I believe that if finger strength is not directly related to the climbing grade, the model must be poorly constructed or there must be an error in interpretation. Despite this, I think it's great that there is starting to be more data analysis in climbing. Nice video!
Extraordinary claims require extraordinary proof. I've taken a look at their blog post, and the biggest issue I see is that most of the variables they study are HEAVILY correlated to each other. For bouldering their top 3 variables are "max hang, max weighted pull up, days of experience climbing outdoors". You can see how someone who has spent more time climbing will have gone outdoors more days, and tipically will be stronger in both the back and fingers than a less experienced climber. For sport climbing the max pull up falls to top 5, and an aerobic endurance exercise on the campus board springs to the top. You can see how a stronger climber will be able to develop more critical force, thus leading to more endurance. The post is a bit lacking in the study of the variables themselves: no cross-correlation of the variables, no PCA (principal component analysis). It's no use saying "this has an influence of 25%, this has an influence of 20, and this has an influence of 19" when all 3 variables are deeply correlated to each other, and measure basically the same thing, thus hiding the effect of still important variables that aren't as influential, but are not repeated. Also, no mention of variable scaling, which might induce some errors: low variance variables will show less effect if the variables are not scaled. Sampling bias is also a problem they allude to in the video, but overall, the claims are really adventurous for the analysis that's been done. Variables are way too related to each other, and simplifying the dataset into the main factors (finger strength, back strength, endurance, experience, BMI) rather than a different number of variables measuring each of these would lead to more insightful results. I would love to have a look at a paper detailing the analysis, I'm sure the blog post is leaving stuff out, but as I say, big claims.
@@paulgaras2606 the claim that most of the factors that have been commonly linked with climbing harder aren't so is quite extraordinary. It's a claim that goes completely against the whole body of conventional knowledge of decades of climbing training, even science based training. You not finding the claim extraordinary doesn't mean it is not so.
@Allen_lena they say in the blog post that the accounted for collinearity between the variables. Would this not refute the claim that the variables are too correlated with each other? Unless of course they simply didn't care about the results of the test of collinearity and chugged along with the analysis anyway.
@@elliotbrown3897 Nah. They say "we accounted for collinearity". They don't say how, they show highly correlated variables as individual prediction variables, and in turn, they minimize the influence of these collinear variables. Them saying "we took a look at this" changes nothing.
When we speak about genetic/bodycomposition (height, weight distribution, etc.) in sport, we talk about 2 things. Firstly, the ceilling an athlete can go. In this case, is there a body that help you climb v17 in our case. We know than anybody with enough experience and hard work can climb v10, but there is a point where it can become impossible. For exemple, I don't know any climber who weight more than 200 pound climbing V15. Secondly, the speed you can get good at climbing. This one is complex, because there is no methodology that can discriminate between the factors that effect your progression. Factors like your genetic (body composition, muscular capacity, etc.), your background (athlete from another sport), your historic (the age you started climbing), just the place your are mentally (capacity to learn, how passionate you are, which training protocol do you use, etc.) In other terms, its hard to make those claims because there is many variables and we cannot easily discriminate between them, what did what. But we know that climbing is strength/weight sport and we know from other sport how body composition affect the performance at the highest level (Gymnastic, Olympic weight lifting, and many other athletics sports). It would be weird that climbing is an outlier. That said, its true that with a good training protocol and enough time we can all climb hard and enjoying our sport.
I feel like the context of weight isn't stressed strongly enough here. No, going from 22 BMI to 17 BMI is probably not going to be great for your climbing; no you don't need to be anorexic to be a strong climber; but there's no way I would believe that going from 27 BMI to 22 BMI couldn't be a great thing for your climbing, especially if that 27 BMI is fat and not muscle.
So after all that nonsense weight is important. Ffs. Talking for the sake of talking. So many variables....soooooo many....being lighter will help. Fact.
