I figure a good way of testing gold recovery is to take a number of lead pieces that resemble the size and shape of gold you’d like to recover (a hammer could help I’m sure), paint them up in a bright colour, then drop them in front of the nozzle at various points while dredging. During clean-up you should get a good idea of your recover rate.
It depends on what you aim is as far as amount of concentrates. If you want to really fine tune your dredge then tungsten is your friend for testing. Lead has a SG of 11.34, Gold has a SG of 19.32, and tungsten has an SG of 19.22. A few other things that I’ve learnt through recent testing are that one size doesn’t fit all - in other words your fine gold recover won’t be as good if you have unclassified larger rocks etc going over the same capture material. I learnt about this from a Roman text about mining and in that they would run alluvial material unclassified through the first sluice, then classify the cons down to a certain size and run it through a second sluice, then classified the cons again down further to run through a third sluice. I ran an experiment where I classified down to half inch and then ran that through a sluice, and then classified the cons down to 1/8th inch and ran them again and got a lot more fine gold.