When determining the best approach for getting to space, the first consideration is the mission objective. For example, placing a small satellite into low Earth orbit (LEO) is a very different mission from setting up a supply chain to a colony on Mars.
Most missions today involve sending small payloads to LEO, but future missions may be significantly larger in scale.
Let's start with the smaller mission. Let's say that you already have a sub-orbital capable rocket, and your challenge is to make it orbital. Your scientists figure out that to do this you need to increase the delta-v of your rocket from 8500 m/s to 9500 m/s.
The performance of the current rocket can be characterized by a curve that tells you the "payload mass" to "initial mass" ratio versus delta-v. This curve is based on the rocket equation but also takes into account things like staging and the ability of your engineers to achieve a certain weight while maintaining reliability.
The chart below shows such a curve for a two stage rocket with a first-stage exhaust velocity of 2,875 m/s, a second-stage exhaust velocity of 3,018 m/s, and a variable mass ratio* of 0.05.
(*The "Variable Mass Ratio" is roughly the dry mass of the rocket divided by the rocket’s initial mass. Think of it as the mass of all the stuff you need to engineer and optimize to make the rocket work over the total mass of the rocket.)
This curve tells you that, if your target is a delta-v of 8500 m/s, the current rocket-plus-payload mass will be 39 times the payload mass. It also tells you that if you want a delta-v of 9,500 m/s (1000 m/s faster) with this technology, then the ratio will increases to 85.
One way to add the needed 1000 m/s of delta-v is to make the rocket 85/39=2.18X larger. Alternately, you can reduce your payload's mass by a factor of 2.18X. A combination of the two, such as a 48% payload reduction and a 48% initial mass increase, will also work.
Another proposal might involve a ground-launch-assist strategy that involves building a railway track and a vehicle that will carry the rocket up a mountain at 1000 m/s and release it.
In the context of this mission the team that proposes to make the rocket bigger and the team that proposes to make the payload lighter are likely to win the engineering debate in the conference room. Both teams have years of experience making rockets bigger or payloads lighter.
The ground-launch-assist proposal, on the other hand, will involve a lot of R&D, capital expense, acceptance of risk, and just won't add much benefit for this kind of mission. It's simply easier to make the rocket bigger and the payload lighter.
But if the nature of the mission were to change significantly, then it could be worthwhile to reevaluate the various ground-assist launch technologies that many people have proposed in the past.
Let's suppose that your goal is to impart enough delta-v to payloads to send them to Mars (you need ~17,120 m/s of delta-v for that - ref) and your mission involves sending thousands of tons to Mars over the next decade.
In this case, the teams that can make rockets larger and payloads lighter are going to be hard pressed to come up with viable proposals.
In fact, to be able to place anything on Mars they are going to need to drastically change the shape of the curve - which means the way that they have historically done things. It means switching to engines that burn hydrogen to achieve a higher exhaust velocity, using lighter-weigh materials (and doing lots of testing) to reduce the rocket's variable mass ratio, adding more stages, not attempting to recover anything, etc. Whatever they come up with, it will be expensive (that is, it will probably look a lot like the SLS). From experience we already know based on past missions that it costs in excess of one billion dollars per ton to place payloads on the surface of Mars.
In this context, a ground-launch-assist technology has a much better chance of being the most technically and economically viable proposal, even if it involves: a) Maturing some low Technical Readiness Level (TRL) technologies, and b) Is relatively capital-intensive. This will be especially true if its proponents can successfully demonstrate the "cost-per-kg versus delta-v" curve overtakes a similar curve for a state-of-the-art chemical rocket-based system.
