Collectible Card Game Headquarters

I see from time to time, on random forums people who have plans to start "a new MTG game, with AI, a history mode, and multiplayer". Most of the time, these remain at the state of "projects" because these people have no clue how long it takes to have an engine that handles even only 5% of the rules.
My piece of advice for people who want to try such a huge project is to start small, and clearly define what they want to do: a simulator with perfect respect of the rules ? A video game with an AI ? A multiplayer game ? A campaign mode ? How many cards would you like to handle ? etc... start small, don't aim for campaign mode + 10 000 cards + AI + multiplayer straight away, or you'll end up with nothing.

So I'm asking the people who have something that "works", how long have you been working on your projects? How long before it was a tiny little bit playable? How many cards does your game handle? What are the key features of your project (what makes your project different from the others)? What huge features would you like to see before you are satisfied with what you have (a.k.a. missing features)? What features are missing from your implementation that you think are not too important?

I'll start with Wagic
- How long: I've been working on Wagic for 18 months now (it's taking most of my free time)
- How long did I actually take to get the first playable version: 9 months
- How many cards: the game handles 1000~3000 cards (don't know how many exactly, haven't counted the user-created sets)
- Key features: AI, addictiveness because of the credits/booster system, compatibility with several platforms including Windows and the PSP, possibility to create your own cards with a simple text editor, highly customizable (sounds, graphics...)
- Missing key features: better AI - campaign mode (shalara like) - wifi 2 players game - online mode to exchange/buy/sell cards (with fake money obviously) - internationalization (language files)
- Missing features I don't really care about: 100% respect of the official rules/ Handling all the cards ever created.

Looking forward to other creators' answers

I'll go next...

Mox (My Own eXperiment? )

- How long: Been actively working on it for about 8 months but been thinking about it and following "the scene" for a couple of years
- How long did I actually take to get the first playable version: about 6 months to get Shock working with UI + network framework
- How many cards: 6 Basic lands + Shock
- Key features: Heavily unit tested (code coverage > 90%), WPF/.Net 3.5 based, Network (WCF)
- Missing key features: More cards obviously, AI (though I'm starting this next), UI polish, the infamous campaign mode (which is definitely the last thing I'm going to do)
- Missing features I don't really care about: The possibility to define cards in some script format (xml or other). I think it's an error to spend time trying to devise a perfect scripting system for defining MTG cards/rules. IMHO, traditional code can be as clear as so called "scripts" (even more... Firemox anyone?) and it's so much more flexible. And easy to integrate with automated tests (unit tests).

I know all the newcomers think they can create "the program to rule them all" in 2 months. That's not my case; I'll do it in 4 Jokes apart, I think I have good programming notions and a good approach to durable development. It's my real life job, after all (I'm a lead tools developer for a major game company). So I hope I'll be able to get to a more "finished" product.

Frank
http://moxdev.wordpress.com

BlackMamba wrote:I think it's an error to spend time trying to devise a perfect scripting system for defining MTG cards/rules. IMHO, traditional code can be as clear as so called "scripts" (even more... Firemox anyone?) and it's so much more flexible. And easy to integrate with automated tests (unit tests).

In Incantus, the cards are coded in straight Python code (the same as the rest of the program). However, there's no reason you couldn't have an interpreted language like Python be your scripting language. Firemox is a terrible example... XML is awful for game logic (and the people who code cards for Firemox don't know the rules right anyway, at least when I took a look at the program). Using an interpreted language gives you all the advantages of having complex logic structures, like conditional branches and for-loops, without having to hack them into something not meant for them.

BlackMamba wrote:So I hope I'll be able to get to a more "finished" product.

So do we all, and good luck to you.

I won't give the data for Incantus, since I didn't start the project (although for those interested, I've been working on it (on and off, in my spare time) for about nine months).

Hm...3 years. About 1 year to get things running, I kept adding new cards so I had 50 or so cards when everything worked. And two more years of updates and adding new cards.

- How long: 3 years in my free time

- How long did I actually take to get the first playable version: Seems like 12 months but maybe only 10.

- How many cards: 1440

- Key features: AI, sealed, drafting, quest mode, edit booster packs,

- Missing key features: better AI (of course), better user interface (I'm afraid MTG Forge falls behind Incantus and Wagic.)

- Missing features I don't really care about: None really, cards bugs are usually funny and the AI is just a generic opponent that does "good enough" and is not expected to be a genius. I think of playing the AI like playing against your little brother. My goal for the AI is to have it sometimes surprise you.

I started Incantus in November of 2007, and I've been working on it off an on since then (although I haven't done much in the last few months).

