Author |
Topic: Essay by Christian Freeling on inventing games (Read 522079 times) |
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender: 
Posts: 804
|
 |
Re: Essay by Christian Freeling on inventing games
« Reply #495 on: Jul 3rd, 2011, 9:37am » |
|
We've put Nick Bentley's game Ketchup in the Pit at mindsports. There's no applet yet because we've developed a substantial pipeline filled with competing priorities. But we're working on it.
P.S. Sometimes people summarize ideas accurately and concisely, like J. Mark Thompson in his 'Defining the Abstract'. I'm a great admirer of that capability, that's why I want to emphazise Nick's summary of his approach of Ketchup:
Quote:| "I was trying to design a game which is short, intuitive, unintimidating, addictive and deep. The difficulty with these criteria is that the last one is often in conflict with the first four. Deep games tend to be hard to figure out, and that quality can make them intimidating and not-addictive. But Ketchup avoids that trade-off better than most of my other games." |
|
In the process he mentions a couple of drawbacks that J. Mark Thompson fails to mention, not so much of "depth" itself, but of the way it may present itself to a novice. I can well remember my first encounter with Go, in the late sixties (the stone age in terms of spreading a game). 'Hard to figure out' and 'intimidating' certainly came to mind.
I agree with Nick that Ketchup stikes a fine balance between depth (as yet unmeasured, here's where a good intuition may be required) and accessibility.
|
| « Last Edit: Jul 4th, 2011, 5:52am by christianF » |
 IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #496 on: Jul 6th, 2011, 8:00am » |
|
A good summary is one thing, a good metaphor is another. In trying to explain the significance of Sygo, please envision the appearance of the first helicopter.
Balancing the basic a-symmetry of Sygo's first player advantage by an a-symmetry that runs counter to it, and that is embedded in (and thus provided by) the mechanism itself, is like the tail rotor of a helicopter. Not only is the flight mechanism different (that is: the move protocol), but the same mechanism is used to balance the inherent deviation.
Apart from its working perfectly, I find that aesthetically pleasing.
Sygo is fast, so a new game between RY and me is well underway. Note that RY pushes the button on his second move, which I consider too early.
However, you can't create eyespace by capture in Sygo, so you generally need more 'body' to live. RY's invasion of the upper side, combined with the black stone on P12, forced me to invest in defensive placements in both upper corners. Now his center influence looks promising. My last move was 7. G8, strengthening the left side and eyeing the center.
Additional: Black's 8. ... M15 starts a local tactical battle in which I had to grow and attack it, or K14 would be lost. Black cannot afford placement of a single stone either so here we have a full-fledged growing battle.
Additional: Black 11 is a single placement again. Now white, in order to capture the hooked 3-group, cannot grow, because black then could connect with h14 (growing one group) and i14 (growing the other). That would be fairly fatal. So must white play 12. H14 himself.
|
| « Last Edit: Jul 6th, 2011, 10:43am by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #497 on: Jul 8th, 2011, 1:40am » |
|
After 17 turns in our latest game, the cake was divided, and white is a couple of points behind. Bummer.
However, black's top group surrounds a 3x5 area. Certainly enough to get one eye, but if I can construct a seki in the other, I might just have enough to win.
Maybe black can capture the invading stone(s) while keeping enough eyespace, but it's worth a try, hence 18. I18.
Edit: On his 17th move, black should not have connected at H16, but instead should have grown both groups separately, for instance at G18 and J17, to create eyespace. As it is he has a bad shape for eyespace and only one growing option.
You can't get eyespace by capture, so building 1-point eyes is important. Eyes consisting of more points may be prone to 'seki-invasion'. Consider a group with a 1-point eye and a 2-point eye, then the latter can be invaded by the opponent. His stone cannot be captured, nor can it capture, so the opponent grabs one point and spoils two (because the vacant point in the seki is neutral). That's the general idea behind seki invasions.
So considering the score this should be interesting.
|
| « Last Edit: Jul 8th, 2011, 10:11am by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #498 on: Jul 9th, 2011, 3:02am » |
|
game
With 19. ... J18 Black threatens an eye at K19 with 20. ... J19.
If white moves 20. J19 himself, black can make a 'seki-eye' by moving 20. ... I17, because a white move at K17 would then be suicidal. Given that the remaining part of the area would now bring white no more than a seki too, white decided to allow the eye at K19 by growing both white stones. Now, if white can turn the remaining area into a seki, which appears possible (but I'm not a very good player and this is uncharted territory), he does so with at least an extra white stone. Immediate capture of the two white stones would prevent black from completing the eye, and would thus probably lose.
The growing move also allowed white to strengthen his 'shape' elsewhere, in particular in the top right section, to prevent invasion (whether it would actually have a chance of success or not - it's uncharted territory so I'd rather be on the safe side).
Additional: after white 21, black needs both G18 and H17 to make a second eye, so white's invasion succeeded. The seki points are just enough to tip the scales.
P.S. Luis Bolaños Mures, Luigi at RGA, directed me to a page at Sensei's Library about the othellonian method of capture in Go, its advantages and problems.
