|
atg200
Jan 12, 2006, 3:02 PM
Post #101 of 240
(47312 views)
Shortcut
Registered: Jul 27, 2001
Posts: 4317
|
the stats weenies are pretty ridiculous.
the test is important for one reason - the units failed at the braze. if they were built properly, they wouldn't. i bet the failure results are all over the map - it would be pretty hard to braze something incorrectly in exactly the same way over and over again.
the load under which they failed is pretty much irrelevant. much more interesting will be the analysis about why the braze failed.
|
|
|
|
|
scottquig
Jan 12, 2006, 3:04 PM
Post #102 of 240
(47312 views)
Shortcut
Registered: Feb 10, 2005
Posts: 298
|
In reply to: There is no safety factor built in to the rating of climbing gear.
Why not?
|
|
|
|
|
heiko
Jan 12, 2006, 3:04 PM
Post #103 of 240
(47312 views)
Shortcut
Registered: Mar 3, 2005
Posts: 1505
|
In reply to: In reply to: You should give more than a rat's ass. When part of an object must stand up to a certain weight, a factor of safety is used. For something like a cam it's pretty safe to assume a factor of safety of 2 or higher There is no safety factor built in to the rating of climbing gear.
Unless you care to enlighten us regarding your claim, I will stick with the following:
Climbing gear are engineered products. If you want/have to guarantee a certain strength, the only way to achieve this is to work with safety factors. This is normal engineering practice.
If a cam says "10kn" this reads "will hold AT LEAST 10kn" and not "will hold AT MOST 10kn". Otherwise this rating would be utter nonsense, because "at most" would also include a failure at ZERO kn.
|
|
|
|
|
heiko
Jan 12, 2006, 3:06 PM
Post #104 of 240
(47312 views)
Shortcut
Registered: Mar 3, 2005
Posts: 1505
|
In reply to: In reply to: You should give more than a rat's ass. When part of an object must stand up to a certain weight, a factor of safety is used. For something like a cam it's pretty safe to assume a factor of safety of 2 or higher There is no safety factor built in to the rating of climbing gear.
Unless you care to enlighten us regarding your claim, I will stick with the following:
Climbing gear are engineered products. If you want/have to guarantee a certain strength, the only way to achieve this is to work with safety factors. This is normal engineering practice.
If a cam says "10kn" this reads "will hold AT LEAST 10kn" and not "will hold AT MOST 10kn". Otherwise this rating would be utter nonsense, because "at most" would also include a failure at ZERO kn.
|
|
|
|
|
patto
Jan 12, 2006, 3:15 PM
Post #105 of 240
(47312 views)
Shortcut
Registered: Nov 15, 2005
Posts: 1453
|
In reply to: If you want to convince people I'm wrong, you should at least attempt a counter-argument.
I didn't correct you the first time but the fact remains what you said was wrong. If you understood the other posts or understand what three sigma rating is then you would know that what you wrote was incorrect.
In reply to: For a cam rated at 15 kN (similar to the aliens) with a factor of safety of 2, it should not fail at lower than 30 kn for the majority of the samples (99.9%). By this method of safety, the 1 cam in 1000 that fails below 30 kN will nearly always fail at a slightly lower force, much higher than the rated strength.
This is incorrect. In fact Black Diamond say that for a cam rated at 15KN that 99.87% will be stronger than the rating, certainly not 99.9% stronger than 30KN.
Learn more about the THREE SIGMA RATING SYSTEM at:
http://www.bdel.com/about/3_sigma.php
EDIT:
In reply to: In reply to: There is no safety factor built in to the rating of climbing gear. Why not? |
|
|
|
|
patto
Jan 12, 2006, 3:25 PM
Post #106 of 240
(47312 views)
Shortcut
Registered: Nov 15, 2005
Posts: 1453
|
In reply to: Climbing gear are engineered products. If you want/have to guarantee a certain strength, the only way to achieve this is to work with safety factors. This is normal engineering practice. If a cam says "10kn" this reads "will hold AT LEAST 10kn" and not "will hold AT MOST 10kn". Otherwise this rating would be utter nonsense, because "at most" would also include a failure at ZERO kn.