- How long did I actually take to get the first playable version:

The first thing i worked on was getting the stack interaction working (basically priority). After that I implemented a simple 2D interface and started with the basic phases. Finally I worked on combat. I wish I had used a version control system in the beginning, so I could recreate my early progress, but after about a month and a half I had a basic playable version (with Honor Guard, Plains, and a made up Toolbox card).

- How many cards: I think the current count is about 3500 tested cards

- Key features: The eventual goal is 100% rules accuracy. Cards are implemented in python, so you get the full benefits of a full programming language when implementing cards. Also, the program defines a mini DSL (domain specific language) for magic.

- Missing key features: Any kind of AI - currently the only way to play is solitaire mode (gold-fishing mode)

- Missing features I don't really care about: Not sure yet. Probably AI is furthest on my todo list.

Incantus wrote:- How many cards: I think the current count is about 3500 tested cards

Current status:
Total cards: 4130
Untested: 162
Dubious: 24
Error: 115
Tested: 3829

They just keep going and going and going...

By "tested", do you mean automated tests? If so, I'm impressed

BlackMamba wrote:By "tested", do you mean automated tests? If so, I'm impressed

tested means that the card has been run in incantus and that its abilities have been tested at least once, for example a trigger msut have triggered at least once, an activated ability will have been activated.
Of course there will be some mistakes in there, but most cards actually work they way they should

The Scientist wrote:

BlackMamba wrote:By "tested", do you mean automated tests? If so, I'm impressed

tested means that the card has been run in incantus and that its abilities have been tested at least once, for example a trigger msut have triggered at least once, an activated ability will have been activated.
Of course there will be some mistakes in there, but most cards actually work they way they should

In other words, someone put time and effort into making sure each of those cards at the very least does what it says on the tin, which ensures no simple coding errors slip by, and usually means they follow the comprehensive rules, as well.

Also, the number went up again.

Hmm, in one way or another I've been working on my project since about 2002. It originally started out as a way to play the Middle Earth Collectible Card Game, morphed into trying to handle multiple games with all kinds of bells and whistles, and now is targeted at just Dreamblade http://ww2.wizards.com/dreamblade/home.aspx.

I had originally written it in python but the AI was just waaaaay too slow.

I have AI implemented (quite good on a fast computer), most abilities and mechanics, some testing, replay support (although it's currently not working AGAIN). I need to do a proper UI for it and would like to have the ability to specify miniatures via text files.

- How long: The past year or two has been the bulk of actual coding and I'm getting closer to being finished. I hope to be putting the source up and doing some sort of release soon.
- How long did I actually take to get the first playable version: maybe 3 or 4 months, barebones though
- How many cards: 391 pieces out of the 396 released
- Key features: AI, flexible engine with (very little) special casing, almost all minis supported
- Missing key features: Any sort of networking or multiplayer, non text based UI
- Missing features I don't really care about: Multiplayer/networking

~telengard

telengard wrote:I hope to be putting the source up and doing some sort of release soon.

Looking forward to that!

Dreamblade sounds really fun, I enjoyed playing the online demo. I haven't played any sort of miniature game before. Keep up the good work telengard.

Hmm, in one way or another I've been working on my project since about 2002. It originally started out as a way to play the Middle Earth Collectible Card Game, morphed into trying to handle multiple games with all kinds of bells and whistles, and now is targeted at just Dreamblade

Narrowing the project really helps you get done. It is really hard to cut features but if you don't, you will never get anything working. It is tough being a computer programmer,

mtgrares wrote:Narrowing the project really helps you get done. It is really hard to cut features but if you don't, you will never get anything working. It is tough being a computer programmer,

Projects are never released, they escape.

A statement which seems oddly accurate about Incantus, actually...

mtgrares wrote:Dreamblade sounds really fun, I enjoyed playing the online demo. I haven't played any sort of miniature game before. Keep up the good work telengard.

Hmm, in one way or another I've been working on my project since about 2002. It originally started out as a way to play the Middle Earth Collectible Card Game, morphed into trying to handle multiple games with all kinds of bells and whistles, and now is targeted at just Dreamblade

Narrowing the project really helps you get done. It is really hard to cut features but if you don't, you will never get anything working. It is tough being a computer programmer,

It's so true. Focusing on one game made it a lot easier to a) make some meaningful progress and b) realize my initial goals were pretty insane.

~telengard

I've been working on my (unnamed ) project for about 2 months in total, though I started 3 months ago..
So far I've only been implementing the infrastructure, such as the event system, the layer system and the stack. I have however implemented a few cards in order to get a better sense of what kind of events and systems they need to function properly. (First non-land was Dingus Egg)