You have the right idea. For very small payloads launched by very small rockets, it becomes increasingly attractive to get the rocket above as much of the atmosphere as possible.
And you are in good company, Virgin Galactic's Space Ship Two, Orbital ATK's Pegasus, and Scaled Composites' Stratolaunch are all current or near future air-launch platforms. Incidentally, Wired Magazine's article about Microsoft co-founder Paul Allen founded Stratolaunch points out that it is going to be the largest airplane in the world.
There are several helpful answers to the question Can a miniature Saturn V get to the moon and back? and you should read them all, but I'll summarize the issues as aerodynamic drag and "gravity-drag", and just say here that as you scale a rocket down in size, the thrust decreases faster than the drag. This ends up requiring the rocket to accelerate more slowly in the dense part of the atmosphere to decrease the drag, which causes a secondary "gravity drag" penalty, the fraction of the thrust that is needed to simply keep the rocket from falling back to Earth. It's a little like having to push down constantly on the pedal of a bike just to keep from rolling backwards down a hill, except that rockets have to push by spending active thrust because they aren't touching the ground.
This is why small rockets tend to be very long and skinny, to pack more thrust behind them while minimizing area. This only works to some extent because the sides of the rocket also generate drag, otherwise we'd have ridiculously long, skinny needle-rockets.
So for a 500 or 100 gram payload and a small rocket (which seems like it might be affordable until you do the numbers), getting above the air is a great idea!
The big problem is that when you make rockets smaller, you end up making other compromises by making it cheaper and simpler. This means the performance is lower. It won't be a "mini Saturn-V". This comment by @Deimophobia links to the analysis done by Dorin Patru, Jeffrey D. Kozak, and Robert J. Bowman at The Rochester Institute of Technology, and presented at the 20th annual AIAA/USU Conference on Small Satellites as A Custom Launch System for Satellites Smaller Than 1 kg.
The LEO orbital velocity, i.e. 200-300 km, is $v_{LEO}$ = 7600 m/s. A total $\Delta v_{LOSS}$ of 1600 m/s is assumed, due to (1) thrust-atmospheric loss, (2) drag loss, (3) gravity loss, and (4) maneuvering and launch window allowance. The thrust-atmospheric and drag losses will be much smaller compared to a sea-level launch, due to the very high altitude of the entire powered flight.
Also you end up making it relatively heavier. What does that mean? A really high performance big rocket can have well over 90% of its mass as pure propellant. I think the record is 94% or 95% but I'm still looking for a source. But when you build a small rocket, this gets really difficult. Making all the structural components, containers, and the engine as small and lightweight becomes a real challenge, and that means expensive materials and manufacturing techniques not available to hobbyists. The RIT work explains that the test stage they built never aimed for the "structure to propellant mass ratio of 1/10" since it was for ground tests. This would have been even harder.
They show that for a smaller scale rocket, using a fairly safe hybrid $I_{SP} = 230 \text{s}$ engine design that is still way beyond "hobby level", you'll need at least $100,000, a team of people, some radio equipment, and about a 200 kg, four stage rocket to put 1kg into LEO.
Space-capable rockets are much more difficult and expensive to make than hobby-rockets, as illustrated by the engine comparison in this answer.
You also need to file several documents and receive approval from some serious government agencies.
The same rocket would have to be much much bigger to launch from Earth, I'm guessing 500 to 1000 kg, but the point is it's still not going to be a "hobby rocket" to get to orbit from a balloon. It is still quite a technical challenge, and if you can build this kind of challenging rocket, then when you are faced with the choice of the hassles of the balloon versus just making the rocket bigger, growing the rocket becomes more and more attractive.
If you'd like more explained about the RIT paper, leave a comment and I'll expand the discussion.
below: Screen shots from A Custom Launch System for Satellites Smaller Than 1 kg.
Tags
menu and then type "balloons" and see what what balloon-related questions look like. You may find helpful answers there. $\endgroup$