Kolin, head of quality assurance over at Black Diamond, just did some more belay loop testing in response to Todd's accident, check this link for more info. One of the reasons I really like working with BD is that many of the people there are truly obsessive climbers, with access to stacks of high-tech lab equipment. As a climber and professional "tester," Kolin was obviously directly interested in belay loop failure and so immediately did a stack of tests on belay loops with various "issues." Right on Kolin, thanks. What Kolin's tests showed is that even extensively damaged NEW belay loops are still very strong. As I read through both his belay loop tests and then his previous tests and reports (lots on that page) I noticed that Kolin kept repeating some varation of this quote from his tests: "Regardless—swap out old crappy gear—the heartache avoided could be your own..." Most of his reports were written before Todd's accident.
I recently wrote on this blog about teaching clinics a couple of weeks ago in Maryland where several people were belaying by clipping through their leg loops and swami belt instead of the belay loop. This accident has made me think about that, but I'm still far more worried about a biner breaking through cross-loading or through the carabiner flipping over and having the gate pressed open than I am about belay loops breaking while belaying. Carabiners can break when cross-loaded, or when loaded with the gate opened, I've had it happen several times now and watched the lab tests, it does happen. I've only ever heard of one belay loop ever breaking, I'll go with the belay loop as it reduces the odds of biner failure dramatically.
For rappelling the decision is a little more murky. The forces involved on the average rap are generally pretty low (although they can be surprisingly high when rapping on double ropes, remember that's two strands to share the force, resulting in less rope stretch and therefore higher loads on the carabiner if "bounced"). Most carabiners should be more than adequate to handle this sort of load even if cross-loaded or if the gate gets pressed open, so those arguments are slightly less compelling. Overall I'm still inclined to use the belay loop to rap. For starters, it's simpler, cleaner and easier to see what's going on with the device and belay loop connection. Human error happens, a belay loop is just simpler to see those errors. I also don't like having the ropes sliding through the carabiner so close to my tie-in points. It's pretty easy to have the harness bunch up in such a way that the rope runs on the nylon tie-in point during the rap. Nylon running on nylon is really dangerous as most of us know, it melts and cuts very quickly. One rap with a rubbing rope on the harness isn't likely to cut a tie-in point, but it sure could make it a bit weaker, and I'm not confident in knowing how many raps like this would equal failure. A belay loop keeps the device well away from the tie-in points and prevents "bunched up" rubbing on them, the tie-in points need to be absolutely strong for harsh falls.
I also don't like reducing the distance between the "hot" part of the carabiner and my tie-in points. With a belay loop the distance or "heat sink" is the full size of the carabiner, but may only be a few cm of the carabiner if the carabiner is through the tie-in points. Likely not a big deal but, over years of use, ?
The final thing I don't like about using the tie-in points instead of the belay loop for raps is related to the closeness of the belay device--it's a lot easier to get clothes, gear etc. stuck into the device. While this is not normally fatal, I do remember getting my shirt stuck into my belay device years ago. At the time I felt fortunate to be carrying a knife; in retrospect, any blade near a rope I'm hanging on is a really stupid idea... I've also seen people rapping with the device on a long sling, this isn't good as it's easier to get hair stuck into the device, I've seen that happen a few times too. A belay loop seems about right.
One thing I am considering using more often is a backup prussic knot of some kind. I've always thought these added more complexity than they were worth--I've seen all sorts of cluster fucks on rap with people using backups. Some were pretty funny, some were potentially life-threatening (dark, -20, the prussic freezes to the rope after weighting it and the climber is left hanging there on an ice climb trying to sort it all out). I've seen far more potentially "bad" situations than I have situations where the person was potentially "saved." I do put knots in the ends of my ropes on "mystery raps" in the dark. In my mind knots in the end of the rope on any "suspect" rap are way better than a prussic, most people's reaction to falling is to grab the prussic knot, which then just slides uselessly down the rope. I've seen a few accidents where people rapped off the ends of their ropes, a prussic knot wouldn't have done anything as the rope goes through the device so fast that the person would have to be thinking not to "squeeze" the knot as the ropes went through the device and toward the knot, I just don't see that happening. Prussic knots may be useful for those raps where the ends of the ropes are hanging in space, but in that situation the prussic knot is only useful if you know the ends of the ropes are hanging in space and stop early, before the ends of the ropes. If you know that then why would you take it to the ends of the rope anyhow? And if you're rapping into suspect terrain/rope combinations then you ought to have knots in the ends of the ropes anyhow... As for falling rocks etc. the equation comes down to how many problems I've seen with backups vs. how many times I've been smacked stupid on rap by a rock. Lots, and never are the answers. Wait, I just talked myself out of a backup prussic knot, and yet Todd's accident has still got me thinking it might be a good idea...
One of the things Todd was good at was getting people to think, that's reportedly why his corporate presentations were so good (I've only ever seen his climbing talks, which were awesome). So thinking about all of our systems is a good thing, I just wish it didn't take a guy like Todd dying to get me thinking about it this stuff for hours.