Encyclopedia of Failure Modes

Some great thoughts here.

With many others in this thread, I share an interest in risk analysis, accident prevention, and human decision-making, especially as pertains to climbing etc. I think it is right to be asking ourselves as members of the climbing community how we can continue to address these concepts effectively as the sport continues to grow and evolve. I think the motivation behind this encyclopedia of failure modes is great.

However - as a non-expert and fellow observer of the same trends - I’m not confident that this is likely to bring any meaningful change, if reduction of accidents is the goal. If the mere creation of another educational resource is the goal, then that might be just as well. To me, while countless “avoidable” accidents continue to occur, it isn’t for want of readily-accessible educational material. Many resources exist across a pretty wide range of platforms and formats, institutional and unaffiliated, paywalled and free, ripe for consumption by those who are looking for training and ed. I don’t really see how an encyclopedia would offer something new here. I also see challenges to the organization of that information (it seems like much of the utility of the encyclopedia would be in its ability to expose and disclose unanticipated ‘modes of failure’ - that’s going to entail a lot of entries!) but some have already alluded to this.

I do think that the basic problem you’re describing deserves to be addressed, that as a community our training, education, etc can be better, and that there’s a ton of room for improvement here in an exciting way. I just don’t think that the exposure to data is the problem. As we know, people sometimes make odd decisions even with all the relevant data. I wonder if thought around AAC’s Accidents has evolved now that that resource is so well established. Maybe some from deeper in those circles can chime in. Looking forward to the continued conversation on this.

So if the above goal is correct, the question seems not to be how to present more information (or the same information differently), but rather how to better capture the non-information seekers. Does that seem right?

(For further on decision making, behavioral psych, etc, I learned a lot from reading Kahneman’s Thinking, Fast and Slow and also Nudge and Noise by Sunstein, Thaler, and others)

I don't think we're trying to solve the non-information-seeker problem.  Rather, to make existing information more easily available to "casual seekers".  I.e. people who want to learn and do the right thing but won't subscribe to / read pages of accident reports, or watch 20-minute videos with ads for 10 seconds of info, or look for insights on page 6 of forum pissing contests and semantics debates.

You may well be right about "meaningful change" in terms of numbers.  I try to think of this as helping the people we can help.  If somebody figures out a way to help everybody else, I don't think the existence of this encyclopedia would stop them.

D Bauman wrote:

So if the above goal is correct, the question seems not to be how to present more information (or the same information differently), but rather how to better capture the non-information seekers. Does that seem right?

Thank you for your comments.  I agree with part of the above. I do think it might be valuable to present known info differently,  I think that capturing noninfo seekers is a next step, "marketing" stage. My main goal right now is to get the actual "product" developed.  

One of the management & leadership courses I took a long time ago had 5 "basic principles".  My favorite was "Take initiative to make things better".  That principle is what is motivating me to spend time on this.  Because I'm so tired of seeing young, clueless people die. Honestly, if I end up spending 100 hours on this but it saves even one person's life, I will think it's time well spent.

In a way though it's kind of ironic that I'm doing this because, as we all know, most accidents happen because experienced climbers make deliberate decisions that have an unhappy result.  I'm one of those climbers. I consider myself a very safe climber whi decides to do "unsafe" things. I choose to climb runout routes.  I sometimes choose to climb things with sketchy pro. And I will do other things that technically are frowned upon.  I and people like me are not the target audience I'm thinking about. 

Julian, I sent you a message through the MP system on 2/15.  If you didn't see it, please check your spam folder.

If you are still unsure about where this information might be hosted, wikimatrix.org is a good resource to compare the different wiki options available. It also has a step-by-step wizard that will provide a filtered list of the wikis that fit your needs.

As a newer and mostly self-taught climber, I can say that this kind of resource would be very useful for me and others like me. If you need any assistance on the technical side of things I would be happy to donate some time to this project.

Will Charbonneau wrote:

If you are still unsure about where this information might be hosted, wikimatrix.org is a good resource to compare the different wiki options available. It also has a step-by-step wizard that will provide a filtered list of the wikis that fit your needs.

As a newer and mostly self-taught climber, I can say that this kind of resource would be very useful for me and others like me. If you need any assistance on the technical side of things I would be happy to donate some time to this project.

Thank you! I will contact you. 

Though still a skeptic, I do see the appeal of the project, from the creator’s standpoint. I wish I knew more about the technical aspect of things and applaud those who do.

Maybe my rants would be more at home in the politics page (so far there is no pedagogy page).

