For the first year or so after I relocated to Squamish, my efforts at sport climbing were blatant grade-chasing. I sieged my way up a 5.13a and then a 5.13b. Embarrassingly I then realised that the underlying improvement in my performance didn't really reflect that level; I was still finding routes at a much lower grade hard. So, from mid-2014 onwards I decided to go into reverse and "build a pyramid" - a process recommended by many climbing coaches. Essentially this meant climbing a lot more 5.12d sport routes. I made a list of eight popular 5.12d's across a range of styles and focused on completing them all through 2015 and 2016. I ticked the last one off, the horrifically thin Vorpal Sword at Murrin, in September last year. I also climbed several more at the grade that weren't in my initial list, including a couple of new routes confirmed by others as 5.12d.
My main observation from this exercise was that the 5.12d grade is not a very useful guide to difficulty. I came close to flashing one on my list but needed seven days to complete another, the brutal Mr Negative at Chek. Discussing this with other climbers, I heard a common opinion that in Squamish there is resistance to admit sport climbing routes into the 5.13 zone, which is seen as much more prestigious than 5.12; consequently 5.12d is "broad", as it has become over-populated with routes that might elsewhere be given 5.13a. Being an empirical kind of person, I wondered if this theory was supported by any data?
Helpfully the excellent sendage.com route database supports filtering by area and grade to create route lists. Furthermore, the grades are consensus from the site's users, so usually correct obvious anomalies in guidebook grades. I copied and pasted from sendage into Excel and then did some easy calculations to create a frequency distribution of incidences of routes across a range of grades. For further context, I ran the same analysis for Skaha, the other area in BC with a large number of sport routes, also heavily visited by sendage.com users.
These are the results in chart form, Skaha first:
And Squamish:
A few observations from the data:
My main observation from this exercise was that the 5.12d grade is not a very useful guide to difficulty. I came close to flashing one on my list but needed seven days to complete another, the brutal Mr Negative at Chek. Discussing this with other climbers, I heard a common opinion that in Squamish there is resistance to admit sport climbing routes into the 5.13 zone, which is seen as much more prestigious than 5.12; consequently 5.12d is "broad", as it has become over-populated with routes that might elsewhere be given 5.13a. Being an empirical kind of person, I wondered if this theory was supported by any data?
Helpfully the excellent sendage.com route database supports filtering by area and grade to create route lists. Furthermore, the grades are consensus from the site's users, so usually correct obvious anomalies in guidebook grades. I copied and pasted from sendage into Excel and then did some easy calculations to create a frequency distribution of incidences of routes across a range of grades. For further context, I ran the same analysis for Skaha, the other area in BC with a large number of sport routes, also heavily visited by sendage.com users.
These are the results in chart form, Skaha first:
 |
| Frequency per grade for Skaha sport routes with two or more reported ascents, with exponential trendline (data extracted from sendage.com) |
 |
| Frequency per grade for Squamish sport routes with two or more reported ascents, with exponential trendline (data extracted from sendage.com) |
A few observations from the data:
- The Skaha chart shows the frequency distribution I assume to be "normal" for a fairly large and diverse sport climbing area: that with each increment in grade there are fewer routes. The only data point that doesn't fit the curve neatly is the 5.13a grade, but the variance isn't pronounced.
- As anticipated, and in contrast to the Skaha data, the 5.12d grade does appear to be over-represented in Squamish.
- However, the 5.13a grade doesn't appear under-represented in the Squamish data. It fits the curve.
- The 5.12c grade looks significantly under-represented in the Squamish data. (So does the 5.13b grade to some extent.)
So where does that leave the theory that people are reluctant to give routes 5.13 in Squamish? Seemingly not very well supported. The simplest conclusion is that there is a "12c issue"; 5.12d is broad because a proportion at the soft end really belong in 5.12c. I have not spoken to anyone who thinks that make sense, but the data "is what it is" (a Canadian platitude that irritates me - but that's a whole other topic!).
Alternatively, if you stare at the chart for long enough, it is possible to imagine the data fitting the curve through a more general "smoothing": push about five 12b's into 12c, similarly about five 12d's into 13a and a few 13a's to 13b. In other words: there isn't a specific "sticky" grade - Squamish sport climbing is systematically sandbagged. This seems to resonate with many people's experience. Significantly it is much easier to get suggestions for upgrades (*) from people than the converse: suggestions for downgrades. So I am inclined to run with that conclusion.
For completeness, here are those two rival theories re-plotted:
Before anyone says it, I should add that I am fully aware that this analysis rests on a few questionable assumptions. The main one being that the plot of grade frequency for an area should have some specific shape at all. For an individual cliff, it would obviously be ludicrous to expect that each increment in grade is less populated. In fact it is easy to think of counter-examples. On the other hand, I am reasonably sure that it is a correct assumption for the total data set of sport climbing routes worldwide, at least down to the grade below which people tend not to be interested in developing routes. It has been decades since I last studied statistics but, as far as I recall, sampling theory - in particular the minimum size of a sample that can be expected to represent the whole - is well established stuff. So it should be possible for someone to do the math and state whether the data set of Squamish sport routes is a large enough sample. But not me.
I should also state again that this topic is only about sport climbing grades in Squamish. The style differences between sport and trad here are so great that I don't see much point in in trying to compare grades between the two. People, especially americans, will endlessly tell you how some popular Squamish trad crack would be three grades easier in the "Valley" or the "Creek". Whatever.
* For what it is worth: my upgrade list would include Ty Man from 12b to 12c, Mr Negative from 12d to 13a and Ibiza from 13a to 13b.
For completeness, here are those two rival theories re-plotted:
 |
Frequency per grade for Squamish sport routes with two or more reported ascents, with exponential trendline
ADJUSTED with five routes moved from 5.12d to 5.12c
|
 |
Frequency per grade for Squamish sport routes with two or more reported ascents, with exponential trendline
ADJUSTED with 5 routes moved from 5.12b to 5.12c, 5 routes from 5.12d to 5.13a and 3 routes from 5.13a to 5.13b
|
Before anyone says it, I should add that I am fully aware that this analysis rests on a few questionable assumptions. The main one being that the plot of grade frequency for an area should have some specific shape at all. For an individual cliff, it would obviously be ludicrous to expect that each increment in grade is less populated. In fact it is easy to think of counter-examples. On the other hand, I am reasonably sure that it is a correct assumption for the total data set of sport climbing routes worldwide, at least down to the grade below which people tend not to be interested in developing routes. It has been decades since I last studied statistics but, as far as I recall, sampling theory - in particular the minimum size of a sample that can be expected to represent the whole - is well established stuff. So it should be possible for someone to do the math and state whether the data set of Squamish sport routes is a large enough sample. But not me.
I should also state again that this topic is only about sport climbing grades in Squamish. The style differences between sport and trad here are so great that I don't see much point in in trying to compare grades between the two. People, especially americans, will endlessly tell you how some popular Squamish trad crack would be three grades easier in the "Valley" or the "Creek". Whatever.
* For what it is worth: my upgrade list would include Ty Man from 12b to 12c, Mr Negative from 12d to 13a and Ibiza from 13a to 13b.
