The excellent Squamish Select route guidebook, now in its 4th edition, has had a Top 100 list in its last two editions. Stating that a climb is "definitely top 100" is a frequent comment in online and even real-world conversations. These lists are based on one person's subjective opinion, or maybe the pooled opinion of a few people, but I have sometimes wondered if there is a way to do it more objectively, sourcing a much wider range of opinion?
A month or so ago, my friend Tim B pointed out that the sendage.com database actually computes an average star rating for every entry, though you can only see it by browsing the source. For example, Grand Wall's average, based on 94 logged ascents, is 4.8841 (more accurately: is right now, as the average is recalculated every time someone logs their ascent of the route in the database and gives a star rating).
 |
| Grand Wall's sendage.com page as the user sees it |
 |
| And the source code for the same page, note the 4.8841 average star rating |
More usefully, the website's useful filtering tool allows sorting by rating, using the average ratings. So you can select an area and then filter with criteria like this:
In plain english: show me all the routes with at least four stars and twenty logbook entries, then sort them by average star rating. This gives 115 results. It is a fairly trivial exercise to then copy and paste these results into a spreadsheet (here, if you are interested) and play around with the list further, for example, dropping the bottom fifteen entries (the ones with the lowest average star ratings) to give a Top 100. When I did this I found a couple of duplicates that I thought uncontroversial to simplify: Spirit of Squamish and Klahanie Crack, and Smoke Bluff Connection and Wonderland. I also dropped a route from Porteau Cove as that is in the Vancouver not Squamish guidebook. Here is the final list, (re)sorted by grade:
A few comments:
- In case you are wondering, the other popular databases (Mountain Project, 8a.nu) don't have the same star rating sorting tools (as far as I can see).
- The list is what it is. It is the consequence of analysing widely-sourced subjective data in an objective way. I am not saying it is "better" than a list constructed in a more conventional way (but I am not saying it is "worse" either).
- The only arbitrary choice is the minimum sends in the filter. Leaving this at zero gives very weird results, biasing especially to new routes only logged a handful of times which the FA'er and their buddies have all given five stars. I chose 20 sends - you will get slightly different results if you lower to, say, 10.
- On the subject of number of logged sends, I assume it is not necessary to point out that actual sends will be a large multiple of logged sends? Because many people don't use sendage or, like me, may only log routes selectively. However I think it is reasonable to assume that logged sends are roughly proportional to actual sends. In other words: if a route has been logged many times it has been climbed many times, and if it hasn't been logged often it isn't getting sent often.
- Unfortunately some very significant but hard routes get screened out by the minimum sends filter. Spirit of the West and Queen Bee are good examples. They have only been logged 3 and 12 times respectively.
- Similarly there are no trad routes in the list above 5.12a. The main reason is that the harder trad routes simply don't get logged very often on sendage. For example, even the uber-classic 5.12c Flight of the Challenger has only been logged 18 times. Cobra Crack has only been logged once.
- There are a disproportionately large number of 5.12d's. I have written about this before.
- I have highlighted new routes established since the last guidebook came out in gold. These constitute 10% of the list, which is a startling proportion, given that the guidebook came out only three years ago. Arguably it is something everyone should celebrate, as it implies the best may yet be to come.
- Sceptical people may also spot that many of the highlighted new routes are from cliffs that I have discovered and/or routes that I have established. Well, that is what the data says. Feel free to do the analysis yourself and see if you discover something different.
- Some exceptional moderate-ish routes established in the last few years aren't in the list because they haven't been climbed often enough yet but I am sure soon will be. Life in Space, Natural Perfection and Zen Garden spring to mind.
- In the Squamish Select books, it is fairly obvious that the Top 100 list is being used to entice people to visit new or overlooked cliffs. That is a very valid reason to highlight some specific climbs over others, but clearly that can't be done with a strict data-driven list.
