Whether you call it gaming the system or skewing the results, it destroys the validity of the test. The purpose of the test is to allow institutions to make an apples to apples comparison of kids no matter where they come from. The results lose there reliability when some kids take it multiple times. Let’s say Student A takes it once, scores a 1500, and never takes it again. Then you have Student B, who takes it six times, and improves their score from an 1100 to a 1350. Then you have Student C who scores a 1350 on their first and only test. Are student B and C the same? Not at all. Student B inflated their score by taking it multiple times and benefitting from guessing right across multiple exams. If Student C takes the exam multiple times and increases their score to a 1500, are they the same as Student A? Again, no. Even if Student A wanted to put in the effort and spend the money to maintain his or her separation from other students, they can’t as the exam maxes out at 1600. So taking the exam multiple times results in a consolidation of scores near the top with an inability for the true top performers to maintain that separation. The supposed purpose of the test (besides making a ton of money for the college boards) is defeated by the practice we are discussing.
I have argued before that subsequent test score should be discounted by an increasing percentage the more times you take it.
Super scores are a complete joke and perversion of the testing process.