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

        Arimaa >> Off Topic Discussion >> 2012 Havannah Challenge Match
        
(Message started by: omar on Mar 30th, 2012, 11:44am)
Title: 2012 Havannah Challenge Match
Post by omar on Mar 30th, 2012, 11:44am
Christian, I had some questions relating to how the challenge match should be handled.

1. If there are more than one program that want to challenge you, should there be some tournament between the programs to see which one will be allowed to challenge you? Or were you planning to play 10 games against each of them. Or perhaps, once you've defeated a program then any challenger has to also defeat the program you most recently defeated before challenging you.

2. During the challenge match is the programmer allowed to make any changes to the program between games.

3. During challenge match games, can the programmer operate the program or must it be operated by an independent referee.

That's it for now.
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Apr 1st, 2012, 6:19am

on 03/30/12 at 11:44:18, omar wrote:
Christian, I had some questions relating to how the challenge match should be handled.

1. If there are more than one program that want to challenge you, should there be some tournament between the programs to see which one will be allowed to challenge you? Or were you planning to play 10 games against each of them. Or perhaps, once you've defeated a program then any challenger has to also defeat the program you most recently defeated before challenging you.

I play ten games total, and I don't care which bot(s) as long as I know 'who' I'm playing against. I've also insisted to play against every participating bot a couple of times in the weeks before to the contest. Castro is not Wanderer, and the Germans either pulled out or they may be preparing their secret weapon, I don't know.

on 03/30/12 at 11:44:18, omar wrote:
2. During the challenge match is the programmer allowed to make any chances to the program between games.

3. During challenge match games, can the programmer operate the program or must it be operated by an independent referee.

That's it for now.

No fiddling during a game, but in between games is ok by me, plus a credible guarantee that I'm indeed playing against the bot.
Title: Re: 2012 Havannah Challenge Match
Post by omar on Apr 2nd, 2012, 2:35pm
Thanks, and what did you have in mind for the time controls?
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Apr 2nd, 2012, 3:02pm

on 04/02/12 at 14:35:13, omar wrote:
Thanks, and what did you have in mind for the time controls?

Something like an hour per side and maybe 30 sec plus per move would suit me. Albert Vasse of DGT (http://digitalgametechnology.com/site/) is also involved (they're based in Enschede where I live) so anything goes.
Title: Re: 2012 Havannah Challenge Match
Post by omar on Apr 3rd, 2012, 8:27am
Tony would like to know whether a program bug which makes the program need a restart, automatically means disqualification. Or can the programmer intervene and restart the program.
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Apr 3rd, 2012, 9:30am

on 04/03/12 at 08:27:28, omar wrote:
Tony would like to know whether a program bug which makes the program need a restart, automatically means disqualification. Or can the programmer intervene and restart the program.

Depends on the bug. If it causes a bad move I don't care. If there's a system crash I think it's fair to have it fixed in its own time, not in mine :)
Title: Re: 2012 Havannah Challenge Match
Post by omar on Apr 4th, 2012, 10:47am

on 04/03/12 at 09:30:29, christianF wrote:
Depends on the bug. If it causes a bad move I don't care. If there's a system crash I think it's fair to have it fixed in its own time, not in mine :)


Well maybe if the program was restarted again it might pick a different move and continue the game, but if it continues to pick a bad move or continues to crash then after a few tries the only alternative would be for the developer to try and fix the code during the game. Are you saying that you are OK with the developer fixing the bug during the game as long as the bots clock is counting down?
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Apr 4th, 2012, 2:18pm

on 04/04/12 at 10:47:30, omar wrote:
Well maybe if the program was restarted again it might pick a different move and continue the game, but if it continues to pick a bad move ...

If it moves that move counts, so how would we know a move was 'different' ???



on 04/04/12 at 10:47:30, omar wrote:
... or continues to crash then after a few tries the only alternative would be for the developer to try and fix the code during the game. Are you saying that you are OK with the developer fixing the bug during the game as long as the bots clock is counting down?

That might be problematic. Does a clock continue if the program crashes? I'm no expert. I wouldn't mind if any technical problems were solved while the game pauses, if it doesn't upset things to a degree that might have a negative impact on concentration. I don't have much of an "off/on" switch when it comes to playing. I prepare mentally for one game, the next one, and that's about how far my powers of concentration will go. I dislike distractions or interruptions during a game. In case of a serious computer crash, a game might as well be postponed till the next day, as far as I'm concerned.
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Apr 28th, 2012, 10:12am
I had another invite from wanderer_c (http://www.littlegolem.net/jsp/info/player.jsp?plid=21160) and there are some improvements - not quite enough to beat me yet (http://www.littlegolem.net/jsp/game/game.jsp?gid=1443784), but some anyway :)
Title: Re: 2012 Havannah Challenge Match
Post by Katsunami on Apr 28th, 2012, 10:40am
Sorry to bust in, but is this "Havanna Challenge" for Go bots?

