If they are truly independent and/or random, it would be inevitable that all six would be on simultaneously at some point. 42 Amps for a few seconds might will open a 20 A OCPD.
Now if you could interlock them to insure no more than two were on simultaneously, you would be able to put all six on a 20 Amp circuit.
Let us imagine that these loads operate in such a fashion that in any given second each load has a 1% chance of operating, totally randomly, and that the load operates for the entire second starting and stopping precisely on integral second boundaries.
In this case the chance of all 6 being on simultaneously is 1/10^12 in any given second. Wait long enough and it is a virtual certainty that they will all operate...but 'long enough' in this case is on the order of 32 thousand years. (Assuming totally random, then they _might_ all operate at the same time in the first second, and after 32K years there is no guarantee that they will have ever operated at the same time, but if you wait long enough then it will eventually happen.)
The chance of 4 being on at the same time is 1/10^8 in any given second, or about 1x per 3.2 years.
My point is simply that _if_ the CMP thought it were worth while to provide the appropriate calculation there would certainly be a 'diversity factor' permitted, just like there is for things like electric ranges. However there is no code provision for permitting this sort of calculation, so to be code compliant you need to fall back to assuming there is no diversity factor unless you have some sort of interlock arrangement.
IMHO if these loads are as described, and all were placed on a single 20A circuit, there would never be a problem. I'd bet a single bottle of good beer on this outcome.
IMHO if all these loads are as described, to be code compliant you need to limit to 2 per 20A circuit.
-Jon
