[jwi]

This is by far the best instructional video I have seen on Front Levers, and anyone who is thinking of incorporating this hold into their training would be well advised to check it out.


(In polish, with English subs)

Since the guy in the video is knowledgable, and can hold a front lever in perfect form for 50 seconds, he correctly identifies the lats as the main mover in the exercise.

Board-bros, climbing's rival to gym-bros, like to prescribe front levers for “core training”. Which is technically correct I guess, since the lats undoubtedly are a part of the trunk rather than the branches of the human body.

For some reason people seem to think that this hold, which is fairly achievable for anyone not having ridiculously long arms or legs posting a decent number of pullups, is a good measurement of trunk stability.

A recent example, in this very strange bench-mark test:



they use the following progression as test 3 out or 4:
Exercise 3/4:
- 1 Point = 10 sec L-sit (bend knees)
- 2 Points = 20 sec L-sit (bend knees)
- 3 Points = 30 sec L-sit (bend knees)
- 4 Points = 10 sec L-sit
- 5 Points = 15 sec L-sit
- 6 Points = 20 sec L-sit
- 7 Points = 5 sec Front Lever
- 8 Points = 10 sec Front Lever
- 9 Points = 20 sec Front Lever
- 10 Points = 30 sec Front Lever

L-sit clearly test the ability to flex the hip, whereas Front Lever test the lats (again!, they are also tested in test 2/4) and to a lesser extent the abs, I guess.

(Also in the test: 10 points for holding a 20mm edge with 220% of bodyweight? Isn't that achievable for a lot of people who have nowhere near world class strength? I would not be surprised to find out that there are climbers holding over 250%. 10 points for pullups with  220% of bodyweight, which is not that far off the world record of 240%... Test 4: Hang on a bar for time? What on earth does that test that has anything to do with climbing? World record is 20 min.)

[AMorris]

God, I had the same thoughts when I saw this weird benchmarking system. My favourite thing about it is that in the Magnus vid they say they based the top score on Alex Megos' stats. I love the idea that they managed to extrapolate an entire system of values largely from the most extreme outlier they could find, and which Ondra would score relatively poorly on. Makes lattice look watertight!

The core thing got me. Front levers are one of those exercises which are significantly easier for the short. I suspect it would not take a great amount of training for a decent shorter climber to smash out 10s, but as a taller climber (181cm) I have to put a lot of time into it to get anywhere near 10s let alone 20  thats my co-rant for you!

[User deactivated.]

Absolutely spot on. After reading the first few sentences of your post I was genuinely about to cite the Magnus training video before seeing you'd posted it.

The demand on the 'core' for a front lever is comparable to that of an L-sit. Perhaps the test should start adding ankle weight on the L-sit if they want something somewhat objective (lever lengths aside).

At 6'2 and 83kg, I could do 1 arm pullups from a full deadhang and circa 10 muscle ups before I could hold a full front lever with arms locked out for 1 second! The back strength demands are massive for the tall.

[duncan]

I was about to post this. N=1: I'm stuck at home at the moment so did their tests; their model underestimated my grade by 3 levels (and I hardly redpoint so you can probably add another one or two to that). 

It's easy to compare the face validity unfavourably with the Lattice assessment. However, since neither the Climbing Bible or Lattice have published the predictive value of their models we have to assume they are equally rubbish (Lattice have the raw data to do this and as they have chosen not to I think we can guess the answer).

[jwi]

In the case of the test battery they do on Midtbø there is no reason to do any statistical reliability testing since it the test fails on first principles, more or less.

BTW, I am pretty sure that one of the biggest confounding variable that is rarely included in these test is number of hours a week of training.

[teestub]

L Sits seem like such a non climbing specific position too, I guess the only time you need to generate tension in that position is passing a stepped roof where you feet are in the back wall? I guess the ability to raise your legs is useful, but not the ability to hold the position for x amount of time.

[Fiend]

I gotta say that anything entitled "+ jwi rant" promises a worthy read.

Mitbo has gone from being  respectable 9b (?) climber to being one of the top 5 instantly-ignorable clickbait chodes on Youtube. Not surprised the benchmarking is off-target and quite different from actual climbing results.

