

I like the looks of this better:
47g Cg3n Cg4n rg6w Cg5n

Have fun with the challenge games, you should do well!


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