So just yesterday I stumbled upon the "improved sliding x" thread and learned that the cordalette doesn't work as it's supposed to. I learned about the equalette, but honestly, I didn't like it very much, so I decided to do some experimenting on my own.
I stayed up 'till 3 in the morning trying different setups, and this is my best result so far.

You basically start with a "craig short" rig, that is, a sliding x adjusted to incorporate 3 pieces instead of two, with NO limiter knots.


Next, you attach the middle strand to the masterpoint with a clove hitch and a quickdraw (or dedicated sling + carabiner(s))


This middle strand-masterpoint attachment must be LOOSE, in order for the equalization to take place. It works with any 15cm quickdraw, or with a carabiner (possibly locking) and a doubled 30cm runner.

This is what provides the limited extension, if one of the outer strands fails, the setup is going to extend until the quickdraw or runner tightens. If the middle strand fails, the setup is going to extend until the clove hitch reaches the masterpoint carabiner.

This rig provides a true three-point equalization, in which the load is shared approximately evenly among the three pieces (friction and other factors play a part, of course). One of the best parts of this rig is the incredible angle range it has compared to the equalette with limiter knots, as long as the quickdraw remains loose, the setup is equalized. On my tests, this yielded nearly 180 degrees of equalizing range
Here's a pic of the rig with force to one side (note how the side of the bed is at an angle)


Now, here's a pic of the anchor with a failure on one of the outside strands, the middle strand takes all of the force, with the other placement acting as a backup, much like if you blow the figure-eighted placement on a three point equalette.


Now, here's a pic of the rig with a failure on the middle strand, note how the rig extends until the clove reaches the masterpoint biner. The force is equalized between the two outer strands


A detail of the masterpoint when the middle strand fails:


Here's what happens when an outside strand and the middle strand both fail, there's BIG extension, because the failed outside strand is entirely absorbed by the strand that held, still, the knot in the middle strand gets caught on the masterpoint, so the extension is not as big as it would be without it.


Now for the discussion:
Pro's
- All three pieces share the load equally ALL the time, minimizing the chance that one fails
- It's simple and easy to set up (IMO, very subjective)
- You don't have to unclip the sling from the pro, tie limiter knots, and clip again
- The masterpoint is created with ONE carabiner, not two.
- the angle range for full equalization is pretty big, as I said, nearly 180 degrees.
- good redundancy, you'd need to cut one of the outside strands, the middle strand, and the quickdraw for everything to fail
- (your input is appreciated)

Cons:
- an extra quickdraw /carabiner with runner is needed
- If two pieces pop, you're going to have lots of extension.
- If one of the outside strands fail, the center piece will take all the force, with the other one acting merely as a backup.
- (your input is appreciated)


with all that in mind, I think this is a pretty good setup to know. I would think it works especially well with one bomber piece (middle) and two not-so-perfect pieces (sides).

I would very much like to know what everyone (especially the experienced ones out there) thinks.
I would also like to have an opinion from John Long, if he's still around. I will try this setup on the field next time I'm climbing trad.

BTW: The knot attached to the quickdraw in the pics is not a clove hitch, this was a mistake.
I was going to post this on the sliding x thread, but it's pretty old, so I decided to start fresh.

well, there goes my hope for good responses.

did you read the whole thing J. Long posted about the myth of shock loading?
Unless there has been any new data to say that a little bit of extension gives you a "shock" load, I'm willing to go with it.

well, there goes my hope for good responses.

If either one of the outside legs go, like you said, its down to one. This makes the whole system have a backup of one, as opposed to two with a normal 3 legged anchor. I'm not real sure I'd like to have a small shock load go to just one leg, maybe its fine, but seems like the amount of shock load isn't clearly defined.

Basically you've got a sliding W with the center leg backed up. I don't think I'd go with this since there are other rigs close to this with better redundancy. Some might complain about carrying extra gear and extra time. Seems simple enough to me, but also a little to simple for my safety. Keep trying though, maybe you can improve on it some more.

OK, I' gave the whole deal some more thought, and read the thing Craig posted about friction thoroughly.
Now, it may be that it's 3am once again and I'm being a complete idiot, but here's what I came up with.

In this setup, because of friction, the loads will be 20% for the outer placements (which slip in order to "equalize") and 60% for the middle placement (which has a knot preventing it from slipping)

this means that there is less possibility that the outer strands fail and leave you hanging on only one strand.
However, since the middle placement still gets 60% of the load, why not tie an equalette instead? where you load two placements at any one time, one of them with 60% of the load.
In an equalette, if the figure eighted strand fails, it leaves you hanging on only one piece, with another one for backup, and it also happens that this figure eighted strand gets 40%, which means it has the same possibility as any one of the two outer strands of my setup failing.
Using this line of thought, my setup is not more or less safe than an equalette, so you're essentialy trading more gear for more angle range.

crap, my head is all CFed from thinking about anchors too much, I hope people understand what I wrote above, although I doubt it. I'll try to fix it up later today.

Edit: BTW majid and moose, the amount of extension if only one leg fails is OK, it's not even 5 inches. If two fail, however, it suddenly goes up to a lot more. Still, if you believe that it's only like falling a couple more feet on dynamic rope, as Craig Conally says, it's no big deal.

you will DIE

if you blow one side of the equalette, the remaining anchors act just like a cordalette.
Once again, I'd like to remind everyone that the load on the carabiners IS along the major axis, IF you tie the clove hitch correctly (load strand closest to the spine)
Anyhow, seems like nobody likes my anchor, for some very good reasons, If what majid said about jt512 and rgold is true, I can't wait to see what they have to say. Maybe it's just something to keep in the toolbox, maybe it's not even worth it.

if you blow one side of the equalette, the remaining anchors act just like a cordalette.
Once again, I'd like to remind everyone that the load on the carabiners IS along the major axis, IF you tie the clove hitch correctly (load strand closest to the spine)
Anyhow, seems like nobody likes my anchor, for some very good reasons, If what majid said about jt512 and rgold is true, I can't wait to see what they have to say. Maybe it's just something to keep in the toolbox, maybe it's not even worth it.

Why dont you just take what you have in the first picture and tie a big overhand knot with it. If you do it right it is all equalized with no extension and avoids that catastrophe waiting to happen. That way you would be able to see the knot and not be left wondering if your quickdraw-jumble was done properly.

If you blow all the placements of an equalette it doesn't hold you at all

I agree with everything you say, except the point you're trying to bring across with this sentence. By "blowing one side of the equalette" I meant that, in a THREE placement equalette, you only have to blow ONE placement (the one without clove hitches in it) to make the whole system act just like a cordelette and put the entire load on only one strand.
with this system, the force that you put on the placements that leave you hanging on one strand is 20% of the total, as opposed to 40% in an equalette.

Active gear, not so much. All it takes is a little moss in a crack or some wetness you don't notice to spit a cam out with hardly any force at all. Remember, friction is absolutely critical for cams to work. Look for cracks with
rough texture that will fit between the cam's teeth. Make sure the crack is free of anything that could lubricate it, dust, moisture, plantlife, etc. This is optimal.