Im sure we could go further and break down by body type (height, torso length, femur length, wing span, ape index,...), grade, and style of climbs (and discipline) and see how it distributes.
This post provides misinformation. I’m assuming, as it’s unclear, the interviewee wants to make these big claims rather than helpfully explaining the nuances of the data. Using myself as a case study, I lost twenty pounds and found a set of exercises, like pull ups, and the V4 climbing grade much easier. Then I added 20 pounds back in a weight harness to remember what it was like, and pull ups for example and certain V4 climbs were harder or not possible.
I don’t think you understand the point. No one is denying that climbing is a strength to weight ratio sport, but as long as you are a healthy weight you can screw that ratio by getting stronger instead of lighter. What they claim the data shows is that weight is not strongly correlated with max grade. So if you have climber A who is 5’11” 160 lbs and climber B who is 5’11” 180 lbs, you cannot predict who will climb harder.
@@aaronhauptmann869 Can't you, though? We really don't know. Maybe it's not as simple as a linear regression. Maybe looking at weight alone isn't enough. But of course, their "study" doesn't delve deep into these issues. For one, it uses the variables of weight and height independently, which is stupid, because it's akin to saying studying if "lower weight is always better", which everyone knows it isn't. A 6' climber at a 130lbs probably won't climb that hard, same as a 5'1 climber at 190lbs. You're saying you can change the ratio by getting stronger, which is absolutely true: they claim that isn't that important, due to some shoddy math, and the important thing is "days outside". Of course within a healthy weight lighter isn't always better, and we have things like RED-S and training capacity to think about. But to minimize the importance of overall body composition and even finger strength (as they do) is just selling snake oil.
It could mean that a lot of people could be spending time better doing something other than hang boarding. Would be interesting to look into that better. Would also be interesting to do the survey on a larger group.
I believe that if finger strength is not directly related to the climbing grade, the model must be poorly constructed or there must be an error in interpretation. Despite this, I think it's great that there is starting to be more data analysis in climbing. Nice video!
Extraordinary claims require extraordinary proof. I've taken a look at their blog post, and the biggest issue I see is that most of the variables they study are HEAVILY correlated to each other. For bouldering their top 3 variables are "max hang, max weighted pull up, days of experience climbing outdoors". You can see how someone who has spent more time climbing will have gone outdoors more days, and tipically will be stronger in both the back and fingers than a less experienced climber. For sport climbing the max pull up falls to top 5, and an aerobic endurance exercise on the campus board springs to the top. You can see how a stronger climber will be able to develop more critical force, thus leading to more endurance.
The post is a bit lacking in the study of the variables themselves: no cross-correlation of the variables, no PCA (principal component analysis). It's no use saying "this has an influence of 25%, this has an influence of 20, and this has an influence of 19" when all 3 variables are deeply correlated to each other, and measure basically the same thing, thus hiding the effect of still important variables that aren't as influential, but are not repeated. Also, no mention of variable scaling, which might induce some errors: low variance variables will show less effect if the variables are not scaled.
Sampling bias is also a problem they allude to in the video, but overall, the claims are really adventurous for the analysis that's been done. Variables are way too related to each other, and simplifying the dataset into the main factors (finger strength, back strength, endurance, experience, BMI) rather than a different number of variables measuring each of these would lead to more insightful results. I would love to have a look at a paper detailing the analysis, I'm sure the blog post is leaving stuff out, but as I say, big claims.
I don’t find the claims that extraordinary.
@@paulgaras2606 the claim that most of the factors that have been commonly linked with climbing harder aren't so is quite extraordinary. It's a claim that goes completely against the whole body of conventional knowledge of decades of climbing training, even science based training. You not finding the claim extraordinary doesn't mean it is not so.
As a data scientist myself, I wholeheartedly agree with your comment.
@Allen_lena they say in the blog post that the accounted for collinearity between the variables. Would this not refute the claim that the variables are too correlated with each other? Unless of course they simply didn't care about the results of the test of collinearity and chugged along with the analysis anyway.