1. Climbing gear rating system DOES NOT guarantee a certain strength.
2. Safety factors are not the only way to achieve strength goals.
(In fact they are not a very good scientific way of achieving safety goals. A rating system based on statistics is a much more scientic way of making statements about a product's strength.)
A rating on a climbing product doesn't guarantee anything. All it says is that according to our testing 99.87% of the products will fail at a load greater than the stated load.
|
|
|
|
|
dingus
Jan 12, 2006, 3:29 PM
Post #107 of 240
(47312 views)
Shortcut
Registered: Dec 16, 2002
Posts: 17398
|
In reply to: But you are trying to convince me, the reader, of your opinion.
Actually I'm not. You may continue to assume climbing gear is built to withstand twice the maximum force rating, despite the words of a BD QA manager who emphatically stated they do not.
Please... carry on.
DMT
|
|
|
|
|
wideguy
Jan 12, 2006, 4:01 PM
Post #109 of 240
(47312 views)
Shortcut
Registered: Jan 9, 2003
Posts: 15046
|
In reply to: In reply to: If a cam says "10kn" this reads "will hold AT LEAST 10kn" and not "will hold AT MOST 10kn". Otherwise this rating would be utter nonsense, because "at most" would also include a failure at ZERO kn. 1. Climbing gear rating system DOES NOT guarantee a certain strength. A rating on a climbing product doesn't guarantee anything. All it says is that according to our testing 99.87% of the products will fail at a load greater than the stated load.
OK so 99.87% maybe does not meet the definition of "Guarantee" in your book, that's fine. But you're argueing semantics and splitting the finest of hairs.
It is generally accepted understanding in our community that a piece of gear will hold at least what it's rated, safety factor or not. It's also commonly understood that these products would test at the same level in independant tests and most all of them do, when tested. (Does that happen often enough, I don't know)
But of the 9 tested by Mgear and the one posted by the climber who had the failure we all started posting about, you're talking 40% failure.
To paraphrase from earlier... if you picked up you car from the brake shop and the guy looked at you and said, "OK, you're all set. Those babies oughtta stop you, or at least slow you down, a good 60% of the time.", would you pay the man and accept it as is?
Whether it meets your rather anal definition of "guarantee," the rating on climbing gear is meant to imply certain assurances of minimum performance. Ask the manufacturers and they'll tell you that, and poll 1000 climbers and they'll say that was their understanding too.
|
|
|
|
|
tradclmbr
Jan 12, 2006, 4:11 PM
Post #110 of 240
(47312 views)
Shortcut
Registered: Sep 4, 2002
Posts: 238
|
Not to rain on JT512's attempt to impress you all with his knowledge of entry level statistics - but one problem.
(from Mgear's statement) - "We took nine units of various sizes and date stamps dating back to the earliest we had in stock which is 0605 for pull testing"
......assumption of random sampling not met. As such, findings apply to that sampling universe (sampling frame) of 9. Since all in the frame were tested (total enumeration), thats a census and findings are what they are (and applicable only to that population). Applying statistical significance to census data is inappropriate. 3 of a population of 9 failed. Period with a degree of confidence of 100%
|
|
|
|
|
bobruef
Jan 12, 2006, 4:25 PM
Post #111 of 240
(47312 views)
Shortcut
Registered: Jul 22, 2005
Posts: 884
|
I think its more than fair to assume the aliens mgear recieves to be representative of aliens recieved by other businesses. The population in question is not all aliens (though, I'm sure many are very interested now in the safety of all aliens), but rather those produced after june 05.
In taking one radom alien from each size within that range of production dates (or population), they conducted a stratified random sample. Which is still a random sample, and valid enough as a basis for said statistical analysis.
Are you implying that mgear's stock of aliens differs in some way from that of others in a way that could affect the outcome of the testing?
|
|
|
|
|
j_ung
Jan 12, 2006, 4:33 PM
Post #112 of 240
(47312 views)
Shortcut
Registered: Nov 21, 2003
Posts: 18690
|
In reply to: In my reply to an earlier post I stated that Mountain Gear would test a sample of our in-stock Aliens and post results on this forum. Following are our findings. Please realize that Mountain Gear’s tests results are not “official,” they were not independent, nor did they represent a random sampling of all Aliens.
