On the College Confidential forums, there are whole sections where applicants ask others to “chance me” (a rather odd use of “chance” as a verb, but anyways). They post their stats and desired target colleges, and want others to tell them how likely they are to get an offer. It is primarily U.S. college focused, so I thought I would develop a system where you can “chance” yourself for Waterloo Engineering, as an extension of what I discussed in the post about cut-offs.
In the brochure for Waterloo Engineering, there is a table showing your chances at an offer, reproduced below:
This is good, but not necessarily very precise since it’s based on broad grade ranges. We can improve the “granularity” of the information by doing some interpolation within each range. For this process, I assumed the following:
- The probabilities are averages for the range and occur at the mid-point. For example, the 60% probability (first line) is for Environmental Engineering with an admission average of 82.5%.
- The probabilities vary linearly over each range.
- For higher grade ranges where no data is given, the probability is 100%.
- The probability below grades of 80 is 0%.
- The probabilities in each range are “piecewise continuous”, meaning (for example) that the probability at the top of the 80-85 range is equal to the probability at the bottom of the 85-90 range.
So, with these assumptions and an Excel spreadsheet, I created the chart shown below, where you can look up your admission average and read-off the applicable probability.
There, much more precise. However, not necessarily accurate. The data in the brochure is accurate for last year, but nobody knows how this year will be, and the linear interpolation used here may not be accurate. Remember the difference between accuracy and precision: 2+2=5.0000 gives a very precise answer (4 decimal place precision), but it is still wrong. Anyways, a fun little exercise in interpolation, and a bit more visual representation if you want to “chance” yourself.