Signal Strength Randomizer(flat and binomial):
-Outputs random signal strengths (0-15)
-Can be distributed flatly or binomially(either 2 gates or interface option)
-Otherwise functions like randomizer

3 outputs: random value from 0-15
1 input: changes values on signal, should function exactly like randomizer
Uses: minigame construction, various random assignment of items for multiplayer means, etc.
Current construction (these only produce one output):
Flat: 2 randomizers, a 4by16 cell array, 16 inverters, line of redstone
Binomial: 6 randomizers, 30+ repeaters, 16 pulse formers, 16 counters, line of redstone, has about a 2 sec delay

Bundled Randomizer
-outputs basically act as 16 randomizers into three bundled cables
-otherwise functions as randomixer

3 bundled outputs: each color of each bundle random
1 input: changes values on signal, should function exactly like randomizer

Uses: same as signal strength randomizer

Current construction: 16 randomizers, a very large amount of wire.

This is the first proposal involving analog gates that has included
reasonable examples, explanations and use cases. I will plan a couple of
gates that should be useful for these scenarios.

I can see good uses for the first one, but not sure if the second one is necessary if the first one is implemented (ei, a single randomizer can be used per needed color).

Am I correct in understanding that a binomial randomizer would be if each
signal had a 50% chance of being set and then the highest level was output?
So 15 is 50% 14 is 25% etc

@Chicken-Bones I meant a binomially distributed chance: if there are 15 random bits, how many return true. http://en.wikipedia.org/wiki/Binomial_distribution

@MrTJP The second one is mostly for uses treating a bundled cable as a 2 byte number, but that one is much less obnoxious to build with what already exists. Thank you both so much for your work on this project.

I cannot see how an analog gate with the following output would be useful.

I can however see uses for one with diminishing probabilities like I suggested earlier

The binomial distribution is primarily useful for more important gear in the multiplayer distribution example to bring a focus towards the center so that a wider range of rewards/starting items can be used, without making certain ones too common. many videogames use either this or a triangular distribution to achieve this effect. I personally would primarily use the values from 3-12 and treat the outer range slightly separately. The distribution you suggested earlier would be easy to reproduce from 5 randomizers or a bundled randomizer as I suggested.

I feel that if you want something as complex as a binomial distribution, you should use a mod like computercraft, or design some redstone circuit that works well enough for your purposes.

I will however add a bundled randomizer (with single and 16 bit modes) and an analog<->digital converter

That should do enough of what I need. I can make do with a triangular distribution or some such until i figure out LUA. Thank you so much for all this trouble.

Gates added