We took nine units of various sizes and date stamps dating back to the earliest we had in stock which is 0605 for pull testing. Of the nine Aliens tested, three brazes failed and the cable pulled out at below the CCH-stated max strength for that size. The failures were from different sizes and date stamps, and included both Hybrid and regular Aliens. Since we reported these findings to Dave at CCH, he has been working diligently to identify the source and scope of the problem. I expect that we will see information from him soon. According to Dave, CCH is also reporting the issue to the CPSC (Consumer Products Safety Commission), and will work in concert with them to remove the potentially affected Aliens from the Market. You will have to rely on this process to get the final results of the investigation of these products. We have also informed REI of our findings and they reacted quickly to work with CCH and Mountain Gear so that there can be a coordinated effort to get the potentially affected Aliens off the market and off climbers racks. If you bought potentially affected Aliens from Mountain Gear: We will contact you with instructions on getting them returned, inspected, repaired and/or replaced. We will also post information by Friday 12 p.m. PST on returning potentially affected CCH product at http://www.mountaingear.com/aliens. Mountain Gear wants the best results for the climbing community. I wish I could tell you more but this is the best information we have at this time. We will continue to encourage CCH to act quickly to thoroughly complete their investigation, and will post more information on our website as we receive it from CCH or the CPSC. Paul Fish President Mountain Gear, Inc |
|
|
|
|
j_ung
Jan 12, 2006, 4:42 PM
Post #113 of 240
(47312 views)
Shortcut
Registered: Nov 21, 2003
Posts: 18690
|
In reply to: In my reply to an earlier post I stated that Mountain Gear would test a sample of our in-stock Aliens and post results on this forum. Following are our findings. Please realize that Mountain Gear’s tests results are not “official,” they were not independent, nor did they represent a random sampling of all Aliens.
We took nine units of various sizes and date stamps dating back to the earliest we had in stock which is 0605 for pull testing. Of the nine Aliens tested, three brazes failed and the cable pulled out at below the CCH-stated max strength for that size. The failures were from different sizes and date stamps, and included both Hybrid and regular Aliens. Since we reported these findings to Dave at CCH, he has been working diligently to identify the source and scope of the problem. I expect that we will see information from him soon. According to Dave, CCH is also reporting the issue to the CPSC (Consumer Products Safety Commission), and will work in concert with them to remove the potentially affected Aliens from the Market. You will have to rely on this process to get the final results of the investigation of these products. We have also informed REI of our findings and they reacted quickly to work with CCH and Mountain Gear so that there can be a coordinated effort to get the potentially affected Aliens off the market and off climbers racks. If you bought potentially affected Aliens from Mountain Gear: We will contact you with instructions on getting them returned, inspected, repaired and/or replaced. We will also post information by Friday 12 p.m. PST on returning potentially affected CCH product at here. Mountain Gear wants the best results for the climbing community. I wish I could tell you more but this is the best information we have at this time. We will continue to encourage CCH to act quickly to thoroughly complete their investigation, and will post more information on our website as we receive it from CCH or the CPSC. Paul Fish President Mountain Gear, Inc |
|
|
|
|
tradklime
Jan 12, 2006, 4:44 PM
Post #114 of 240
(47312 views)
Shortcut
Registered: Aug 2, 2002
Posts: 1235
|
Another thing for the folks all caught up in the actual test results and strenght rating etc:
CCH's strength rating is a "MAXIMUM" strength. What the heck does that mean? It is not defined. It could mean that the best you could hope for is 10 kn (as in the maximum strength), but you could also get 2 kn. Who knows? It could also be an attempt to capture that CCH knows the strength of their cams to be 10 kn, but the rock will likely fail before the cam will break, therefore the "maximum strength". Who knows. I find no comfort in an undefined strength rating.
