Testing regime

Instead of kill count-based testing, here's a more win-focused method.

Our current method of testing works inconsistently at best. Some types test extremely easily, other types literally cannot pass tests. This is due to a compound-style scoring where one major mistake can make you lose the entire match/test. Instead, here's an idea: Focus on wins instead of kill counts.

Change the testing regime to one of the following, or stagger it based on the position:
a 2/3 wins with a major disadvantage, minor disadvantage, and neutral.
a 3/5 with 1 major disadvantage, 2 minor disadvantage, and 2 neutrals.
a 5/7 with 1 major disadvantages, 2 minor disadvantages, and 4 neutrals.
a 6/9 with 2 major disadvantages, 2 minor disadvantages, and 5 neutrals.
Or even a 7/10 with 2 major disadvantages, 3 minor disadvantages, and 5 neutrals.

And now here's the kicker. Instead of relying entirely on pure type matchups, classify the major disadvantages, minor disadvantages, and neutrals by how the type actually fares against the other type. For example, a test for Normal with the 3/5 method would be:
Steel (MaD), Ground (MiD), Rock (MiD), Fairy (N), Bug(N).

For the most part, the type matchup list can be used, but certain matchups (Dragon v Normal) aren't correct, because Normal roflstomps Dragon teams, no matter how good the Dragon user.

Pastebin of type matchups: http://pastebin.com/RfM0WjsP