[cheque]

clickbait

  Mitbo’s business model for his YouTube videos is surely based on people sharing them and saying “look at this rubbish” as much as it is on people taking them seriously. I saw the thumbnail of another of his clips and it was titled something like “1-5-9, the Hardest Move in Climbing” 

[Davo]

It's easy to compare the face validity unfavourably with the Lattice assessment. However, since neither the Climbing Bible or Lattice have published the predictive value of their models we have to assume they are equally rubbish (Lattice have the raw data to do this and as they have chosen not to I think we can guess the answer).

I found this interesting. I thought but clearly am mistaken that Lattice claimed predictive powers for their assessment? A couple of friends of mine have had Lattice assessments and their reports /findings are completely bizarre and bear little resemblance to their actual climbing strengths.

Dave

[AJM]

It's easy to compare the face validity unfavourably with the Lattice assessment. However, since neither the Climbing Bible or Lattice have published the predictive value of their models we have to assume they are equally rubbish (Lattice have the raw data to do this and as they have chosen not to I think we can guess the answer).

I found this interesting. I thought but clearly am mistaken that Lattice claimed predictive powers for their assessment? A couple of friends of mine have had Lattice assessments and their reports /findings are completely bizarre and bear little resemblance to their actual climbing strengths.

Dave

I think what Duncan means is how much explanatory power the model has i.e. how much of what you observe is explained versus being the "random noise" of individual skillsets. High explanatory power indicates that the variables they're inputting are really important predictors of climbing ability, low explanatory power indicates that at best there's a loose relationship.

Anecdotally, I've always found the route:boulder comparisons a bit odd (I definitely had something in the past which suggested an equivalency of finger strength between 8a and V9 which to me at least feel a world apart, and recently I saw a screen grab which was suggesting that a V4 boulderer was significantly overachieving at 7a RP. That might be what the data says, but I'm more interested in what the relationship "should" be than what it is in the sample data. Having said that, I've usually benchmarked about right for my route grade, it's just the huge disparity in boulder grades that puzzles me.

[Oldmanmatt]

I can muscle up, 3 sets of 3 with 30 second rest.
I can hold a ten second front lever (I do sets of three. From dead hang, toes to ceiling, lower into lever, hold, return to hang. As slowly as I can).
I only started campusing again last week, but I can 1-5-7 on the large rungs again by my second session.
Finger strength is still shit and I can hardly last a full minute of laps on the Lattice board, in either direction.
I can’t even manage 6C on the Moonboard, but have done a few 7A’s on the wall. I haven’t been outside since August, but couldn’t touch anything harder than 6C (despite a few attempts and some promising individual moves) outside.

Ten years ago, when I lived in Spain and training meant climbing (indoors and out) I was warming up on 7A and projecting 7C most sessions blah blah blah...

( I was lighter, younger and hadn’t had five years of unenthusiastic pottering and desultory forays, with repeated injuries etc etc).

My point being, I could probably get quite a bit better at lots of gymnastic feats (I can hold an unassisted Iron Cross on the rings for a heartbeat now, soon it’ll be two heartbeats and an “ow!” (My ambition is two heart beats, an “ow!” and a “fuuuck” before requiring medical assistance*)).

None of it would make a jot of difference to my climbing.

Even if I put in the hours on the board, as well as lose around 5kg and climbed two grades harder indoors; it would still only make a modest difference to my grades on rock. I know plenty of people who hardly train, couldn’t hold a lever at all and have never even attempted a muscle up, who dance around on problems I struggle to pull on to atm, like they’ve got super glue fingers and an anti-grav belt.

I am fit, strong and workout/train 5 days a week for around 2 hrs a day, but my climbing is still shit (muscle memory-schmemory! Oi Veh!).

So, that vid seems shit to me.

* I’m anticipating an odd snapping noise in there at some point, but since it will probably coincide with the “ow!” or, worse, the “fuuck”, I can’t use it in the timing assessment.

[moose]

I think what Duncan means is how much explanatory power the model has i.e. how much of what you observe is explained versus being the "random noise" of individual skillsets. High explanatory power indicates that the variables they're inputting are really important predictors of climbing ability, low explanatory power indicates that at best there's a loose relationship.

I once directly challenged Lattice to guess my grades based only on their test methods, with no additional information from me.  Tom also had to give an assessment of my strength's and weaknesses that was compared with assessments from people who climb with me - if he was wrong the assessment was free. 

