Hard to say on both AG near future and on ranges for AU. I have not watched and tried to determine settings for algorithmic behavior in AU, some because for my first couple decades in the game AU was out of range except as a buy and hold reserve whereas AG was in range to adding/liquidating (pre-ETF days mind you) but mostly as AG has such exaggerated volatility (opportunity). The ranges you conjecture are certainly familiar, but it would take careful tracking to fine-tune ranges, bands, and thresholds. With silver the frequency of "spikes" is higher, which are a main factor in this whole thesis. Something similar is as you guess probably discoverable in AU but I guess with signals not as easily discerned and also not as percentage-wise large or frequent (?) plays (although I will admit to mostly playing this "info (?)" for moves between ranges than oscillations within a range).
(Why is there a mid-morning 10 minute spurt of apparent not-for-profit selling chopping $0.50/oz off the price? why when "demand" has been floating the price consistently around that half-buck higher mark all morning, and returning it just as quickly as it spiked down? Once or twice I could chalk to random behavior of the market attempting to be rational and efficient - yea, right - but repeatedly? I just have not noticed this with AU enough to conjecture.)
It is imo worth noting that the Friday close did take AG to the space between the low (23) subbands, albeit barely, since much of Friday was spent in the lowest subband and bumping along the low edge of A (23.3), so we had a week starting with falling from the upper side of A to just maintaining A but showing some resistance to giving up the 23.4-23.7 "home ground" despite living mostly around the bottom of A. Seeing the week - if the cusp stuff does not perturb trading too much - tend to A center, at least as you see on the up side of 23.8 would imo solidify the hold on A and making moves to or penetrating an A edge reinforce the "planning" cost testing idea for a push that way when market psychology plus stops/triggers profile is ripe.
JMO
