Tuesday, February 24, 2009

Magic Statistics

Magic seems to attract a large number of computer and math geeks. I, myself, am a computer guy that has taken way too many math classes. So naturally computer people want to program Magic and math people want to generate statistics.

Occasionally I get an e-mail talking about how they would like to write a program that would play two decks against each other and determine which deck is the best. I like this idea in theory and while I can write a program to solve every Sudoku puzzle, I’m not sure that Magic falls into the same category.

Hypothetically let’s say that you wrote a program to test two decks against each other and one deck wins 80% of the time and this was tested for 1000 games. You and your friend now use those two decks and you use the winning deck. Let’s say you played 10 games and you lost 6 of them. Is the 80% winning percentage wrong? No, because statistics only applies to a large number of games. Your measly 10 games is too small.

I’m not saying that Magic statistics can still be fun. I wrote a program to figure out land counts in your opening hand for decks with 2 types of basic lands. I believe a computer program could really help determine the number of land/mana sources as well as when to mulligan.

p.s.
I got the card art from this article about Magic’s 2010 Core Edition which will have 50% new cards, yeah.

2 comments:

Gando the Wandering Fool said...

Not only that but Magic is not simply a game of mathematics but also of psychology and look ahead. It is not simply enough to respond to what your opponent is doing but you must know what they will do as well if you have hopes of succeeding. Alot of what we as humans do while playing is not entirely conscious. We chunk tasks and do logical calculations so fast they seem intuitive rather than reasoned. Now it is true some players play only the deck but they are either very disciplined or very bad. So you and I can play the same decks the AI plays and have entirely different outcomes with them even if we play the same 1000 games. In my opinion you can't really compare PVP to Computer VS Computer. Though such games could still be useful in determining how a PVP match should be played.

Forge said...

Pyschology always matters when it comes to humans.