The main problem is that you can't create eyespace by capture, that's why Luis' own new variant Goncrete introduces a simple but somewhat artificial mechanism to remedy that.
That in itself isn't new either: othellonian Go variants like Medusa and Lotus also employ an artificial 'life insurance'.
Sygo escapes the need for anything artificial because a player can combine several moves in one turn to create eyeshape, inherently reducing the problem to proportions that fit the general strategy and tactics.
|
| « Last Edit: Jul 9th, 2011, 9:29am by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #499 on: Jul 9th, 2011, 11:13am » |
|
game
Unfortunately for white the position had a detail I had overseen: black's access to H19, making G19 his second eye. So white 22 is a formal move of recognition, and sad as it is to lose the game, the tactics were hopefully instructive.
As usual I got only myself to blame. Instead of looking better I assumed I had it figured out, and I hadn't.
I could have saved the seki by not connecting on move 21 (I17), but moving at G18 instead, at the risk of losing the two stones at J/K17.
Greed blinds, I feel like Dumb & Dumber all in one  .
Yet a nice beginner's game, fast and pretty clear cut in the strategic phase, where black did obviously better. Then he made a mistake at move 18, allowing me to launch a seki invasion, and I blew it at move 21.
-----------
On another note, my opponent surprised me with what I take to be the Taiwanese wiki entry of Congo, a Chess variant made by my son Demian. The game is going on thirty while its inventor is going on thirtyseven.
|
| « Last Edit: Jul 10th, 2011, 4:11am by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #500 on: Jul 10th, 2011, 1:32pm » |
|
I've assembled a few thoughts about 'othellonian capture' in Go variants, a kind of provisional overview open to additions, corrections and/or suggestions.
Please have a Go @ it: Othellonian capture in Go variants.
|
|
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #501 on: Jul 11th, 2011, 3:44pm » |
|
on Jul 3rd, 2011, 9:37am, christianF wrote:| I agree with Nick that Ketchup stikes a fine balance between depth (as yet unmeasured, here's where a good intuition may be required) and accessibility. |
|
We've seen charaterizations like "a flailing tweakfest" here, but as long as there's a real game to discover, tweaks may be the only way to approach a theme and a mechanism that is reluctant to unveil itself.
I thought Nick's last version of Ketchup made a pretty neat and balanced game, and so did Nick, judging from the title.
Fortunately there's some wisdom floating around in the abstract games community like "If you find a good move, look for a better one", and that applies to inventing too.
I admire Nick's intuition and the courage to change what "woudn't change anymore".
A slight change of perspective, a simplification actually, brought to light the more organic core of the game, because the right to conditionally move three instead if two stones now doesn't hinge on a dead count anymore, but on a specific action of the opponent.
So I thought I'd mention the final version (no sarcasm implied, I really believe so):
Ketchup 4.0 - I've changed the rules again. This is just stupid.
|
|
IP Logged |
|
|
|
SpeedRazor
Forum Guru
Skepticism is the Immune System of Science
Gender:
Posts: 72
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #502 on: Jul 12th, 2011, 2:33am » |
|
When I saw your post Christian, my first inclination was "No, Ketchup is [finally] perfect! You can't change the rules, Nick, unless you change the name!"
(Salsa? LoL)
But after reading them, I can't wait to try this variation! In the Depth vs. Clarity debate, Ketchup was very deep ... but not so clear. This simple new rule might even that out some.
In the Decisiveness vs. Drama debate - (earlier Christian post) - I didn't like that Ketchup seemed maybe not so balanced here. Too decisive: I don't want to drop any stones at all, until I can see the victory. Think Nim. This new rule might add some Drama to the game (and time on my clock!)
Also, incidentally, it might be construed that it was not a rule added, but taken away.
"Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away."
|
|
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #503 on: Jul 12th, 2011, 7:17am » |
|
on Jul 12th, 2011, 2:33am, SpeedRazor wrote:| "Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away." |
|
That's a nice quote and a deep truth. The implication is that there must be something to "take away from" in the first place.
I think Ketchup is as interesting as its process of invention. Some games pop up ready. I won't bore you with examples in my own career, but it is clear that Piet Hein and John Forbes Nash came up with the same Hex because there is only one Hex (for variants, consider the quote).
Other games may be far more reluctant to reveal themselves, either in their spirit or in their mechanics. Few doubt the spirit of Go, but after three milennia or thereabouts, people are still tweaking at its fringes to have it behave properly, even if the players won't.
There are many inventors and they make many mistakes (yours truly not excluded) and consequently there's a lot of tweaking going on.
If a notion of a new game, in terms of a theme and mechanics, leads to a solid core once "there's nothing left to take away", then tweaking may lead to that core.
Or it may not. Star, Superstar, *Star and YvY were all off the mark, and if Benedikt hadn't literally talked me into finding the core of the group penalty theme, Symple would probably not have existed. Note that Symple emerged because I took something away! As did Emergo, for that matter.
If a notion of a new game is wrong, and there's nothing to tweak towards, then you may end up with Frankenstein's monster: all seperate parts and no life of its own.
That's why tweaking has a bad reputation.
