Some of the problems you cite with the balls bunching up and acting as a brake would no doubt happen but I think the same happened to an extent with the Stonehengineers' logs, which are comparable to the balls in some respects - but the net effect was that the logs (and I suspect the balls) would be helpful in toto. The secret, we decided, was pulling damn hard and then all the localised inefficiencies were overcome by the main effect. We even had logs getting stuck fast and the rest of the logs and a 15 ton stone simply sailing over them.

But then again, that "pulling hard enough" mantra makes me think balls and logs aren't necessary, you just mount the stone on two log runners with pointed ends and off you trot, flanked by lever-men each side to do a bit of adjustment, even a bit of stone rowing maybe, when the terrain gets rocky or marshy. The truth is, isn't it, whatever system you suggest they used it can be done twice as well with twice the number of people and 4 times as well with 4 times the people. Stonerowing was based on Burl saying there were only a few families available to do the job. Durrington suggests there may have been zillions available and that would cast doubt on an awful lot of the experimental archaeology on the subject up to now IMO. Who needs efficiency or cleverness if you've got a hundred rugby teams?