Results were mixed - substantially underestimated my sport ability and slightly over-estimated my bouldering (so I got a free assessment).  But it was boradly correct about my mass of weaknesses (brute strength, power, core strength etc.).  I put it down to me being a bit of a freak with one great aspect (length and crimp strength / weight ratio) and an ability to find sequences that play to that strength (i.e. turn most moves into a straight arm reach and hang). So, a dataset full of more well-rounded climbers was always going to have predictive problems.

https://ukbouldering.com/board/index.php?topic=29636.0

[User deactivated.]

I can hold a ten second front lever

...

Finger strength is still shit

...

I haven’t been outside since August, but couldn’t touch anything harder than 6C

Surely this means just seek out compression boulders and pretend that finger holds don't exist? AKA: my UKC logbook wish list.

[AJM]

I once directly challenged Lattice to guess my grades based only on their test methods, with no additional information from me.  Tom also had to give an assessment of my strength's and weaknesses that was compared with assessments from people who climb with me - if he was wrong the assessment was free. 

Results were mixed - substantially underestimated my sport ability and slightly over-estimated my bouldering (so I got a free assessment).  But it was boradly correct about my mass of weaknesses (brute strength, power, core strength etc.).  I put it down to me being a bit of a freak with one great aspect (length and crimp strength / weight ratio) and an ability to find sequences that play to that strength (i.e. turn most moves into a straight arm reach and hang). So, a dataset full of more well-rounded climbers was always going to have predictive problems.

https://ukbouldering.com/board/index.php?topic=29636.0

Yeah, I remember seeing that and I thought it was interesting. I suppose some of it maybe depends how much weight you put on getting your weaknesses broadly right (your peers could have told you that anyway, it sounds like?) versus it's numbers based predictions of your grades.

To your point on outliers, I sort of agree - I wouldn't regard myself particularly as an outlier on routes, and it's generally tended to align reasonably closely with the observed reality - but I don't know how to square that with the boulder:route equivalences, unless potentially it's by concluding that their "centre point" is someone I would view as massively overstrong for their sport grade.

Or maybe I am just an outlier in bouldering. In which case it'd be great to work out why so I could fix it 

[duncan]

I once directly challenged Lattice to guess my grades based only on their test methods, with no additional information from me.  Tom also had to give an assessment of my strength's and weaknesses that was compared with assessments from people who climb with me - if he was wrong the assessment was free. 

Results were mixed - substantially underestimated my sport ability and slightly over-estimated my bouldering (so I got a free assessment).  But it was boradly correct about my mass of weaknesses (brute strength, power, core strength etc.).  I put it down to me being a bit of a freak with one great aspect (length and crimp strength / weight ratio) and an ability to find sequences that play to that strength (i.e. turn most moves into a straight arm reach and hang). So, a dataset full of more well-rounded climbers was always going to have predictive problems.

I remember this too. I’ve never met Moose to my knowledge but from his posts here I knew he was long and thin, a very experienced and tenacious redpointer, and climbed like a sloth. It wasn’t difficult to guess his route grade would be high relative to his physical attributes (experience) or his finger strength and endurances would be adequate but he’d be low in power and possibly core strength (build, climbing style, favoured routes over bouldering).

More generally, what AJM said. I’ll expand when I’ve more time - I’m looking after a household of sick people and attempting to home-school at the moment.

[tomtom]

More generally, what AJM said. I’ll expand when I’ve more time - I’m looking after a household of sick people and attempting to home-school at the moment.

Were the tests +ve?

[duncan]

More generally, what AJM said. I’ll expand when I’ve more time - I’m looking after a household of sick people and attempting to home-school at the moment.

Were the tests +ve?

Two out of three were. 

[Oldmanmatt]

More generally, what AJM said. I’ll expand when I’ve more time - I’m looking after a household of sick people and attempting to home-school at the moment.

Were the tests +ve?

Two out of three were.

I wish you all the best.

[tomtom]

More generally, what AJM said. I’ll expand when I’ve more time - I’m looking after a household of sick people and attempting to home-school at the moment.

Were the tests +ve?

Two out of three were.

Eek - sorry to hear that. Hope recovery/isolation goes OK...

[abarro81]

That might be what the data says, but I'm more interested in what the relationship "should" be than what it is in the sample data.