@@elliotbrown3897 Nah. They say "we accounted for collinearity". They don't say how, they show highly correlated variables as individual prediction variables, and in turn, they minimize the influence of these collinear variables. Them saying "we took a look at this" changes nothing.
When we speak about genetic/bodycomposition (height, weight distribution, etc.) in sport, we talk about 2 things. Firstly, the ceilling an athlete can go. In this case, is there a body that help you climb v17 in our case. We know than anybody with enough experience and hard work can climb v10, but there is a point where it can become impossible. For exemple, I don't know any climber who weight more than 200 pound climbing V15.
Secondly, the speed you can get good at climbing. This one is complex, because there is no methodology that can discriminate between the factors that effect your progression. Factors like your genetic (body composition, muscular capacity, etc.), your background (athlete from another sport), your historic (the age you started climbing), just the place your are mentally (capacity to learn, how passionate you are, which training protocol do you use, etc.)
In other terms, its hard to make those claims because there is many variables and we cannot easily discriminate between them, what did what. But we know that climbing is strength/weight sport and we know from other sport how body composition affect the performance at the highest level (Gymnastic, Olympic weight lifting, and many other athletics sports). It would be weird that climbing is an outlier. That said, its true that with a good training protocol and enough time we can all climb hard and enjoying our sport.
I feel like the context of weight isn't stressed strongly enough here. No, going from 22 BMI to 17 BMI is probably not going to be great for your climbing; no you don't need to be anorexic to be a strong climber; but there's no way I would believe that going from 27 BMI to 22 BMI couldn't be a great thing for your climbing, especially if that 27 BMI is fat and not muscle.
Even if that 27 bmi is muscle - you can only put so much weight on your fingertips, being lighter is better.
This is such a good example why we can't always trust data and that data can be munged to fill many obviously incorrect claims.
So after all that nonsense weight is important. Ffs. Talking for the sake of talking. So many variables....soooooo many....being lighter will help. Fact.
Im sure we could go further and break down by body type (height, torso length, femur length, wing span, ape index,...), grade, and style of climbs (and discipline) and see how it distributes.
This post provides misinformation. I’m assuming, as it’s unclear, the interviewee wants to make these big claims rather than helpfully explaining the nuances of the data.
Using myself as a case study, I lost twenty pounds and found a set of exercises, like pull ups, and the V4 climbing grade much easier. Then I added 20 pounds back in a weight harness to remember what it was like, and pull ups for example and certain V4 climbs were harder or not possible.
I don’t think you understand the point. No one is denying that climbing is a strength to weight ratio sport, but as long as you are a healthy weight you can screw that ratio by getting stronger instead of lighter. What they claim the data shows is that weight is not strongly correlated with max grade. So if you have climber A who is 5’11” 160 lbs and climber B who is 5’11” 180 lbs, you cannot predict who will climb harder.
@@aaronhauptmann869 Can't you, though? We really don't know. Maybe it's not as simple as a linear regression. Maybe looking at weight alone isn't enough. But of course, their "study" doesn't delve deep into these issues. For one, it uses the variables of weight and height independently, which is stupid, because it's akin to saying studying if "lower weight is always better", which everyone knows it isn't. A 6' climber at a 130lbs probably won't climb that hard, same as a 5'1 climber at 190lbs. You're saying you can change the ratio by getting stronger, which is absolutely true: they claim that isn't that important, due to some shoddy math, and the important thing is "days outside". Of course within a healthy weight lighter isn't always better, and we have things like RED-S and training capacity to think about. But to minimize the importance of overall body composition and even finger strength (as they do) is just selling snake oil.
It could mean that a lot of people could be spending time better doing something other than hang boarding. Would be interesting to look into that better. Would also be interesting to do the survey on a larger group.
It’s great to see people trying to dig into the assumptions climbers have about bodyweight.