The game you linked looks like Go on a honeycomb shaped board; is it that, or is it actually something else with different rules? I've been searching a bit, but there seem to be many games that are called "Havanna", "Havanah", or similar, with different rules, and sometimes they're completely different games.

Ah, found it.

http://en.wikipedia.org/wiki/Havannah

Developed by yourself I see. Nice. I'll Go (:P) and read the rules.

(It's a bit off-topic, but I find it baffling that whenever there is something to be developed or something to be done in the open source world, that there are always many more Dutch people in it than can be expected from the size of the population.)
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Apr 28th, 2012, 11:01am

on 04/28/12 at 10:40:04, Katsunami wrote:
Sorry to bust in, but is this "Havanna Challenge" for Go bots?

Not quite. I'm playing ten games this summer, against the 4 or 5 top bots, and must win all to win a challenge put out in 2002.


on 04/28/12 at 10:40:04, Katsunami wrote:
Ah, found it.

http://en.wikipedia.org/wiki/Havannah

Developed by yourself I see. Nice. I'll Go (:P) and read the rules.

(It's a bit off-topic, but I find it baffling that whenever there is something to be developed or something to be done in the open source world, that there are always many more Dutch people in it than can be expected from the size of the population.)

Most Havannah info can be found at mindsports (http://mindsports.nl/), best place to play is Little Golem (http://www.littlegolem.net/).

Regarding a bot contests, very shortly now the 2013 CodeCup Challenge (http://www.codecup.nl/intro.php) will be announced and the game to program will be Symple (http://mindsports.nl/index.php/arena/symple/), in case you'd like to compete :)
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Aug 8th, 2012, 7:54am
Here's an update (https://plus.google.com/105607364226084069219/posts) on the Challenge, at Google+

Also, the CodeCup Challenge 2013 (http://www.codecup.nl/intro.php) has started with Symple as the piece de resistance.
Title: Re: 2012 Havannah Challenge Match
Post by Fritzlein on Aug 8th, 2012, 11:54am
Is there a way to link to completed games as they finish (assuming the games are preserved for posterity)?  I'm rooting for you 100%.  Go carbon-based life forms!
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Aug 8th, 2012, 12:16pm

on 08/08/12 at 11:54:09, Fritzlein wrote:
Is there a way to link to completed games as they finish (assuming the games are preserved for posterity)?  I'm rooting for you 100%.  Go carbon-based life forms!

Yes, the games will be played at iGGC and there will be a video livestream, and a YouTube summary after the event. Thanks for your support, I'll try to be a good carbon-based representative.

Actually I got an agenda too: don't trust 'intelligent systems' to make critical decisions (markets, military and the like) if they can't win a pencil & paper game. Amplifying the calculation speed of questionable algorithms doesn't make them any more 'intelligent', just faster :o
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Oct 2nd, 2012, 9:49am
Here's the press release (http://mindsports.nl/havannah-man-vs-machine).
Title: Re: 2012 Havannah Challenge Match
Post by Fritzlein on Oct 2nd, 2012, 8:12pm

on 10/02/12 at 09:49:04, christianF wrote:
Here's the press release (http://mindsports.nl/havannah-man-vs-machine).

Thanks!  I hope I get a chance to listen to live commentary of at least a couple of games.
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Oct 3rd, 2012, 5:45am

on 10/02/12 at 20:12:54, Fritzlein wrote:
Thanks!  I hope I get a chance to listen to live commentary of at least a couple of games.


I hope it's not too boring, an old guy behind a screen, making a move every few minutes :P .

If I win all games of the first four days, I'll consider hanging Kobus around my neck on the fifth day.

http://i50.tinypic.com/98u2wm.jpg

This may add too visual entertainment, and since I'm trying to show that computers still lack 'intelligence' it may also provide some balance by proving without a doubt that humans may lack intelligence too ;D .
Title: Re: 2012 Havannah Challenge Match
Post by Fritzlein on Oct 12th, 2012, 4:11am
3 meters long?  4?   :o
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Oct 12th, 2012, 5:19am

on 10/12/12 at 04:11:55, Fritzlein wrote:
3 meters long?  4?   :o


When I got him, eleven years ago, he was 3.5 m and weighed 20 k. Now he's 4+ m (couple of cm past 4) and weighs 30+ k.

