I have never seen a cold shut that was closed in an anchor just open.
Neither have I but this test was how they are intended to be used. I did this because I thought they might be stronger than a smash link which it seems is true. It would be better to use a cold shut in an anchor than a smash link. Quick links are of course even better.
what's a factor of safety?
A good thing. The factor of safety in this case is just the ratio between the breaking strength and the safe load. It is interesting that the FS seems to be different for the two types. I would have expected them to be the same since they have the same purpose. Factors of safety vary for different materials, load types, etc. Most buildings use a FS of about 2. lifting equipment in an industrial setting would probably have a FS of 2 or 3, lifting above people 4. If you are lifting people, I believe OSHA requires a FS of 7. FS is higher if there are lives at risk, if the failure type is brittle and if there is no redundancy. This paragraph takes the traditional Allowable Stress Design (ASD) approach. Up until the last 10 years or so, metal and wood structures were designed this way.
There is another way to provide for some safety which was for many years called ultimate load method or strength design (SD). In this approach the minimum breaking load is calculated then a factor is applied to increase the design load. Different loads have different factors. If a load is very predictable, the load factor is small. SD design has been used for concrete structures since the mid 1960's.
The version currently in favor is called Load and Resistance Factor Design or LRFD. The loads are factored up by various amounts based on their predictability and the strength or resistance is factored down based on predictability and brittleness. LRFD is becoming common for steel structures and is used for wood structures.
The three design concepts above will give the same size structure (beam, column, whatever) within about 10%. Steel designers like the 10% savings because it saves money. LRFD is more work and book keeping.
Climbing gear is rated according to an approach that is not ASD but has some aspects similar to SD and LRFD. For climbing gear the capacity or strength given (e.g. a 22 kN carabiner) is the minimum breaking strength so there really is no factor of safety incorporated in the strength. The load is unclear. For lead climbing the load is something like 2000 to 3000 lbs on the rope. Not because a falling climber can't generate larger forces but because it is assumed he will be dead if he does. Somplace around 10 G impact and your insides get scrambled. So, if the climber is to survive a 3000 lb force on the rope, the carabiner and anchor have to resist pretty much twice that. Small carabiners are rated at 22 Kn which is 5000 lbs
As far as I can tell, UIAA doesn't have seperate capacities for rappel gear. I think they feel it should be the same as lead gear. It is possible to shock load the rope on rappel but I personally think it is is acceptable to have rappel gear with a lower capacity than lead gear.
If a smash link is used to provide rappel capability at an anchor, The OSHA FS of 7 might be a reasonable place to start. A big climber with bear might weigh 300 lbs. 300x7=2100 lbs. I would hesitate to suggest using anything for rappel with less capacity than this. On the other hand, I have no doubt that I have used anchors with less capacity than that for rappel.
