By wicked darkman on Monday, June 05, 2006 - 07:33 pm:

This is for the drafting people.

Im not sure it will be very helpfull, because there is huge amounts of randomness in drafting, and even the change of one card can break the balance of a manacurve deck. But here it is anyways.

After 678 cycles I have some nice results...

Remember the key to read the manacurves:
lands, 1cc, 2cc, 3cc, 4cc, blanks...

18 7 8 5 2
17 8 8 5 2
18 6 8 6 2
20 4 8 4 4

And some less powerfull ones...
16 10 6 6 2
18 10 4 6 2
18 8 8 4 2

I have focussed on those manacurves that can be split into two equally large sections, so it will be easier to focus on two colors...

If you find it usefull I will bring you more tomorrow...

By Homelands Bookie (Psycho) on Tuesday, June 06, 2006 - 07:33 am:

I like drafting, so you got my attention. However, I do think I've never used so much land as your key suggests.

By Dark Lord (Apprentice) on Tuesday, June 06, 2006 - 09:38 am:

This is quite interesting, but there has to be a way to hone results to make them more flexible/accurate.

Usually I use many 1-2 CC creatures, and a few fatties that cost 5-7 mana. I dont know how you would engineer that tho.

-Apprentice

By wicked darkman on Tuesday, June 06, 2006 - 04:02 pm:

I can change the parameters in the program tio both a minimum and maximum of how many of each card the simulation may use.

If the drafting people can come up with those limits I can run the simulator within them and get the best results inside the limits. Remember to set the limits as flexible as possible if you want many different manacurves. But limit control can also bring focus into the manacurve. Like I am trying to do with a 18land/60 cards/12 blank scenario... (look at the second sphere my top deck)...

APPRENTICE:
Im sorry about the fatties. everything above 4 is too slow for me currently. One day I will probably go that way too. It looks like I'm restricting you as much as you want me to restrict the simulator. That's ironic...

By wicked darkman on Tuesday, June 06, 2006 - 06:26 pm:

There is a reason for as to why I excluded cards costing above 4...

When I started the projekt I figured out that 1/1 and 2/2 critters were able to kill at turn 6. So I decided to include everything able to do that. Both 3cc and 4 cc kills at turn 7 so I decided to include them anyways. But turn 8???

I later learned that it was possible to kill at turn 5 with normal creatures. Something I had never discovered without the simulation. first time I saw that a curve had killed at turn 5 I thought there was a bug in the system...

So the fact that if you each turn cast a land and an amount of creatures with a total power equal to lands in play gives a turn 5 kill was what drove me forward from then on...

5cc and above does not best that speed, so I excluded it. The fact that some creatures at that cost have abilities that could lock down opponents I have arrogantly ignored. Besides, most people like to say that in t1 you really shouldnt use manacosts above 3 so the fact that I included 4cc has already crossed 1 borderline...

By Dark Lord (Apprentice) on Tuesday, June 06, 2006 - 09:54 pm:

Yeah, but the situation is entirely different in draft. Usually you draft a lot of small creatures then 2-4 gamewinners that cost more then 5 mana.

-Apprentice

By wicked darkman on Tuesday, June 06, 2006 - 09:59 pm:

cycle 1623:

16 9 6 6 3
16 7 7 6 4
16 6 8 5 4 1
17 7 7 6 3
18 6 8 5 3
16 7 8 5 3 1

And the splitable manacurves...

18 8 4 8 2
16 8 6 6 4
18 8 6 6 2
16 8 6 8 2
18 6 8 6 2
18 8 8 4 2

You may be wondering why some of them are similar. It is because evolution is slowly crawling towards better and better manacurves, and some decks are only fractionally better than others. But compare these newer results with the ones on the top and you can see the difference...
None of them have been speedtested so you just have to pick the ones speaking out to you the most and hope for the best.

This time there are blanks, which means that I got these samples at a time where the simulation is switching to a new and improved kind of manacurve...

I am thinking on buying a portable computer so I can have two computers running at the same time... This will help researchspeed...

By Dark Lord (Apprentice) on Tuesday, June 06, 2006 - 10:04 pm:

I guess I can take these to a draft, draft my cards and then assemble them to one of the deck parameters.

Can you make it so that every deck has 8 blank spots for spells?

-Apprentice

By wicked darkman on Tuesday, June 06, 2006 - 11:01 pm:

Easilly...

Does anyone disagree with this number of blank spells ???

Otherwise consider it done...

By Dark Lord (Apprentice) on Wednesday, June 07, 2006 - 11:27 am:

It is pretty much accepted that the standard amount is:

16 Lands
16 Creatures
8 Spells

-Apprentice

By wicked darkman on Wednesday, June 07, 2006 - 08:21 pm:

Okay, the last manacurves (two different cycles) before swithching parameters can be read in the post

Research: guess a decendant...