In ten years I'll be 75 (Whoever's-Up-There willing) and he'll be 4.5 m 50 k. By that time, or earlier if I die or get otherwise incapacitated, I've reserved a place for him at a reptile zoo in Vlissingen (http://www.iguana.nl/). I think it would be wise by then not to invoke Murphy's Law, lest I should become an item in "What a Way to Go" :P .
Title: Re: 2012 Havannah Challenge Match
Post by Fritzlein on Oct 15th, 2012, 3:44pm
Ah, shoot, I missed the first two games this morning, even though I was free due to Fall Break.  The standings are

Humanity 1
Computers 0
Internet Failure 1
Title: Re: 2012 Havannah Challenge Match
Post by Arimabuff on Oct 16th, 2012, 1:04pm

on 10/15/12 at 15:44:32, Fritzlein wrote:
Ah, shoot, I missed the first two games this morning, even though I was free due to Fall Break.  The standings are

Humanity 1
Computers 0
Internet Failure 1

You mean there's spring break AND fall break? WOW!  :o

;)
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Oct 17th, 2012, 12:32pm
It's now human 5 - bots 1, so I've lost the match, 4 games to go. The results are very interesting:
  • The bots have become significantly stronger (except Lajkonik in the first game, where it played like version 1.3 instead of 3.1)
  • They also contribute to strategic insight. The 'clusterbombs' they employ give more tempo than most players would anticipate, and these must be defused at an early stage (at the price of tempo of course)
  • The games are of a high standard, and I enjoy them tremendously - and I still want to go for the best possible result of course :)
  • Obviously I was wrong about the level of play of the bots.
My congratulations go to Marcin Ciura for his win in the second game!
Title: Re: 2012 Havannah Challenge Match
Post by Fritzlein on Oct 17th, 2012, 6:00pm
Thanks for the update!
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Oct 18th, 2012, 10:43am

on 10/17/12 at 18:00:59, Fritzlein wrote:
Thanks for the update!

18 October: 7-1
Title: Re: 2012 Havannah Challenge Match
Post by Fritzlein on Oct 19th, 2012, 8:08pm
Final score, 8-2 in favor of humanity.  Well done, Christian!  Too bad you lost the 1000 euros, but you are obviously still way ahead of the machines.
Title: Re: 2012 Havannah Challenge Match
Post by browni3141 on Oct 19th, 2012, 8:42pm

on 10/19/12 at 20:08:26, Fritzlein wrote:
Final score, 8-2 in favor of humanity.  Well done, Christian!  Too bad you lost the 1000 euros, but you are obviously still way ahead of the machines.

Are you sure it's 8-2? I read somewhere else that the score was 7-3.
Title: Re: 2012 Havannah Challenge Match
Post by Fritzlein on Oct 19th, 2012, 11:58pm

on 10/19/12 at 20:42:22, browni3141 wrote:
Are you sure it's 8-2? I read somewhere else that the score was 7-3.

Oh, then it is probably my mistake.  I didn't see the final games live.
Title: Re: 2012 Havannah Challenge Match
Post by omar on Oct 20th, 2012, 2:16am
Christian, congratulations on a 7-3 finish. If it had been best of 10 you clearly would have won. But you structured the challenge to very difficult for yourself. When I was talking to Fritz about it he mentioned that you would need to be about 350 rating points higher than the bots to pull of a streak of 10 wins. Earlier this year I play 10 postal games with bot Castro (the strongest of the current bots in rating) and had a 5-5 record against it. I've been playing Havannah for years now and Castro already seems to be on par with me.

Congradulations to Marcin Ciura also on winning the challenge and creating a strong bot.

Some related links:
http://mindsports.nl/index.php/arena/havannah/641

http://www.dgt.nl/site/index.php/dgtchesslab/activities/339-havannah-challenge-live-stream
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Oct 20th, 2012, 6:39am
That was great fun :)

The games: I hate analysis but here's the overall impression.

The first game Lajkonik presented itself as dumb - that's perfectly legal. It's not reversable either: I can't lure a bot into thinking ...

There was a second game that went wrong due to a connection bug. I wasn't too happy with the position at the time it was aborted. I should have been warned.

The next second game I chose the wrong strategy, plain and simple. You can't outrace a bot when it indicates readyness to engage. I thought I had the decisive tempo. Mea culpa.

Then wins up to and including the 8st game. All based on a strategy of enclosure at two sides because the bots prefer cluster bombing with ring threats for making speed, and will usually keep expanding close to the main structure, and very effectively at that, given the object.

In the 9th game I was already fairly exhausted, but I think I had the decisive tempo when I accidentally submitted a move instead of removing the stone I had been analyzing. I had a direct threat in two to answer and was well aware of that. Analysis will probably reveal I had a won position, but (sh)it happens :( .

