During an episode of "SOS-Survival" Creek Stewart had occasion to tie a rope around a tree to serve as an anchor for extracting a snowmobile from a stream. He used a timber hitch, stating that this was the strongest possible knot.

This is news to me. I have always favored either a bowline, including several of its variants, or the figure of eight, which tests out quite well in this application. I could find no data on strength of a timber hitch, but I doubt it would do as well as either the bowline or figure eight.

Strength is defined as the percentage of strength retained by the knotted rope, compared to the breaking strength of the unknotted line. A degradation of close to 50% is typical for most knots, with the bowline and figure eight testing at 75-80%.

Any thoughts? None of my anchors have failed yet....

I have no idea if a Timber hitch is the strongest, but as I recall a line loses strength where the knots bend tightly and the “timber hitch” wraps around the line and uses pressure against the tree or cylindrical object to maintain its hold. This looks like it would hold very well as long as tension is maintained. I wonder how well it holds once that tension is released and then reapplied...

You can't judge a knot by looking at it. But if I had to give that a shot, I'd say the Timber hitch looks reasonably strong, but not very secure. Two different qualities. It would definitely need constant tension to hold together well.

What puzzled me was that I could find no data on testing of timber hitches for strength, although there are plenty of data for most other common knots. Anyone know of any tests?

Timber hitch would not be my choice for that use.

High angle rescue teams these days seem to favor the ”tensionless hitch” to tie a rope to a tree. The tensionless hitch is clained to retain 100% of rope strength. However, it will not stay in place without a load.

It occurs to me that the word “timber” in the name is significant. The friction of rough bark on the inside of the wrap is what keeps it secure. I’d like to see how well that hitch holds on different cylinder materials. I imagine a steel pipe would have a different value for friction and the hitch could slip. Maybe knot, but I’d be surprised if the cylinder material did not matter.

The friction of the material (tree bark in this case) certainly does matter. However, that can be dealt with. For the “tensionless hitch” I mentioned, rope rescue manuals recommend taking more wraps for smoother material such as a pipe.

The timber hitch holds fine on a pole without bark as well. It needs a fairly large cylinder though and it does not react well to changing loads and vibration. It tends to work lose under those conditions.
I was in an emergency respond team and we were not allowed to use the timber hitch for securing people or for really heavy loads. So strength did not really matter. The main use was for transporting smaller timber (a size one or two people could lift) vertically. It was always used in combination with a half hitch on the other end of the timber.

Just like a girth hitch can vary in strength depending on how much bend in the rope is created by where the intersecting rope surfaces are, so does the timber hitch.

So one would end up with a range of strengths for the same rope, hitch and anchor.

The tensionless hitch is a better way and has the bonus of being able to be released under tension. Just unclip the biner and unwind. (This can be useful in a rappel situation where someone gets their clothes or hair stuck in the rappel device and is also on a backup belay. Hold the person on the belay rope, give a few feet of rappel line, then reclip the rappel anchor. Then the person can clear their system, belayer release the belay, and continue with the rappel. Much like using a Munter/Mule system on the rappel anchor for similar screwup solutions. Cheaper than cutting the rappel rope free of the anchor and safer than trying to use a knife to cut clothes or hair away from a tensioned rope.)

The timber hitch holds better when dragging timber (imagine that) as it cinches down as the log turns, rather than unwinding, with less rope needed to make the hitch.