Estimating your CS414 grade

Warning: Please be aware that your estimate is just an estimate. When we actually make up the final grades the Professor and TA’s sometimes shift things around a bit, and as they do this your grade might be nudged up or down. Moreover, we don’t yet know the results for homework 5 (yet it counts more than any other homework!) and for those who take prelim3, we don’t know your exam scores. Also, there are a handful of people who for special reasons like illness are not accurately included yet but will be when I actually make up the final grades for the class as a whole. All of these factors mean that if your estimate leads you to conclude that you are heading towards a B+, you are probably safe in betting that the grade won’t be a B- or an A, but you really can’t be certain that it won’t shift enough to become an A- or a B. Just the same, it would be unexpected for a grade to shift by more than one third of a grade (eg from B+ you might see a shift to A- or B, but it would be very strange for a B+ to turn into a C unless someone computes something wrong).

Compute your average homework score, using 0 if you didn’t hand in a homework and didn’t get Professor Birman’s written permission ahead of time. Multiply by 7. Now, add your two prelim scores. You should get a number between 0 and 300. Here’s a histogram of the scores when I compute them this way:

Now, CS414 letter grades are typically distributed as follows: a tiny number of people receive D’s, and the remainder of the grades split into three groups: the C’s, B’s and A’s. Within each group the top 1/3 get the pluses and the bottom 1/3 get the minuses – although I do A+ grades by hand, and in general I really look hard at the actual grades and patterns of grades, so some adjusting does occur. In the graph above, the cumulative percentages are shown on the right axis.

So, roughly, you can read letter grades off the above graph by looking at it’s 10% points near the top, and 15% near the bottom: the top 20% are candidates for A grades and a handful of A+ grades (typically not more than 3 to 5 in the whole class), the next 10% for A-, the next for B+, etc. As an example, if your score comes out to 243, you would look at this graph and note that 243 corresponds to roughly the 70% percentile point. You are thus on the borderline between B+ and A- – it could go either way. If your computed score is 195, on the other hand, you would be at 19.2 on the cumulative percentage curve, you are probably in the C- range. People with scores below 150 have real reasons for concern.

Since grades are not computed purely from a formula, there is no way to be more certain of where you stand at this stage of the class.

One last comment: I’m assuming that between 25 and 75 people will take the third exam. I’ll compute the deviation from the mean and will use this to adjust grades for those people – they should move up or down, but in a way that won’t impact the overall distribution of grades in the class. So, prelim3 will impact the grades for people who do take it, but should not impact the grades for those who don’t take it.