In the final game against Castro, I saw a catastrophic move in the replay. Early on I placed a 'defensive' stone under a delusion that I was unware of at the time, like all delusions, but that I immediately spotted on replay: the danger was an illusion. That not only cost a tempo (and I was playing second as it was), but once you stick your foot in it, there are suddenly additional moves necessary, and Castro expoited that very good. From that point on I was running behind the facts.

My congratulations go to Timo and Marcin, and my thanks to all involved, Richard Lorentz, Ton van der Valk, Albert Vasse, Jan Krabbenbos, Frans Faase, Arty Sandler and Ed van Zon.

P.S.
I burned my ships behind me in the 5th game (I think) when I went for a bridge, because I didn't see how it could be prevented. I could have kept the option while securing a forkframe, but because I couldn't see it, putting it all on that one card seemed more aesthetical (hear the bots laughing).
Fortunately the bot didn't see it either. Jos Dekker saw it, and Ton van der Valk confirmed it: the bridge could be prevented. So I got away with it :) .
Title: Re: 2012 Havannah Challenge Match
Post by omar on Oct 20th, 2012, 7:42am
Thanks for the brief analysis Christian.

Here are the links to games. You first have to login to http://www.igGameCenter.com/ for the links to work.

Game 4 is still set to private so it can't be viewed.

To step through the moves you first have to click on one of the arrows under 'Move History'; then you can use the forward and back arrows at the top of the list.
1Lajkonik_botChristian Freeling* http://gc1.iggamecenter.com/gm.php?sid=299932
2Christian FreelingLajkonik_botx http://gc1.iggamecenter.com/gm.php?sid=300074
3Christian FreelingCastro_bot* http://gc1.iggamecenter.com/gm.php?sid=300092
4Christian FreelingWanderer_bot*private http://gc1.iggamecenter.com/gm.php?sid=30037800&lang=en
5Castro_botChristian Freeling* http://gc1.iggamecenter.com/gm.php?sid=300345
6Wanderer_botChristian Freeling* http://gc1.iggamecenter.com/gm.php?sid=300378
7Christian FreelingLajkonik_bot* http://gc1.iggamecenter.com/gm.php?sid=300491
8Christian FreelingCastro_bot* http://gc1.iggamecenter.com/gm.php?sid=300520
9Lajkonik_botChristian Freelingx http://gc1.iggamecenter.com/gm.php?sid=300715
10Castro_botChristian Freelingx http://gc1.iggamecenter.com/gm.php?sid=300773
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Oct 20th, 2012, 7:56am
Here's another source, courtesy of Z (http://www.littlegolem.net/jsp/info/player.jsp?plid=15076) at LittleGolem (http://www.littlegolem.net/jsp/index.jsp):
  • game 1 (http://www.trmph.com/havannah/game/hvc2012-1)
  • game 2 (http://www.trmph.com/havannah/game/hvc2012-2)
  • game 3 (http://www.trmph.com/havannah/game/hvc2012-3)
  • game 4 (http://www.trmph.com/havannah/game/hvc2012-4)
  • game 5 (http://www.trmph.com/havannah/game/hvc2012-5)
  • game 6 (http://www.trmph.com/havannah/game/hvc2012-6)
  • game 7 (http://www.trmph.com/havannah/game/hvc2012-7)
  • game 8 (http://www.trmph.com/havannah/game/hvc2012-8)
  • game 9 (http://www.trmph.com/havannah/game/hvc2012-9)
  • game 10 (http://www.trmph.com/havannah/game/hvc2012-10)
In game 10 the delusional move was black_5 (move 10).
Title: Re: 2012 Havannah Challenge Match
Post by browni3141 on Oct 20th, 2012, 11:56am

on 10/20/12 at 06:39:41, christianF wrote:
That was great fun :)

The next second game I chose the wrong strategy, plain and simple. You can't outrace a bot when it indicates readyness to engage. I thought I had the decisive tempo. Mea culpa.

Was it the wrong strategy because it was against a bot, or because it was just wrong period? Don't you have to trust yourself? In Arimaa I try to play what I think are the best moves no matter who my opponent is. I almost never adjust my strategy based on the opponent, at least not purposely.
Title: Re: 2012 Havannah Challenge Match
Post by christianF on Oct 20th, 2012, 3:28pm

on 10/20/12 at 11:56:26, browni3141 wrote:
Was it the wrong strategy because it was against a bot, or because it was just wrong period? Don't you have to trust yourself? In Arimaa I try to play what I think are the best moves no matter who my opponent is. I almost never adjust my strategy based on the opponent, at least not purposely.
It was wrong because I had counted wrong. In retrospect I might have known. I thought I had a decisive tempo and that turned out not to be the case. I did 'adjust' in that I changed to a 'double enclosure' strategy after the second game.


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