No, Jb weld DOES work excellent and has a shrinkage rate of 0.0 and tensile strength equal to any other steel epoxy, and at a fraction if the cost.
ht tp://jbweld.net/products/jbweld.php
The shrinkage rate is not accurate.
All epoxy resins shrink...that is a known fact. The amount of fillers to resin ratio determines the amount it shrinks. Tensile strength has little to do with this application, it is compressive strength that applies since you are pulling an action (either with or without pillars) into the material. Along with removing any movement between the stock and the action, the secondary purpose of bedding is to provide a foundation that does not compress between the action and the attaching screw/bolt heads. This in turn provides more consistant bolt torque.
There is a term called "specmanship" you ought to familiarize yourself with if you take a companies claims or specs into consideration. They spend umpteen dollars to provide specs on products that the average joe is intended to see and use as comparison to their competitors claims. All posted general specs on any product can be "legally" backed up with specific data that contains the fine print and explanations how the testing was done. Rarely is this procedure of testing agreed upon between manufacturers or mandated to be followed by an overlooking body such as the government.
To give you an example of how shrinkage rates can be reduced to zero or near zero by a testing facility....
Epoxy product is cast into a tube measuring 1 meter long and parralel end points are measured in the casting die with CMM machine to within +/- 3 microns.
Cast product is removed and measured again and shows a 1% decrease in size measuring .9900000 meters (+/- .000003 meters).
Same product is cast in a die measuring .01 meters long, removed and remeasured. At the same shrinkage rate of 1%, you have now exceeded the tolerance of you measuring machines capability and the number gets rounded to 0.0%....bingo, you have succeesfully decieved the public with totally legit data.
I worked specifically in this field for many years producing specs for products that are designed to "sell" the product. Terms such as "accuracy, percentages/ volume, weight, density; tolerances, limits, repeatability and hundreds more are extremely difficult for layment to understand and are "dumbed" up for the public to interpret exactly how we intend them to be interpreted. One of the best ones to date are when a digital caliper is produced with a resolution of .0005 on the scale. 99% of the public thinks it is accurate to half a thousandths of an inch because of this when in fact the caliper measuring faces could be out of parallel more than 10 times this amount.
Something else to consider. Talk to top level benchrest competitors or their gunsmiths. Ask them if they use JB Weld or another product for bedding that is proprietary made for the application. Devcon 10110, Marine Tex, ProBed 2000 and to a lesser extent Accraglass are the go to products used. Personal preference seems to go to the ones that are easiest to mix properly, have the right viscocity, produce the right combination of compressiveness without sacrificing to be too brittle.