Consider this one way to peel the banana. I’ve created a formula (a couple of different formulas, actually) that begin to assign trail ratings for the various Michigan Mountain Biking trails throughout the state. I chose to focus on trail difficulty for now. I started by trying to break down trail difficulty into a few key factors: elevation gain, technical nature, and length. Notice I did not include an ‘enjoyability’ factor because different people find different types of challenges ‘enjoyable’. An argument could also be made for not combining the three factors into one score, but simply giving each trail 3 ratings (one for each characteristic). However, because the factors are related and their combination can create entirely different difficulty levels, I felt combining them somehow was appropriate.
So the first factor I tackled was the elevation. At first I was going to just take overall elevation gain for one lap (the longest possible lap), but because each trail is a different length, I chose to simply calculate the average elevation gain per mile for each trail. Data is available at the Michigan Mountain Biking Association on each trail’s length so that was easy enough. Elevation gain is always hotly contested, though, so what I started doing was going back to my new favorite app, Strava, and grabbed elevation gain data from Strava Segments at each trail. You can double check that the segment matches the trail on the MMBA by matching up the length (which should approximately match). I simply divided the overall elevation gain by the length of the trail (in miles) and calculated the average elevation gain per mile. Then I took the average elevation gain per mile for all trails (this number will change as the spreadsheet is filled in) subtracted the specific elevation gain per mile for the trail in question, then divided it by the average elevation gain for all trails, multiplied by 5 and then finally added 5. Somehow that worked 🙂 In all seriousness, I ran a bunch of trails through that equation and was simply satisfied with the comparative results. Fault: some hills are harder to climb than others so pure elevation gain may not fully be representative of elevation difficulty. Mitigation: Include steepness of climbs in the technical nature rating.
Second I attempted to calculate the technical nature aspect of the rating. This is and will remain somewhat subjective, but I’ll rely on comparison logic and the rule of the masses. For example, Bass River in Grand Haven is what I would consider a rather technical trail. It’s pretty twisty-turny and requires solid steering skills to negotiate with any amount of speed. It has almost zero elevation gain, so it can’t be the highest in technical difficulty, but it still ranks rather high. Egypt Valley (Cannonsburg State Game Area) on the other hand is a much more flowing, cross country trail that has some elevation gain, but one descent and subsequent climb in particular has some nasty roots. How do they relate? Well, I’d gave Bass a 3 (with 1 being the most difficult) and Egypt a 5. But on the overall difficulty, Egypt will end up with a more difficult overall rating because of the lack of elevation gain at Bass River. We’ll see what responses I get to that.
The final factor was really pretty easy. Length in miles. I know some trails have multiple loops, inner loops, short-cuts, etc. that might affect this, but I chose to use the longest available loop without duplicating trail. Someday, someone might dive into this further and break down some of these trails into sub-trail ratings. It just won’t be me right now. To calculate some sort of rating that matched with the way I was rating the other factors, I decided that 5 miles was a good average trail length and worthy to have each trail compared to. So the math is (5/trail length * 10). The multiply by 10 was to move the decimal back over after the division.
Finally, the question became how do I combine the three factors for an overall score. The three ‘Overall Course Rating’ numbers are three different ways in which the individual characteristic scores could be factored. The first is a simple addition of the three factors (E+T+L). The second multiplies the three factors together and divides by 100 ((E*T*L)/100). The third is simply an average of the three ratings ((E+T+L)/3). I obviously haven’t decided on which is best, yet, as all three are still in the sheet. The question is: are each of the factors equal in weighting? Ah, so after all that, I still may not have slain the beast of developing a logical trail rating. I’d love some thoughts on that if you have them.
Here’s the link to the spreadsheet on Google Drive: https://docs.google.com/spreadsheet/ccc?key=0AlIZePieMofJdGl5VndsVE5hOXpfMnk5N0U4eHFvVkE&usp=sharing
It is an open file, meaning anyone can edit and save data. If you see data that is missing that you know, or if you know that trail information is out of date for either trail maintenance or new Strava Segment info, feel free to make the necessary changes.