# Comprehensive Exams at Stanford University Computer Science

One of the rites of passage of a Stanford Ph.D. student in
Computer Science is a suite of exams called the comprehensive exams
("comps").
While old examinations (and often, solutions) are available, it was
not obvious to me how well one has to do on them to pass, so I compiled
a few aggregate statistics for the ones I took:

## The 1998-1999 Comprehensive Exams (October 1998)

- Analysis of Algorithms
- Author: Serge Plotkin
- Passing scores: 35 - 56 out of 60 points
- Automata and Formal Languages
- Author: Rajeev Motwani
- Passing scores: 40 - 58 out of 60 points
- Artificial Intelligence
- Author: Carlo Tomassi
- Passing scores: 35 - 57 out of 60 points
- Compilers
- Author: David Dill
- Passing scores: 42 - 60 out of 60 points
- Computer Architecture
- Author: Bill Dally
- Databases
- Author: Jennifer Widom
- Passing scores: 40 - 53 out of 60 points
- Logic
- Author: Zohar Manna
- Total score: 54 points (3 per correct answer, -1 per incorrect answer)
- Expected value for random guessing: -3.6 points
- Passing scores: 17 - 50 out of 54 points
- Failing scores: -2 - 15 points
- Numerical Analysis
- Author: Andrew Stuart
- Passing scores: 13 - 30 out of 30 points
- Programming Languages
- Author: John Mitchell
- Passing scores: 40 - 57 out of 60 points
- Software Systems
- Author: Mary Baker
- Passing scores: 17 - 30 out of 30 points

Wang Lam - source Tue Mar 9 14:54:20 1999 - generated Thu Aug 5 02:57:14 2004