Nick's game leaves little doubt about his intuitive notion: he definitely sensed a beautiful new game, a natural organism - only thing was how to capture it, and this has been a reluctant cookie to say the least.
Now that it's all cleaned up and polished, we have a new definition of territory and new mechanics to match, and a beautiful game awaiting indeep investigation.
|
| « Last Edit: Jul 12th, 2011, 8:50am by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #504 on: Jul 13th, 2011, 10:29am » |
|
Sygo in the Chinese wiki (courtesy of Dandolo).
|
|
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #505 on: Jul 15th, 2011, 6:07am » |
|
We've presented Symple as a possible next AI challenge in terms of game programming, more in particular in terms of Monte-Carlo evaluation and its refinements.
Part of the reason is its branch density, but that's not all, or Medusa could long ago have been presented as such, and Arimaa was also developed with such a feature in mind. But Medusa is a Go variant with an artificial life criterion (natural as it may feel) and is played on a sub-grid of a hexhex board that does not particularly stand out for clarity. And Arimaa isn't a uniform mechanism, which is perfectly justifyable, but not what I have in mind concerning the 'human versus computer' issue. I prefer uniform games along the lines of Go, Hex, or Havannah, to name a few.
Symple and Sygo both comply, but for the reasons mentioned below I prefer Sygo for humans and Symple for bots. Here's the part we added to 'About Symple'. There's also a reference on our homepage (mindsports.nl).
Quote:Is Symple a programmable?
Without Symple's move protocol the game would probably fall into the same category as Go, Hex or Havannah, to name a few. The Monte-Carlo method and its refinements would provide a firm handle, as they do in the games mentioned. However, the Symple move protocol brings with it two new obstacles:
- The choice between placing a single stone and growing all groups does not seem to align smoothly with a search based on random play-outs.
- The branch density isn't of this world.
The setting of the parameter 'P' would seem to be less of an obstacle since it affects only the counting procedure and not the nature of a Monte-Carlo evaluation (though it does affect the strategic evaluation by humans).
Despite the exploding branching factor, for human players the strategical and tactical considerations differ little from those in Go, Hex or Havannah.
We promote Symple as 'the next AI challenge' for abstract games, not only because of its branch density and move protocol, but because of its simplicity.
We consider Sygo the better game for humans, but although if it may pose AI problems that equal or even surpass Symple's, Sygo isn't quintessential and neither is its theme 'group penalty', which is the cradle of the move protocol.
AI programmers should be provided with the essence, and Symple is the essence. Moreover, concerning Symple's relative lack of drama, computers don't care about it, and for humans the drama is provided by the programs' progress rather than the outcome of a particular game.
So we think Symple is THE game to consider for the abstract games AI community in the years to come. Because if they can't do it, humans can do something computers can't, and the question "how do humans do it?" remains a mystery. |
|
And no, I'm not offering a bet this time - if at all, I'd be 75 by then and I'd have to play a lot to get to an acceptable level.
And I prefer Sygo.
|
| « Last Edit: Jul 15th, 2011, 3:00pm by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #506 on: Jul 17th, 2011, 5:44am » |
|
on Jul 15th, 2011, 6:07am, christianF wrote:Speaking of which, here' a new game that I hope to be entertaining.
Edit: That I hope was entertaining. White 17 was a mistake in more than one sense. As a seki invasion it's a three point difference because then too, the black group lives. So its timing would have to be in the endgame - there are more pressing issues to address at this point.
Moreover it wasn't a seki: after M9 and capture at K7, the black group has two solid single eyes.
|
| « Last Edit: Jul 17th, 2011, 12:06pm by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #507 on: Jul 17th, 2011, 9:41am » |
|
Ed has finished and launched the applet of Nick Bentley's new game Ketchup, so you can play this intruiging new game at mindsports now.
|
|
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #508 on: Jul 17th, 2011, 10:04am » |
|
Meanwhile Dandolo's powers of recuperation are admirable, so here's another new game of Sygo you may enjoy.
|
| « Last Edit: Jul 17th, 2011, 12:05pm by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator
Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #509 on: Jul 21st, 2011, 5:20am » |
|
I've played my first game of Ketchup at mindsports and got crushed. Ed's strategy of keeping a large second group and the option to connect it to his main group was in effect before I fully realized my position was doomed.
This is very much a strategy game, as opposed to a tactical one. Of course I will next time try to apply such naive thoughts on strategy as I now have, as every beginner would, and get crushed again at times, no doubt. Tactics will evolve to a higher resolution as strategies deepen. This game will outgrow hexhex5 pretty soon, as far as I can see. It has an excellent balance between strategy and tactics, feels altogether organic, has a natural simplicity and a nice pace.
It is something of a coincidence that Nick and I more or less simultaneously came up with group-based territorial games that feature a new move protocol with an embedded balancing mechanism. None of them are in need of a pie, if at all able to profit from one.
It must have been in the air.
|
| « Last Edit: Jul 21st, 2011, 6:32am by christianF » |
IP Logged |
|
|
|
|
Home
Help
Search
Members
Login
Register
Notify of replies
Send Topic
Print