Despite being released about 4 years after the TI-82, the TI-86 does not include median-median regression, which can be a useful method of linear regression for data with extreme outliers. Fortunately it's a simple enough algorithm to implement in a small program.
Copy MedMed.86p to your TI-86 via TI Connect (or your preferred method), and run the program. You'll be prompted to enter your X and Y lists (the program does not support frequency lists), which can either be lists keyed directly into the program, or names of list variables. After a few seconds of processing, you'll see the computed coefficients, and the regression equation will be saved to y21.
The program works by first sorting the input data points in ascending order by the X values. Then the data is partitioned into three equally-sized subsets: left, middle, and right, using the X values. In cases where the number of data points is not divisible by three, the program will attempt to make the left and right subsets have equal size, and vary the size of the middle subset by 1. Then the Y values of these three subsets are sorted independently, and the pairs of median values (or the average of two medians when the number of values in the subset is even) from each corresponding X and Y subset are used to create three summary points. The line passing through the left and right subsets' summary points is translated to pass through the average of the three summary points, which produces the final regression line.
Reference: https://bgsu.instructure.com/courses/901773/pages/d7-median-median-line?module_item_id=6364145