|
|
Forums:
Climbing Disciplines:
Trad Climbing:
Re: [curt] Possible legal action against CCH Inc.:
Edit Log
|
|
ja1484
Jun 20, 2007, 12:19 PM
Views: 4337
Registered: Aug 11, 2006
Posts: 1935
|
curt wrote: Just FYI: When posting on a public internet forum, nothing will make you look more stupid than accusing someone else of something that you yourself are guilty of.
....And what exactly makes you think I haven't spent a fair amount of time around statistical math?
jt512 wrote: ja1484 wrote: dynosore wrote: We're not talking about a failure to meet 3 sigma limits here....we're talking about pieces failing at a small fraction of their stated rating. If a cam is rated for 10 kn but fails at 9.5kn, such is life. It should NEVER fail at 1 or 2kn, EVER! Crunch the numbers, if their process is "in control", that failure should NEVER happen, period. Yet is has several times. They should be making toasters, not life saving devices, if their knowledge of manufacturing processes is this poor. You haven't done much in the way of statistics I take it? You're trying to think in absolute terms on one hand, and then relative terms on the other hand. If you crunch the numbers, eventually it is a mathematical certainty that a cam will get through quality control and be primed to fail at 1 - 2kN. Dynosore is right. Under reasonable assumptions, no cam rated to 10 kN should ever fail at 1 or 2 kN. Assume a cam with a 3-sigma rating of 10 kN has a normal distribution and a standard deviation of 0.5 kN. The mean strength of the cam is thus 11.5 kN, and 2 kN would be 19 standard deviations below the mean. The stat software on my computer puts the probability of -19 standard deviations to be on the order of 1 in 10^80. To put 10^80 in perspective, it is about 10 times greater than typical scientific estimates of the number of atoms in the Universe. Thus, 1 in 10^80 would be an event so rare that we would expect it never to occur in human history. Jay
We're in agreement, but that doesn't change the fact that the math is there. Just to be clear where our arguments differ: Dynosore wasn't talking about reasonable assumptions, he was talking about all assumptions, so I laid it out. If you change the delineations to *reasonable* assumptions, which is actually what I advised him to do if you go back and read the post, yes, you'll reach the conclusion that we share.
I climb, and I don't worry about the gear, because I know that this type of failure is not a realistic occurrence. I realize what the numbers mean.
But, I'm also not enough of a goofball to look at things in absolutes, which is a sure path to neurotics if you try to reconcile it with climbing. It's never going to be black and white and come out in such a way that the person seeking absolute safety gets it.
Hence the fallacy of a black and white point of view here.
In reply to: As Jay has already pointed out, you are completely wrong. For one thing, most gear testing (as at Black Diamond) is done at 50% of rated strength. So, it's fairly clear that a 10kN rated piece of gear that has been tested to 5kN, will never ever fail at 1 - 2kN. If you are claiming that the 50% strength testing itself has weakened the piece to that extent, the odds are far better that you will die from a meteor impact. Curt
See, I love this type of thing, because we're in complete agreement, but apparently you don't think so.
The odds that a piece will fail in the above fashion are absurdly small - that would be why I pointed out to the guy I quoted that looking at the situation in a boolean manner
(WILL HAPPEN or WON'T HAPPEN)
was pretty stupid. One of those lovely attributes of mathematics is that when looking at the probability of ANY occurrence, you will eventually reach a point on the bell curve where it does happen.
If one considers the situation as a will it or won't it happen to me type of thing, then you have nothing to worry about. But when you try to cut it down to an on/off result like he did, you'll go crazy because you're seeking a guarantee of safety. His mistake was in saying that something should never happen, ever, instead of saying that it was unimaginably unlikely.
I'm actually rather happy that Jay and yourself misinterpreted my comments (or didn't take the time to read them thoroughly, whichever the case was), and ran the math on through. It saved me the trouble. I wasn't going to bother, as this isn't enough of an issue that anyone should be worried enough to put in the effort. Just buy gear from good manufacturers, and no problemo.
To recap:
- Dynosore was referring in absolute terms, which means all possible occurrences ever, which includes that 1 in 10^80 chance.
- I pointed out that this was a surefire recipe for driving one's self crazy and instead suggested he look at whether or not it was likely to affect him rather than occur at all, eventually, ever.
- I am lambasted by several folks for...I'm not sure why. Having the same view as them but bothering to point out why the opposing view is silly?
- And here we are.
(This post was edited by ja1484 on Jun 20, 2007, 12:28 PM)
|
|
|
Edit Log:
|
|
Post edited by ja1484
() on Jun 20, 2007, 12:23 PM
|
|
Post edited by ja1484
() on Jun 20, 2007, 12:28 PM
|
|
|
|
|
|
