alphametics: mark computationally intensive test as extra-credit

While this test is canonical, it does not technically add additional coverage. This test serves as a test for efficiency (exercism/problem-specifications#1024 (comment)) of a solution, not completeness.

Furthermore, here are the run-times for this exercise from the [latest Travis build]() (at the time of this writing):
| Python Version | Run-time (seconds) |
| --- | --- |
| 2.7 | 3.155 |
| 3.3 | 2.461 |
| 3.4 | 3.567 |
| 3.5 | 7.270 |
| 3.6 | 0.774 |

Notice that the optimized example solution is only "fast" in 3.6.

exercises/alphametics/alphametics_test.py

@@ -66,6 +66,7 @@ def test_puzzle_with_ten_letters(self):
              "S": 6,
              "T": 9})
 
+    @unittest.skip("extra-credit")
     def test_puzzle_with_ten_letters_and_199_addends(self):
         self.assertEqual(
             solve(