Apr 19, 2012 -- 1:18AM, Leftconsin wrote:

Why not make a post to try to explain why this process is beneficial over intelligent design?

Well it's not, since there's no way you can make it evolve faster than a metagame, even if you could do it fast enough to get a strong deck.  It's more just to say you can do it, I'd imagine.

I think it's interesting.  It's worthless, but it's neat to see genetic algorithms at work.

I've had the idea of a genetic cube (every time it's drafted the last picks are replaced by random cards), so this isn't too far off.

LEFTCONSIN:

Id' prefer to call it emotional intelligence because humans are more emotionally involved in the cardpicks than logical. Intelligent design is reserved for "heavier" entities

And there's only one real good reason for running something made by evolution, and that is because it picks something humans would never think of, like fungosaur+unstable mutation or orcish oriflame+prodigal sorcerer. With other words it may come with some fun and refreshing ideas.

ADEYKE:

I play both sides because no computer can play magic at that level. During games I set up rules for the behavior of playing specific cards so I can avoid being biased as much as possible, and I have been doing this a lot and have trained myself to be as neutral as possible regarding the designs.

I have run such projects with populations as low as 20 decks, and in the oldest project evolution managed to produce something that killed at an average of turn 6, while the core deck started with an average of killing turn 10. (Against goldfishes, just for a measure of speed) That project achieved that within 10 generations, but the cardpool had a lot of good cards mixed in between it.

It was not hybrid cards alone, it was including all common multicolored, and artifacts.

How the spawning of new decks happens will probably be covered by the old articles. If not, yell like a troll and I'll try to go into the details with you.

SLEETFOX:

I have been trying to figure out how fast I can run such projects, just to see if I could create something in standard, but so far I'm involved with another project using the top 500 legacycards that was listed at a time. The reason for running it is to study rates of mutations and how many mutations will be best. But one day I will try with standard, based on my experiences with projects like this.

MOIMOI:

For the sake of speeding up this type of projects I chose to go with 4 of each a long time ago, but this is something I might investigate one day, seting up a 4 of each vs 3 of each project just to get the damn answer of what we've discussed so much

I used a cardpool (Which will be mentioned in the article) which I wrote into a program that would print out cards at random so whenever there would be a mutation I would get the location (set 1-15 in the deck) and a cardname. The 4 cards at that location would then be replaced by the randomly chosen card. Just for the sake of doing it right I went for ALL the mutlicolors and artifacts involved, even the crappiest of them all. But details will be in the main-post.

If not yell like you use to and I will vaguely try to answer ALL your questions, and probably fail at it

MTGKAOSHIN:

The winning deck "spawns" it's offspring. No problems with orcs

CHAMALE:

I was interrested in how much blind evolution could achieve more than in taking it to tournaments. I did some tournaments just to try it out and lost a lot. Reactions from players I met was quite funny when they were told that the deck was made by evolution. I had to explain the process a lot of times throughout that day...