Quest Decks Discussions

[mtgrares]

Chris said:
I would like to thank Rares for the quest mode and we realize how hard it can be for just a single person to take on the responsibilities for the entire project.

Well I thank dennis and rob and everybody else who has contributed code and for adding new cards. Everyone loves using fresh cardboard :--) And about the changes to the quest decks files, it sounds fine to me. (I'm not really the owner of questData and I don't mind other people making changes.)

Maro (Mark Rosewater) has some of the best and worst articles. His recent article on lands was great but his other "only pictures" articles or long dialogues (back and forth conversation between two people) are just stupid.

I agree with rob.

the pool should be drawn from at least a single "set" which is designed with a specific distribution of cards by color and type. So pretend you're MaRo for a moment, and imagine that you're designing a set with 350 cards. 50 cards in each color, 50 multicolor, and 50 cards split between artifacts and lands. Each color should be represented by 25 creatures and 15 sorceries and 10 instants. Artifacts/lands represented by a 30/20 split... so choose either the original dual lands, the ravnica duals, or the pain lands, plus the man lands and other assorted lands.

350 cards is just a round number and would be a good starting point. This discussion is similar to "what is a good cube". Cube is where you make up your own set and you include Magic's most powerful cards. (A good intro to cube drafting is here and you can read about more cube decks here.)

Having multiple sets would be nice but practically I don't see this in the near future. Although I can easily change the quest data so it reads from different files, so drafting and sealed is different from questing.

[Rob Cashwalker]

I agree with rob.

350 cards is just a round number and would be a good starting point. This discussion is similar to "what is a good cube". Cube is where you make up your own set and you include Magic's most powerful cards. (A good intro to cube drafting is here and you can read about more cube decks here.)

Thanks for the support! Exactly... a cube, and it can grow as more cards become available, but it should always grow equally across colors, super-type and rarity.

[GandoTheBard]

I agree with rob.

350 cards is just a round number and would be a good starting point. This discussion is similar to "what is a good cube". Cube is where you make up your own set and you include Magic's most powerful cards. (A good intro to cube drafting is here and you can read about more cube decks here.)

Thanks for the support! Exactly... a cube, and it can grow as more cards become available, but it should always grow equally across colors, super-type and rarity.

I did recommend exactly this in the first Quest mode post though I had no notion of the "Cube format". It doesnt surprise me that Ben had a version of this as he has always been intrigued by this stuff.

I don't know about growing it. Again I strongly support the idea of multiple sets. I realize that MTGRares is set dead against it on some principle that I don't understand but I am certain it is codable. I mean even microprose got that part right. But one of the major problems we still face despite the prodigious efforts of Dennis, Chris and Crew is that a lot of "top50" cards are either uncoded(uncodable?) or broken in some very important way. For one thing without a proper mana pool and proper phase stops the game is still not quite magic. It is similar and does a good job of simulating it (Thanks Rares for starting this thing and for sticking with it despite boredom/personal ennui. and Thanks Rob for pushing forward some important fixes).

Since it isn't actually Magic alot of the best cards do not quite fit the top 50 of MTG forge for this purpose. (All the tappers for example, Icy Manipulator on down, do not have a real function yet.)

[mtgrares]

I'm working on my set editor so that you can see statistics. This will make balancing draft/sealed and the quest cards that you win easier. The statistics be similar to below, but with better formatting

Code: Select all

             Creatures    Spells   Total
Green        10            5           15
White        10            5           15
Blue        10            5           15
Red        10            5           15
Black        10            5           15
Artifact        10            5           15
Gold        10            5           15
Land                                     20
Total      50             25          300

             Common  Uncommon   Rare       Total
Green        10            5        15       30
White        10            5       15        30
Blue        10            5        15        30
Red        10            5        15        30
Black        10            5        15      30
Artifact        10        5       15        30
Gold        10            5        15        30
Land                                 20        30
Total      50             25        200       300

[Rob Cashwalker]

Hey - don't double-post!

[Chris H.]

I would like to at this time suggest that I (meaning me! ) add the following decks to the existing decks in the questData04-08-2009.zip archive:

Rogue = Burning Dragons
Catwoman = Cats
King Edward I = Little Ellie
King Edward II = Ellie
Storm = Lucy Rainbow
Terminator = Machines Mind
Cyclops = Sally
Tarzan = Tarzan
Magneto = Vengence Ltd
Scotty = Yo Yo

I would like to replace Rares version of the battle of wits deck with Gando's version.

I have not yet started, other matters are keeping me busy at this time.

I had a thought. I think that gohongohon mentioned mana acceleration in a post and it got me thinking.

Ultimately, we may want to have several pools of AI decks to play against. While it may be hard for us to create and rank enough decks for five groups, we may be able to work up to say, two or even three groups.

As such, we need more early and middle quest decks. I'm considering not having any mox in these two groups. I plan to replace the mox with the appropriate diamonds. What do the other "deckers" think about this, Gando, Vecc, gohongohon?

