<p>Energize, the citation for my above assertion may be found in the paper “The Frog Pond Revisited: High School Academic Context, Class Rank, and Elite College Admissions”, by Thomas J. Espenshade, a sociology professor at Princeton. The work was based upon a dataset consisting of a basket of elite university admissions office decisions. While there are a handful of notable caveats in the data, namely that extracurricular activities were not controlled for in the analysis and that the data set the paper was based on is now 7 years old (and acceptance rates generally higher at the time), the general principle still stands. </p>
<p>While the paper only lists acceptance probabilities up to 9 AP tests per student over four years, after reading the paper I emailed professor Espenshade, who confirmed upon reviewing the data set that the relationship remained linear and statistically significant through 16 AP tests - he only cut it off at nine since at that point nearly the entire applicant body had been accounted for.</p>
<p>You can view the paper at the link below:</p>
<p><a href=“http://www.princeton.edu/~tje/files/Frog%20Pond%20Revisited%20Espenshade%20Hale%20Chung%20Oct%202005.pdf[/url]”>http://www.princeton.edu/~tje/files/Frog%20Pond%20Revisited%20Espenshade%20Hale%20Chung%20Oct%202005.pdf</a></p>
<p>so no, the AP arms race is not irrelevant.</p>