Author |
Topic: Essay by Christian Freeling on inventing games (Read 522350 times) |
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender: 
Posts: 804
|
 |
Re: Essay by Christian Freeling on inventing games
« Reply #900 on: Mar 11th, 2012, 5:21am » |
|
on Mar 10th, 2012, 1:27pm, ocmiente wrote:
Yes, it's been mentioned here before. |
|
That was Luis too, I had missed it, thanks. 
Meanwhile, in the wake of the one for HexSymple, we've made an applet for HexSygo.
I've played a game of HexSymple now and it offers slightly different tactics, tightly interwoven with an obviously deep but somewhat less clear strategy.
A square grid offers more strategical grip and in my experience makes 'visual thinking' easier, at least where territory games are concerned. In connection games it may be different and for me the connectivity issues in HexSymple were less of a problem indeed than staking out territory.
|
|
 IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #901 on: Mar 21st, 2012, 2:20pm » |
|
To keep an eye on its evolving strength towards the 10 game match this summer, I challenged Castro_bot to a base-10 havannah game, me playing second, no swap.
It's still no genius in strategy, but its tactics seem to have grown somewhat more circumfencial and possibly a bit nastier - I was offered a nice selection of well camouflaged mistakes to make, but if it doesn't get much trickier I'm cool with it  .
|
| « Last Edit: Mar 21st, 2012, 2:22pm by christianF » |
IP Logged |
|
|
|
omar
Forum Guru
Arimaa player #2
Gender:
Posts: 1003
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #902 on: Mar 22nd, 2012, 3:09pm » |
|
on Mar 21st, 2012, 2:20pm, christianF wrote:To keep an eye on its evolving strength towards the 10 game match this summer, I challenged Castro_bot to a base-10 havannah game, me playing second, no swap.
It's still no genius in strategy, but its tactics seem to have grown somewhat more circumfencial and possibly a bit nastier - I was offered a nice selection of well camouflaged mistakes to make, but if it doesn't get much trickier I'm cool with it . |
|
Nice win. I have also been keeping an eye on bot Castro and have played it several games this year. It is getting quite strong and I fell for some of it's traps. My win rate is only about 50% against it.
Are your official games going to be postal or live? Is it going to be played online or offline? It would be great if you could post a Havannah challenge page on your site with details.
Also I would like to do a live interview with you maybe a few days before the match and also again after the match is over. Would you be up for it?
|
| « Last Edit: Mar 22nd, 2012, 3:13pm by omar » |
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #903 on: Mar 23rd, 2012, 9:17am » |
|
on Mar 22nd, 2012, 3:09pm, omar wrote:| Are your official games going to be postal or live? Is it going to be played online or offline? It would be great if you could post a Havannah challenge page on your site with details. |
|
If I had 'm . Ton van der Valk is organizing the event and I'm not involved in any way. You can contact him via http://www.hexboard.com/ (or maybe you already did).
on Mar 22nd, 2012, 3:09pm, omar wrote:Also I would like to do a live interview with you maybe a few days before the match and also again after the match is over. Would you be up for it?
|
|
Live like face to face or via Skype or something? In either case, sure, why not .
P.S. Richard Lorentz' Wanderer is more of a one trick pony, juggling ringthreats in and out of season in our current game. Remarkably, keen as it is on creating them, spotting them seems less of a priority. 
|
| « Last Edit: Mar 23rd, 2012, 2:47pm by christianF » |
IP Logged |
|
|
|
SpeedRazor
Forum Guru
Skepticism is the Immune System of Science
Gender:
Posts: 72
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #904 on: Mar 23rd, 2012, 8:31pm » |
|
Quote:Omar:
Also I would like to do a live interview with you maybe a few days before the match and also again after the match is over. Would you be up for it?
|
|
This is a very good idea, Omar  Actually, it Rocks!
|
|
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #905 on: Mar 24th, 2012, 4:22am » |
|
on Mar 23rd, 2012, 8:31pm, SpeedRazor wrote:
This is a very good idea, Omar Actually, it Rocks! |
|
No, the Foo Fighters rock!