Yeah, I think this is quite interesting.
It may just be my impression, but I feel like fewer people are good at onsight vs redpoint than they were 10-12 years back. E.g. I remember thinking that it was quite normal for someone to have a max onsight/flash of 8a and max redpoint 8b (or 7c/8a or 7b/7c etc.), and that 3 grades would be average (so 8a o/s = 8b+ RP). Nowadays I feel like that gap might be bigger because of people's choices around what to focus on. This may be a skewed perception only, and if you'd tracked it via a survey you'd find it's exactly the same now as it was then, but if it is a reality then correlation-type models won't tell you you're a shit onsighter because all your peers are too. Or maybe people used to be shit redpointers 

Ditto with bouldering - if more people are good at boulders but shit at routes then the average changes. Now I always think ~3 grades makes rough sense again (so if you can do 8A then you should be thinking about 8b+s and 8a o/s)... but how do we know if 3 grades is what you "should" have? If it changes, is it because the population is ironing out previous weaknesses in bouldering or just getting less rounded and more shit at route climbing?

Maybe the solution is a dataset that only includes "rounded" climbers - you have to spend a certain split outside vs inside,  a certain proportion of routes vs boulders etc... But you might not get a very big dataset!

[AJM]

That might be what the data says, but I'm more interested in what the relationship "should" be than what it is in the sample data.

Yeah, I think this is quite interesting.
It may just be my impression, but I feel like fewer people are good at onsight vs redpoint than they were 10-12 years back. E.g. I remember thinking that it was quite normal for someone to have a max onsight/flash of 8a and max redpoint 8b (or 7c/8a or 7b/7c etc.), and that 3 grades would be average (so 8a o/s = 8b+ RP). Nowadays I feel like that gap might be bigger because of people's choices around what to focus on. This may be a skewed perception only, and if you'd tracked it via a survey you'd find it's exactly the same now as it was then, but if it is a reality then correlation-type models won't tell you you're a shit onsighter because all your peers are too. Or maybe people used to be shit redpointers 

Ditto with bouldering - if more people are good at boulders but shit at routes then the average changes. Now I always think ~3 grades makes rough sense again (so if you can do 8A then you should be thinking about 8b+s and 8a o/s)... but how do we know if 3 grades is what you "should" have? If it changes, is it because the population is ironing out previous weaknesses in bouldering or just getting less rounded and more shit at route climbing?

Maybe the solution is a dataset that only includes "rounded" climbers - you have to spend a certain split outside vs inside,  a certain proportion of routes vs boulders etc... But you might not get a very big dataset!

Sort of as you say, I guess one way to see the mechanic is that if your dataset tells you someone is unbalanced, you give them a plan to fix it - but if you also fold in their results into the dataset they look a bit less unbalanced than before you did that. There's obviously going to have been changes in the dataset over time - the geographic barrier to entry has fallen massively from when you had to physically be in Toms garden if nothing else.

Your bouldering rule of three still makes me look like a massive overachiever/underperformer (message as appropriate) on my 7A+/8a gap.

I'm still clearly a pro redpointer/shit onsighter on my 4 grade split, but I knew that anyway!

[tomtom]

You have a method for measuring peoples strength - a range of tests - but each one is quite specific. Fair enough.

But perhaps unsurprisingly translating that into route and bouldering grades doesn't work so well - where grades themselves are (a) qualitative and (b) affected by specificity to shape/strengths (mental and physical) as well as (c) conditions... 

[abarro81]

I'm still clearly a pro redpointer/shit onsighter on my 4 grade split, but I knew that anyway!

I don't think 4 is too bad, and certainly not unusual, it's when you get to 5/6 that it looks very disproportionate to me

[Teaboy]

 I remember thinking that it was quite normal for someone to have a max onsight/flash of 8a and max redpoint 8b (or 7c/8a or 7b/7c etc.), and that 3 grades would be average (so 8a o/s = 8b+ RP). Nowadays I feel like that gap might be bigger because of people's choices around what to focus on.

If you go back even further than 12 years though Jibe Tribout had a table of equivalence that was a full letter grade between on sight and in a session and then another full letter grade between in a session and full on project, so 7b OS = 8a RP. Obviously there are too many variables for any of these models to be entirely accurate but this seemed to work for me and most people I know.

Anyone who is in Lattice database is going to be part of a self selecting group who are, on the whole, probably skewed towards single project and indoor training compared to a lot of the people you will see (still operating at a high level) at the crags. Unfortunately I don’t have any data to back this up!

[Nibile]

And there you go, another thread split coming.