I found some minor sites today in Chelmsford. They are already pretty broken down and not too delicate, so I see no harm in identifying the locations.
The first site is a remnant of a few things at the G.W.Wright Reservation off of Parker Rd (lower right blue outline in the picture). Most of the rock pile material I saw was focused around a large boulder in there.
Notice the structure at the left end of the boulder. Here is a closeup. You can see that this is in fine repair and my guess is that this is very recent. In fact throughout the day I was looking at things that included recent structures built over older ones.
In addition to this there was one alignment leading away from this boulder. And nearby one or two broken down platform piles.
Here is a detail of the near platform.
Note the quartz and/or white feldspar in the detail. Quartz "blazes" like this were another theme of the day. Everywhere I went the rock piles had a blaze like this, usually with a single quartz or feldspar cobble.
I was thinking it is likely that quartz (or feldspar) occurs with a certain frequency in the glacial cobbles of this area. So it follows that you expect to see it mixed in with other rocks in a rock pile with about the same frequency. So if you suppose the rocks of a rock pile to be a random sample then there will be a natural variation found in how many quartz cobbles occur per pile. For example at low frequencies you expect a number of piles to have no quartz at all, and a number of piles to have two or more. One can then calculate the likelihood of seeing a pile with no quartz and compare that to what is actually observed. Standard hypothesis testing will, I believe, prove that the "random sample" hypothesis for rock piles is not a good hypothesis. This is one way to make the argument quantitatively - should anyone care to do the work.