Ten years ago, Chess programming was approaching its current status of grandmaster level. This seemed to end an old debate. Programs would soon be better at Chess and by implication would eventually be better at any abstract game. Or so mainstream concensus dictated.
Why did that bother me? Come to think of it, it must have been the teacher in me, or worse.
Let's leave Chess out of it and ask a different question.
Havannah and Explocus are both simple games.
Havannah is very difficult to program. I know of no significant program that uses the traditional method of evaluating positions and running them through an alpha-beta search. In terms of the challenge, Monte Carlo 'came to the rescue'. But Monte Carlo has inherent limitations and additional heuristics are required to improve its efficiency.
Yet Havannah is easy to read for humans.
Explocus is equally simple, but hard to read for humans. I expect this to be a game where both methods would work (the traditional one certainly does), and both would render programs that would be all but invincible from a human perspective.
And without all that much effort.
So even simple games may show huge differences in programmability, and many are the reasons why, but to my knowledge none of them pinpoints a fundamental difference, if any, between programs and humans. 
And that's funny.
|
| « Last Edit: Mar 24th, 2012, 3:27pm by christianF » |
IP Logged |
|
|
|
omar
Forum Guru
Arimaa player #2
Gender:
Posts: 1003
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #906 on: Mar 25th, 2012, 10:54am » |
|
Yes, I did get in touch with Tony back in January of this year, but he didn't have any specific information other than it would be sometime in the second half of the summer of 2012. I think planning and organization of this event should be done soon.
I'll contact Tony again to see if he has any more info.
BTW, how was Tony selected to organize the event?
|
| « Last Edit: Mar 25th, 2012, 10:55am by omar » |
IP Logged |
|
|
|
omar
Forum Guru
Arimaa player #2
Gender:
Posts: 1003
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #907 on: Mar 25th, 2012, 11:05am » |
|
on Mar 23rd, 2012, 9:17am, christianF wrote:
Live like face to face or via Skype or something? In either case, sure, why not .
|
|
Thanks Christian. I was thinking of doing it through the Arimaa radio page:
http://arimaa.com/arimaa/radio/
You would need to install the TeamSpeak client on your computer, instructions are here:
http://arimaa.com/arimaa/gameroom/teamSpeak.html
|
|
IP Logged |
|
|
|
omar
Forum Guru
Arimaa player #2
Gender:
Posts: 1003
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #908 on: Mar 25th, 2012, 11:09am » |
|
on Mar 23rd, 2012, 8:31pm, SpeedRazor wrote:
This is a very good idea, Omar Actually, it Rocks! |
|
Yes, I think this is an important event which should be recorded.
|
|
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #909 on: Mar 25th, 2012, 12:58pm » |
|
on Mar 25th, 2012, 10:54am, omar wrote:BTW, how was Tony selected to organize the event?
|
|
He volunteered and I was only too happy to accept. Ton has a genuine interest in hexgames and is a brilliant player. He also has a far bigger network than I have, though I introduced him to some additional people now involved in the organization, like the games club "Fanaat" (Fanatic) at the University of Twente.
on Mar 25th, 2012, 11:05am, omar wrote:
Great, even works on my old OSX Mac. But I don't think an interview from my homeplace would be all that interesting (though I could wrap Kobus around my neck ). More appropriate would imo. be to do it before or after a game in the Fanaat gameroom, so we would have to install it there. That'll be no problem, I'm sure.
|
| « Last Edit: Mar 25th, 2012, 12:58pm by christianF » |
IP Logged |
|
|
|
omar
Forum Guru
Arimaa player #2
Gender:
Posts: 1003
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #910 on: Mar 30th, 2012, 11:27am » |
|
Good to hear you were able to get TeamSpeak working. I have some questions regarding the details of the challenge match. But I'll start a new thread for that.
|
|
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #911 on: Mar 31st, 2012, 10:39am » |
|
We (that is: Ed) upgraded the Symple applet to show the penalty, the # of stones and groups and the score.
If this particular score appears uneven, please not that the big white has 28 vacancies left, the biggest blacks both 7 (in the bottom one, one will be taken by white).
If white refrains from invading black can plug 3 groups the next 4 moves, then 2 groups for a further 3 moves, that's 18 points.
In these 7 moves, white can plug 8, that's 8 points.