However, if CCH says the cable will break before any other part will, I'm cool with that. I can look at the cable and feel comfortable with the likely outcome. BUT, if the braze is failing, that's no good. No comfort there.
|
|
|
|
|
jt512
Jan 12, 2006, 4:51 PM
Post #115 of 240
(47312 views)
Shortcut
Registered: Apr 12, 2001
Posts: 21904
|
In reply to: Not to rain on JT512's attempt to impress you all with his knowledge of entry level statistics - but one problem. (from Mgear's statement) - "We took nine units of various sizes and date stamps dating back to the earliest we had in stock which is 0605 for pull testing" ......assumption of random sampling not met. As such, findings apply to that sampling universe (sampling frame) of 9. Since all in the frame were tested (total enumeration), thats a census and findings are what they are (and applicable only to that population). Applying statistical significance to census data is inappropriate. 3 of a population of 9 failed. Period with a degree of confidence of 100%
Not to rain on your attempt to impress us, but you're wrong -- partly anyway. You are right that, strictly speaking, that the sample is not truly random; that is, each unit in the population probably did not have an equal probability of selection. However, your designation of the sample tested as the entire population is arbitrary and capricious. It's a false dichotomy: just because the sample isn't truly random doesn't make it the entire population. If you don't believe that the units tested were a sample, rather than the whole population, then do the following thought experiment: imagine that 3 days ago you just received in the mail a CCH Alien you had purchased from mgear. Are you worried?
A more realistic view of samples like this -- and one that is commonly taken in many fields (eg, medicine) -- is that the sample is a convenience sample with characteristics similar to a random sample of the population. If you have a problem with statistics being applied to convenience samples, then you'd better think twice about relying on the dosage of any drug your doctor prescribes, the reliability of any test he uses to make his diagnosis, the RDI for every nutrient you consume, and on, and on, and on, because every one of these depended on statistics that were applied to not-even-remotely random samples.
Now, while the sample in this case is clearly a sample, what is less clear (and this is usually the case in medicine as well) is what the population is. Is the population all units produced during a particular time interval? Is it all units shipped to mgear during a particular time interval? Is it restricted to unpopular sizes? Etc. We don't really know. But let's assume that the population is all units produced duing the time period represented by the units in the sample. Then, if the probability of a unit's being selected for testing is independent of the probablility of the unit failing below the rated strength, then the statistical results will be unbiased (are we into second year statistics yet, Mr. Tradclmbr?). That is a moderately reasonable assumption in this case; after all, it seems unlikely that mgear would have been shipped especially good or bad pieces, or that consumers purchased especially good or bad pieces, leaving mgear with a biased unsold inventory. But even if you don't accept the independence assumption, the fact is that mgear's sample was a sample from some population with an unacceptably high rate of failure below the rated strength. And I, for one, would not want to rely on a unit from that population.
Jay
|
|
|
|
|
iltripp
Jan 12, 2006, 4:52 PM
Post #116 of 240
(47312 views)
Shortcut
Registered: Oct 6, 2003
Posts: 1607
|
In reply to: Another thing for the folks all caught up in the actual test results and strenght rating etc: CCH's strength rating is a "MAXIMUM" strength. What the heck does that mean? It is not defined.
I would wager that a standard industry term like "maximum strength" is defined quite clearly and specifically.
|
|
|
|
|
outdoorsie
Jan 12, 2006, 4:58 PM
Post #117 of 240
(47312 views)
Shortcut
Registered: Jan 15, 2003
Posts: 302
|
Ok, so all you arguing stats can just pass right over this post... it's actually in response to the Orional Posting by Paul at MGear.
I am totatally amazed that there aren't more posts consisting entirely of "I told ya so! I told ya so!" on this board. Maybe that's in a different thread and I missed it? The "Aliens and CCH I told ya so!" thread?
After all those people on all the previous threads arguing loudly that everybody was making way too big a deal out of something that they felt simply wasn't an issue are now, in effect, shown that it is a serious issue, why aren't there more "I told ya so!"s ? Maybe we still haven't established that it is a serious issue? (please... :roll: )
I'm happy that people did make a big deal out of it. If there hadn't been the good pictures and all of the legitimate worry by people on this board, then MGear wouldn't have bothered to do the pull tests and the aliens would not be actually pulled off the shelves as they are now. We'd just be waiting for one of those three that failed at MGear to fail when somebody's life depended on it, and it would have taken a death to cause the reaction that we have now.