[GandoTheBard]

Yeah we can probably do some teching out of the decks to make them weaker and stronger based on presideboarding, removing/putting in strong cards (wrath of god, swords to plowshares, etc) and tweaking the mana sources. We also should decide whether or not we are going to stick with the land weaving model that Rares started us on. Imho the decks DONT need mana weaving. We have some very strong decks that almost play themselves without any need for the AI to "outthink" the human. Sometimes it really is just about the cards and not about the tactics. Though what makes the game fun is the tactical considerations, the strategy (the deck building) is what usually counts towards who wins. With the quest decks the builder is faced with almost insurmountable challenges to start with (unless they get a GOD pool). It seems highly unfair and unnnecessary for the computer to start with a hand full of gas and certainty of drawing whatever mana it needs. It is an interesting way to play for sure. But I don't think we need it. Particularly on the easier levels. Cheating stuff should be saved for the ultra hard level where the player has to get lucky, be good and be a good deck builder to come out ahead.

[Chris H.]

I have been meaning to try a full quest without the land threading. I meant to try this the last quest that I just completed. Sigh, I failed to remember to turn off the flag for most of the games.

I wonder if Dennis could set the land threading flag to off in the default setting. Give us a few weeks to try out the quest mode without land threading.

I just made a couple of minor adjustments to the Tarzan deck. I will post the listing and the deck file to quest decks forum in a moment.

I will also take a look at the "R2-D2 = AI-Black-Vise" deck. I'm sure that I can find a way to tune this deck down. It won't be easy. I think we all get in the habit of creating the strongest deck that we can for ourselves. These skills do not transfer over to creating play-balanced AI decks for quest mode.

Two different skill sets.

[mtgrares]

About the mana threading I should be able to make the checkbox stay checked or not checked through multiple games (by making the variable static.) I should be able to post a minor update of MTG Forge with this change in a few days.

I was thinking about how to generate a random yet balanced card pool for the quest mode. (I'm talking about the cards that you start with and the cards that you win.) I'm going to talk about booster packs since everyone knows that they have 15 cards. My theory is to create a cardpool with 4 cards of each color making a total of 20 cards and then choosing 15 cards from that. This way you will have at most 4 cards of one color and each color should be fairly even and yet a little random. (And of course add a few artifacts, land, and gold cards.)

Updated:fixed crappy math

[Chris H.]

About the mana threading I should be able to make the checkbox stay checked or not checked through multiple games (by making the variable static.) I should be able to post a minor update of MTG Forge with this change in a few days.

Ahh, this indeed would be a nice thing.

[Rob Cashwalker]

My theory is to create a cardpool with 4 cards of each color making a total of 60 cards (4 * 15) and then choose 15 cards from that. This way you will have at most 4 cards of one color and each color should be fairly even and yet a little random. (And of course add a few artifacts, land, and gold cards.)

Um, 4 cards from each color would be either 20, (5 colors) 24, (5 colors and Multicolor) 28, (5c, Mc and artifacts) or 32 (5c, Mc, Artifacts and Lands).

Your set editor post with the stats is useful. A real set doesn't have precise quantities of creatures vs spells, but each color is represented equally.

[mtgrares]

Thanks for pointing out the math. It made perfect sense when I wrote it, lol. Hopefully the set editor will help with the balancing issues but it is a little boring trying to balance a set.

[Chris H.]

The decks that I have posted to the Quest Decks only topic at http://www.slightlymagic.net/forum/viewtopic.php?f=48&t=967 can be imported into the constructed deck Deck Editor and the quest deck Deck Editor. You should be able to play these decks in either format if you chose.

At some point I will release an updated questData file which will include all of these quest decks.

[Chris H.]

Gando,

I have looked at the cards included in the "Data" deck. And I compared this to what we already have in the other quest decks

The "Data" deck includes Baru, Fist of Krosa and this card is already represented in the "Silver Surfer" deck.

The "Data" deck includes Tarox Bladewing and this card is already represented in the "Boba Fett" and the "Comic Book Guy" decks.

The "Data" deck includes Nightmare and Korlash, Heir to Blackblade and these cards are already represented in the "Wolverine" deck. The Wolverine deck has

1x Korlash, Heir to Blackblade
3x Nightmare

I could delete the "Data" deck from the upcoming fourth version of the questData file.

Or we could perform a major rewrite of the "Data" deck. We could replace all of the cards in the "Data" deck except for the 4x Korlash, Heir to Blackblade cards. I could use your help here.

Would you like to redo the "Data" deck?

[mtgrares]

Not to complicate things too much but this version is based on the 04/24 version I posted. This version has the title "beta" and it divides the quest decks into easy, medium, and hard. The file names are questDecks-easy.txt, questDecks-medium.txt, and questDecks-hard.txt These files just hold the deck names. I suggest that you cut-and-paste the deck names. If there is a problem with a deck name (and there will be because even I made a mistake) an error message will show on the screen telling you where the program is at.

The files questDecks-easy.txt, etc... must have at least 3 decks names and can contain duplicate deck names. The quest files (questDecks-easy.txt, etc..) have just one random deck in them listed three times. I don't know which decks are easy, so you will have to make that decision and then add it. The quest files are required and the program won't run without them.

I say this is a beta version because I haven't done much testing but things seem to work. On easy mode you play 3 easy decks, 3 medium decks and 4 hard (I just guessed that these numbers). On medium you play 6 easy, 6 medium, and 8 hard. (I may be wrong about a few of these numbers but they are at least close. I wrote the code but decoding the code into English can be complicated.)

Also in order to make the medium, hard, and very hard easier, you will get new cards after 2 wins or 2 losses (the wins or losses don't have to be consecutive.)

Have fun and I'll check back on Tuesday.

Download Beta

I'll notify DennisBergkamp about these changes once we figure out how to divide up the decks and know that the code is working.