So the total difference up to that point will be black's current 70 points lead, plus another 10 (the 18-8 )
At that time black must start invading with still some 21 white holes to plug, say ten each, with penalty 8 that's 80 points penalty for black ...
Thrillsville!
cly ring (US) - christian freeling (nl).
|
| « Last Edit: Mar 31st, 2012, 10:40am by christianF » |
IP Logged |
|
|
|
clyring
Forum Guru
Arimaa player #6218
Gender: 
Posts: 359
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #912 on: Mar 31st, 2012, 9:15pm » |
|
The score is indeed not very uneven and the side that looks 70 points behind actually wins.
Your count of black leading by 80 points and then losing exactly that many to the penalty is obviously incorrect since symple is drawless- there were two errors in your counting.
1) White earns one point from growing their K5 group inside the black territory on their current move.
2) The big white group is actually two groups separated at F16 and G15. White will earn 8 points more than you predicted by connecting them on their current move, making the final score a white win by 9 points on the line you give.
However, that line is suboptimal for black (!) and with correct play on both sides from the position you show, white only wins by 5 points. I'll leave you to figure out exactly why, but I will give you a hint: Your very next move, black 35, was a mistake.
EDIT: I'm also willing to perform a chain undo to that point if you want to use it as an example game.
|
| « Last Edit: Mar 31st, 2012, 9:25pm by clyring » |
IP Logged |
I administer the Endless Endgame Event (EEE). Players welcome!
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #913 on: Apr 1st, 2012, 5:27am » |
|
on Mar 31st, 2012, 9:15pm, clyring wrote:The score is indeed not very uneven and the side that looks 70 points behind actually wins.
Your count of black leading by 80 points and then losing exactly that many to the penalty is obviously incorrect since symple is drawless- there were two errors in your counting. |
|
For me that's good  . I hate analysis because for me thinking from the general to the specific is like riding a bike uphill, while generalizing away from the specific is something I can do freewheeling handsfree.
Thus I became an inventor rather than a player
In this case I had a feeling of impending doom, but I just wanted to show that the 'score' displayed during a game can be very misleading to the casual observer. I think the game illustrates that splendidly - no lack of drama!
on Mar 31st, 2012, 9:15pm, clyring wrote:1) White earns one point from growing their K5 group inside the black territory on their current move.
2) The big white group is actually two groups separated at F16 and G15. White will earn 8 points more than you predicted by connecting them on their current move, making the final score a white win by 9 points on the line you give.
However, that line is suboptimal for black (!) and with correct play on both sides from the position you show, white only wins by 5 points. I'll leave you to figure out exactly why, but I will give you a hint: Your very next move, black 35, was a mistake.
EDIT: I'm also willing to perform a chain undo to that point if you want to use it as an example game. |
|
That won't be necessary, but I appreciate the analysis of a player and I hope it will draw more interest in this strategically mysterious game. Thanks for the game and congrats with a victory well deserved .
P.S. I'm starting to get to grips with Marcel's bot too and even with my slow learning curve I think I will start winning on a regular basis soon.
|
| « Last Edit: Apr 1st, 2012, 7:31am by christianF » |
IP Logged |
|
|
|
clyring
Forum Guru
Arimaa player #6218
Gender:
Posts: 359
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #914 on: Apr 1st, 2012, 8:25am » |
|
Thanks. I understand your not wanting to do analysis, but if it were 'just' analysis, I would do it myself and tell you instead of asking that you do it yourself. My goal for the endgame has largely been to find the general by looking at the specific. In this case, it illustrates an endgame concept that I don't think you understand yet- one that allows you to make a 4-point mistake even so late in the game that your intuition would tell you that you have no need to think. Nothing is ever that symple...
Fortunately, in this position you only have two distinct options- placing in my territory and growing in your own territory. I've told you that growing is not correct- I think it's best if you find out why placing in my territory for the next 14 moves should be better on your own. You just might learn something.
|
| « Last Edit: Apr 1st, 2012, 8:27am by clyring » |
IP Logged |
I administer the Endless Endgame Event (EEE). Players welcome!
|
|
|
|
Home
Help
Search
Members
Login
Register
Notify of replies
Send Topic
Print