I do still wonder about the basic methodology here: 

It seems that what a beginner needs most is to understand a few core concepts, and then how to abstract from them. Rather than learning 100 ways not to do something - learning the 3 principles for performing X skill correctly, that capture the essential elements required to ensure safety/success.

Imagine teaching any new skill this way. There are infinitely more incorrect ways to perform a skill than there are fundamentals of performance.

A common deficit among all learners, and relevant practical concern here, is getting “lost in the sauce” - being unable to distinguish critical errors (violations of core principles) from insignificant ones (I’m not sure exactly where I’m supposed to be tied into, but at least I have a big long stopper knot).

It’s in this way that big long lists can actually, in some instances, make the education problem worse or harder.

I see two basic failure modes. The complete noob that doesn't know what they are doing yet and the seasoned pro who gets complacent. A novice who knows the basics is going to be extra careful. An intermediate will do the safety checks and try not to die. A total bad ass will do all kinds of sketchy stuff simply because they are so used to the vertical environment. One of my friends who has been on the cover of the AAJ rapped off the end of his rope last summer, the ice climber who was not clipped in a month or so ago and slipped off the ledge while making a v thread, Lynn Hill forgetting to finish her tie in knot.. The list is long and storied. 

I agree with DB long lists do more harm than good for beginners - beginners can't judge the relative priorities of long-tail items.  I have no intention of showing my list to beginners.

I'm looking at this as "things I learned after more than 10 years of climbing that I wish I'd learned sooner".  I know there is a thread for that, but MP threads aren't always readable (at best you have to click next every 20 items, at worst you get pages of semantics on a tangent).

Examples of knowledge that can reduce superficially simple common problems:

 - distraction by conversation while tying in is a common cause of unfinished knots - a useful footnote to "finish your knot"

 - rap end knots have a way of coming untied when a pile of rope is forcibly yanked off a ledge - a useful footnote  to "always tie end knots"

 - forgetting to clip in as a common unconscious mistake on big ledges - a useful footnote to "always clip in"

I learned all 3 many years after learning the best practice they interfere with - sooner would have been better.  This is the niche I'm trying to fill.

Currently we're thinking Phyl will publish a well-structured article and I will publish a flat spreadsheet (google sheets).  I will open it to editing by interested MP users with google accounts and a history of useful forum activity.  One thing I'm not sure of is whether submitted changes should get published immediately or reviewed by me first in case they contain Viagra links (the latter would on average take a couple of days).

My "article" is almost ready to publish. Probably a few more weeks. I want to add a few more links.  (but travel and climbing for myself is a higher priority!)  Will update here where that ends up...

https://climbingdangers.net/

In the end, I decided to publish my thoughts as an article on a Wordpress site. I bought a private site so the content would be ad-free.

You can leave comments here, thru PM to me here, or to the email address on the site.

I apologize in advance that there is not currently a table of contents so that you can jump to sections. If I have the energy and time in the future I will install the plug-in to generate  one.

Many thanks to the people who read this in some form and gave helpful feedback: Bob Gaines, Serge S and Will Charbonneau.  Obviously they are not responsible for any mistakes I made, or for my dry and humorless writing style.

If this helps keep even one person from getting hurt, by sending them to learn about something they didn't previously know, it will be worth the time spent.

Great Read.  Much better than what I would have been reading.  Thank you for pulling this together.  I think a bit more on the proper way to clean would have been very useful.  I have caught others who were nearly in very bad situations.  I think this ultimately is on the inexperience more than anything else.  I believe there was a bad fatality in 2022 in North Carolina over this.  

Nice work!

Great piece of work. Thank you. I hope newcomers to our activity read this from start to finish.

Alois Smrz wrote:

Great piece of work. Thank you. I hope newcomers to our activity read this from start to finish.

Thank you, Alois and others,

I will make a specific post about this for the Beginners Forum after a few days.  Just waiting to see if I get any comments that lead to modifications.

Phyl

Minor, but for SEO and readability I would use headings instead of bold text and also add anchor links to them so folks can link directly to specific causes.

Tradiban wrote:

How about Google sheets?

There are hundreds of “failure modes” but only a few that matter. Won’t they get lost in minutia of a spreadsheet?

It's a special type of paradox when one misspells "minutiae".

phylp phylp wrote:

https://climbingdangers.net/

Phylp, as I’m new to outdoor climbing, I’ve been looking for something exactly like this so thank you x infinity. A compilation of sorts would’ve saved me so much time already and I still know so little.

I’ll read it over this evening or next and give you my thoughts.

On the other hand, I’m also an internet nerd so I may be able to help organize this in a better way, spruce up the seo so it can be found organically, and have some ideas to get it to new and experienced climbers.

