There are no hard and fast rules for sample panning a creek - just a lot of sound principles. One important principle is that sampling should be systematic, not random. Sampling determines two things: 1) is there 'worthwhile' gold being deposited in that location; and 2) where is the highest concentration of deposition.
Fly-poo sized gold is not a good indicator of localised deposition because it tends to be widespread and, if deposited in any quantity, is more highly concentrated in the first 5-15cm. Where I prospect, I could take a sample pan from the surface just about anywhere and find 1-4 specks of fly-poo gold 8 times out of 10.
On the other hand, if a sample pan contains 20-30 fly-poo specks, then I'll stop and investigate further because this indicates a localised deposition rather than the general wide-spread deposition. I will sample close by to try and identify the highest concentration, and then sample deeper to see if there are any heavier bits also being deposited. If I do not then find anything larger than fly-poo specks, I will move on because there is not much to be gained by highbanking for just ultra-fine gold.
Therefore, I use sampling primarily to find where fine gold (or larger) and not fly-poo is being deposited.
If your sample pan throws up a few bits of fine gold (or larger), then continue sampling close-in around that spot, at intervals less than 30cm and in all directions, again to try and find the highest concentration. As a general guide, such sampling may cover an area of several square metres, and may go down more than 30-50cm. I look for sample pans with more than 10 bits of fine gold, before I would consider breaking out the highbanker.
If your localised spot throws up some fine gold (or larger), your sampling should systematically cover the entire immediate vicinity, perhaps going out more than 5-10 metres (and possibly even more depending on sample results).
Even if I have a bonanza sample pan (20-30 bits of fine gold), I will continue to sample pan to define the extent of the deposition before breaking out the highbanker. And even then, there is no guarantee that the highbanker results will be a true reflection of the sample panning results - "gold is where it is."
And I agree all that Twapster has written in the previous post.
Good luck.