Thankfully, this whole issue is being investigated and on it's way to resolved without anybody getting seriously hurt... yet. Again, I'm glad we made a big deal out of it.
|
|
|
|
|
tgreene
Jan 12, 2006, 5:01 PM
Post #118 of 240
(47312 views)
Shortcut
Registered: Oct 22, 2003
Posts: 7267
|
In reply to: CCH's strength rating is a "MAXIMUM" strength. What the heck does that mean? It is not defined. It could mean that the best you could hope for is 10 kn (as in the maximum strength), but you could also get 2 kn.
Bzzzzzzzzzzzt!
In reply to: Minimum Strength (kN): The minimum holding force for Alien cams is 5 kN.
http://www.aliencamsbycch.com/instructions.html
|
|
|
|
|
|
|
|
tradklime
Jan 12, 2006, 5:03 PM
Post #120 of 240
(47312 views)
Shortcut
Registered: Aug 2, 2002
Posts: 1235
|
In reply to: In reply to: Another thing for the folks all caught up in the actual test results and strength rating etc: CCH's strength rating is a "MAXIMUM" strength. What the heck does that mean? It is not defined. I would wager that a standard industry term like "maximum strength" is defined quite clearly and specifically.
How is it defined then?
CCH makes no claims of a 3 sigma rating system or anything similar.
They do claim to be CE and UIAA certified, whatever that means.
|
|
|
|
|
reg
Jan 12, 2006, 5:03 PM
Post #121 of 240
(47312 views)
Shortcut
Registered: Nov 10, 2004
Posts: 1560
|
girldrifter (heilko??) wrote:
" Climbing gear are engineered products. If you want/have to guarantee a certain strength, the only way to achieve this is to work with safety factors. This is normal engineering practice.
If a cam says "10kn" this reads "will hold AT LEAST 10kn" and not "will hold AT MOST 10kn". Otherwise this rating would be utter nonsense, because "at most" would also include a failure at ZERO kn"
i like that - very well put!
by the way: how does that work - girldrifter - with the dog face (on the post not the person!) and heilko also with the dog face! ?
|
|
|
|
|
crimpstrength
Jan 12, 2006, 5:06 PM
Post #122 of 240
(47312 views)
Shortcut
Registered: Nov 10, 2004
Posts: 285
|
tgreene,
5kn is less that the 00 tcu at 6.something
|
|
|
|
|
tradklime
Jan 12, 2006, 5:06 PM
Post #123 of 240
(47312 views)
Shortcut
Registered: Aug 2, 2002
Posts: 1235
|
In reply to: In reply to: Minimum Strength (kN): The minimum holding force for Alien cams is 5 kN. This is the very last line of technical info at the bottom of the page, and is the only place that such info is even listed on their site. http://www.aliencamsbycch.com/instructions.html
Thanks, I stand corrected.
That seems quite a bit different than a 3-sigma rating.
|
|
|
|
|
tgreene
Jan 12, 2006, 5:09 PM
Post #125 of 240
(47312 views)
Shortcut
Registered: Oct 22, 2003
Posts: 7267
|
In reply to: I am totatally amazed that there aren't more posts consisting entirely of "I told ya so! I told ya so!" on this board. Maybe that's in a different thread and I missed it? The "Aliens and CCH I told ya so!" thread? After all those people on all the previous threads arguing loudly that everybody was making way too big a deal out of something that they felt simply wasn't an issue are now, in effect, shown that it is a serious issue, why aren't there more "I told ya so!"s ? Maybe we still haven't established that it is a serious issue? (please... :roll: )
I'M SURE EVERYONE WAS HOPING AND PRAYING THAT THIS WAS ACTUALLY AN ELOBOARTE HOAX, OR AT THE VERY LEAST AN EXTREMELY ISOLATED INCIDENT! :cry:
|
|
|
|
|
|
|
|
|
