Quest Mode - Post your AI tuned decks

[mtgrares]

I'm probably going to have around 10 AI opponents, each one will always use the same deck. So you can figure out that Sam plays blue control and Henry likes dragons. (Sam and Henry are lame names, so feel free to suggest better ones.)

For the 10 decks I'll probably just choose stuff that people have posted in the decks forum, feel free to post your AI tuned decks, wackiness and fun count.

[mtgrares]

Forge's first Quest opponent could have 15 Life and the Life of the last opponent might progress to 60 Life.

Shandalar occasionally lets your opponents start with a creature in play, which makes things much harder. Even having Llanowar Elves in play at the beginning is hard and Hypnotic Specter is brutal if you don't have any burn. I may be able to code this later.

Something that would help make difficulty levels meaningful, is if all canPlayAI methods used some randomness to simulate skill level.

Maybe, I hate to gimp the AI any more since it is so bad to begin with, lol.

[Chris H.]

I'm probably going to have around 10 AI opponents, each one will always use the same deck. So you can figure out that Sam plays blue control and Henry likes dragons. (Sam and Henry are lame names, so feel free to suggest better ones.)

For the 10 decks I'll probably just choose stuff that people have posted in the decks forum, feel free to post your AI tuned decks, wackiness and fun count.

Gando and I have been hard at work and there is a few decks already for you to look at. We realize that you will not be able to add each and every one of them. Thats OK.

With the current Forge build we may have to adjust somewhat but I think everything will work out OK.

[mtgrares]

Sounds great. All of the quest data is separate from the rest of the program, well, as separate as it can be. Quest mode will have to use a custom version of the deck builder. And when you win cards, it will use the booster pack files common.txt, etc.. to generate the mini-booster packs of 7 cards. (7 cards instead of 15 in order to be more challenging but the exact numbers can be changed later.)

[DennisBergkamp]

These decks Gando and Chris have been brewing are indeed pretty mean, in a game last night the AI had a Doran, the Siege Tower, Rolling Stones and 3 Shield Spheres in play, by turn 4 I think it must've been

[Chris H.]

Quest mode will have to use a custom version of the deck builder. And when you win cards, it will use the booster pack files common.txt, etc.. to generate the mini-booster packs of 7 cards. (7 cards instead of 15 in order to be more challenging but the exact numbers can be changed later.)

Currently the deck builder uses a set of rarity files that Gando adjusted for playability. Having the computer pick randomly from these cards might result in the computer having a stronger random deck.

In order to handicap the Human player you could choose to generate the Human's booster packs using the full set of rarities which are at this time located in a subdirectory named originalRarities.

[GandoTheBard]

Sounds great. All of the quest data is separate from the rest of the program, well, as separate as it can be. Quest mode will have to use a custom version of the deck builder. And when you win cards, it will use the booster pack files common.txt, etc.. to generate the mini-booster packs of 7 cards. (7 cards instead of 15 in order to be more challenging but the exact numbers can be changed later.)

This is not a good idea. Ideally you want to have an independant source for your card rarities because the player may want quest mode for one thing and drafting for another and the two will not necessarily meet in the middle. also I recommend having 4 rarities for the quest mode so that uber cards that you do not want showing up in every prize pack (damnation/wrath of god for example) are extremely rare. so qc.txt, qu.txt, qr.txt and qe.txt for example should be different from the common.txt, uncommon.txt, and rares.txt files we get with the regular game. Also Im of the mind that these files might not even be .txt files as your quest mode might be better as something only you can tweak once we figure out the power levels.

[Rob Cashwalker]

"Extremely Rare" = Mythic...

Wrath is probably going to be printed as mythic rare in M10 (can't stand that name, but that's another story...)

[GandoTheBard]

Rob my hope was to avoid confusing people regarding Official rarities by using a none WOTC word that actually makes a little more sense than "Mythic" which is a great marketting term but not so great for desciptive purposes.

Anyway I am doubting Wrath will even be in M10. Aaron clearly states that we can derive 5 of the 11 mythics from the article. That means since he only mentioned one actual Mythic which is a planeswalker the other 4 we can assume are the other 4 planeswalkers from that cycle. The 6 remaining cards will be brand new. Also in the article he states that M10 will have a lot of new cards never printed before. That leaves alot of the older cards out in the cold.

Personally Im a little disgusted with WOTC on a huge number of levels so I will not comment on the points in that article other than what I mentioned above.

[mtgrares]

I think it will be OK to have mean AI decks because the user can choose to not play against them. My goal is to make it challenging and lengthy. I want to quest mode to mean something when you actually get to the end or when you get the card you want.

(Mythic rares and other Wizard's activities are annoying but I guess you have to take the good with the bad. Personally it is still better than the artifact dominance of Mirrodin, Arcbound Ravager was just insane.)

[GandoTheBard]

I think that Mythic rares are not in and of themselves a horrible idea. I think it gives wotc a chance to reprioritize certain types of cards which in theory is a good thing. I don't like the way it was done but thats my personal opinion not based on business accumen. And that is the primary thing here...WOTC is a business so they make decisions based on the bottomline for businesses (profit margins). What they really needed to do imho is change their antiquated sorting mechanism but *shrugs* I guess they don't mind that deals have a method to steal the mythics out of boxes.

Anyway I really think 4 or even 5 rarities would best for the quest mode particularly if you see it as a lengthly gauntlet affair. Cards like Wrath of God in combination with recursion and search can just rule. As can Counterspells, cards that remove from game (plow), extremely cheap direct damage and removal (bolt etc), Ancestral is just a game winner if you get it early on or when needed in a clinche. Time Twister while double edged really should almost always help you more than your opponent. Black Vise ouch. See? lots of cards that need to be uber rare imho. Maybe only available as prizes or not at all in some cases. Well its up to you but I think it is something to think about.

[mtgrares]

My goal is to just finish a basic version of the quest mode, that way other people or myself can tweak it later. Quest has been in the back of my mind for awhile.

[Chris H.]

Please post your decks that the AI plays well. For quest mode I'll probably have 10 "ai tuned decks". I want each deck to be as different as possible like ... land destruction ...

I was curious and found that the computer was not playing the cards like Sink Hole and Stone Rain against me. I looked at the CardFactory.java file and found that these types of spells will only target a Basic land.

And i found a comment that states:

// target basic land that Human only has 1 or 2 in play

I was hoping to make a major modification to my old land killer deck. I don't think it is going to work.

[mtgrares]

Well I'm making good progress on quest mode. I'm 80% done, all the screens work and the quest data can be saved and loaded. I mention tomorrow (Wednesday) on my blog, because I know MTG Forge is buggy the old save file is never deleted, it is just renamed something like "questData-11". While this will generate a number of old files, it will mean that your game is never lost.

I'm working on the last thing which is building the menu for the deck editor. (I'm not changing anything about the deck editor except the menu.) This screenshot is correct and the data is gotten directly from the new class QuestData. Easy requires that you win 50 matches, medium 100, hard 150, and very hard 250. During that time you also win new cards at various times, easy gets new cards after 2 wins, medium 4 wins, hard 6 wins, very hard 8 wins. And right now the booster packs that you win have only 7 cards. (Wizards is cutting costs, lol.)

All the numbers are just guesses and can be tweaked later.

[mtgrares]

I was curious and found that the computer was not playing the cards like Sink Hole and Stone Rain against me. I looked at the CardFactory.java file and found that these types of spells will only target a Basic land.

When I wrote the AI routine for Sink Hole and Stone Rain nonbasic lands weren't anywhere in my head. My goal was if you had 1 Mountain and 2 Forests in play, that the AI would target your Mountain.