Here are some thoughts of how I’d go about this:

-reach out to AAC and see if they’d consider making something widely available as a free resource. We, mtn project and reddit users can compile failure modes but they seem to be a primary source of data and could add statistics, which could be used to filter by priority.

If not, here is another way to go about it:

-compile failure modes from communities like reddit for mtn project. List them in a Google sheet, as you’ve done. 

-scrape google results (extract all html text from the top ten pages) for each failure node term + rock climbing (optional but may provide better results)

-export that sheet as a .csv and potentially the scraped results for each, and then feed it into ChatGPT/ Open AI’s API. This will generate additional columns for a) categorizing each as TR, Lead, TRS, Trad etc b) experience level like beginner-advanced and c) write a short description how to avoid it, based on its current knowledge or the knowledge we feed it. 

For displaying the data, there are many data table plugins but you could also import each row into Wordpress, either using the default post/categories and/or tags so it can be sorted and have potential to rank for organic — I’m sure you don’t care about traffic of monetization, I’m with you there, but the posts would it’d help get it in front of people googling. I’d suggest all of the below, a master table, exportable exportable table, and categories/posts/tags.

Afterwards I’d reach out to mtn project or the mods if there are any, the main climbing subreddits and AAC to get their thoughts, proof read it (since gpt can make mistakes) and of course share it.

I have python scripts I can share which deal with all the open ai api stuff. I have one that can scrape Google results for a single term as well and it could be expanded to search for all terms within the sheet.

The only thing I’m unsure of is the data table Wordpress plugins since I haven’t looked in awhile and mass importing into Wordpress, though there is probably a plugin for that too.

API credits are relatively cheap, I’d estimate no mo than $100-200, Google/client scraping apis are cheap but dk the pricing, plugins are free, and I’d be willing to chip in/ run the scrapes.

User suggestions/adjustments can be submitted through github or another pipeline as well. Although I don’t have the spare time to constantly moderate/ approve but I’m sure someone would be on board if they shared the same vision. 

having worked with number databases and gpt/bard AI- DO NOT RELY ON IT. It’ll add random pieces that look similar to your data and completely and utterly screw everything up. Spent hours chasing numbers back and forth and only realized after tons of cross checking it was inserting gibberish into my data. Be very very careful and always double check. Not sure if the same will happen with text but just a warning. 

phylp phylp wrote:

https://climbingdangers.net/

Phylp, as I’m new to outdoor climbing, I’ve been looking for something exactly like this so thank you x infinity. I’ll read it over this evening or next and give you my thoughts.

On the other hand, I’m also an internet nerd so I may be able to help organize this in a better way, spruce up the seo so it can be found organically, and have some ideas to get it to new and experienced climbers.

Here are some thoughts of how I’d go about this:

-reach out to AAC and see if they’d consider making something widely available as a free resource. We, mtn project and reddit users can compile failure modes but they seem to be a primary source of data and could add statistics, which could be used to filter by priority.

If not, here is another way to go about it:

-compile failure modes from communities like reddit for mtn project. List them in a Google sheet, as you’ve done. 

-scrape google results (extract all html text from the top ten pages) for each failure node term + rock climbing (optional but may provide better results)

-export that sheet as a .csv and potentially the scraped results for each, and then feed it into ChatGPT/ Open AI’s API. This will generate additional columns for a) categorizing each as TR, Lead, TRS, Trad etc b) experience level like beginner-advanced and c) write a short description how to avoid it, based on its current knowledge or the knowledge we feed it. 

For displaying the data, there are many data table plugins but you could also import each row into Wordpress, either using the default post/categories and/or tags so it can be sorted and have potential to rank for organic — I’m sure you don’t care about traffic of monetization, I’m with you there, but the posts would it’d help get it in front of people googling. I’d suggest all of the below, a master table, exportable exportable table, and categories/posts/tags.

Afterwards I’d reach out to mtn project or the mods if there are any, the main climbing subreddits and AAC to get their thoughts, proof read it (since gpt can make mistakes) and of course share it.

I have python scripts I can share which deal with all the open ai api stuff. I have one that can scrape Google results for a single term as well and it could be expanded to search for all terms within the sheet.

The only thing I’m unsure of is the data table Wordpress plugins since I haven’t looked in awhile and mass importing into Wordpress, though there is probably a plugin for that too.

API credits are relatively cheap, I’d estimate no mo than $100-200, Google/client scraping apis are cheap but dk the pricing, plugins are free, and I’d be willing to chip in/ run the scrapes.

User suggestions/adjustments can be submitted through github or another pipeline as well. Although I don’t have the spare time to constantly moderate/ approve but I’m sure someone would be on board if they shared the same vision.