In my last question about methods of deflecting a rogue planet, there was a fascinating answer from Willk.

Willk basically suggested that if you knew the location, mass and velocity of objects in the solar system, moving for example a small asteroid could alter the path of a larger asteroid which could then alter the path of a small moon... right up to altering the path of a planet.

This set off a whole new story idea for me.

Assuming current or near future levels of technology, except for the computing power which could be a lot further future, is this method feasible for moving a planet, either shifting its orbit, or sending it rogue, or capturing an incoming rogue?

And if it is possible, what timescales would we be looking at?

  • $\begingroup$ If I have made etiquette mistakes in referencing my previous question and an answer or in not correctly acknowledging the answer of another user please let me know. $\endgroup$ Commented Oct 24, 2019 at 9:47
  • $\begingroup$ Referencing previous questions is encouraged (within reason) as it gives context to your current question and shows you are building on previous answers. :) $\endgroup$
    – Tim B
    Commented Oct 30, 2019 at 0:20

6 Answers 6


Short answer: probably not usefully.

You're probably going to end up with a molten space-puddle in the process, which in turn is going to take billions of years to cool, the time required to throw enough kinetic energy at it to get it going in the first place, and then doing it all again to slow it down not withstanding. I can't make an estimate of how long that would take because I know nothing about orbital mechanics, but at blind guess, millions of years minimally when you consider how vast space is, how many objects are going to be involved, and the fact that none of them are under power.

But even then there's the problem of having enough rocks around to do this, and to do it with any synchronicity. Remember our own moon was another planet hitting Earth, and it barely effected our orbit at all, mostly just the direction of spin. You need a lot of mass hitting your planet (and if that mass isn't removed somehow, each hit makes the net process harder and harder—the end state of your planet is going to have a helluva lot more gravity than it started with). You probably want many, many, millions (probably billions, maybe trillions; I don't have a real sense of the maths here, and the size of the rocks is really going to matter) of kinetic impacts all timed to gradually achieve your goal. But let's say millions; millions of destination objects are going to add up to many, many more billions of smaller objects involved in these chains of gradually scaling impacts; and not just untold billions of them, but billions of the right size and right trajectories to make this all possible, which in turn means they'd need to be selected from trillions and trillions of objects that weren't useful (or you'd have to explain why everything in this region of space was so auspiciously arranged, anyway). This is turn is going to raise the question of how a technological society emerged in a region of space this ludicrously dangerous in the fist place.

A lot of computing power is one thing, but the society that could even power the computers that could do this math have much, much, much better ways of achieving the same end.

It's also worth noting, though, that this basically is how everything in the universe is currently arranged—gravity and kinetic energy are mostly how everything's moving anywhere as it is.

Sorry if that's not a very fun answer!


Millions of years.

Yes, we can potentially influence planetary orbits, but this method relies on close approaches between celestial bodies, which happens rarely. Earth has such approaches with asteroids maybe once in a century, end effect of each approach is almost negligibly small. We need a thousand of approaches like that, all working in concert, to affect Earth's orbit to a smallest degree.

A sizeable asteroid can give a stronger "kick" to a planet, but properly directing it would take thousands of years, just because it's big. And it would be able to "kick" just once. After that, we have to start the work over, spend thousands of years again to deliver the next "kick".

Also, we need a surgical precision with operations like that. We don't want an asteroid to hit the planet instead of buzzing close by. For asteroid orbits, there is a "window of uncertainty". We don't know its exact trajectory until time gets close. Maybe advances in computing and science will close this "window of uncertainty" - but maybe not, and then artificial moving of planets would remain a science fiction story.

Potentially, we can speed up this process by relocating a sizeable portion of asteroid belt so we can have a large team of "kickers". This may speed up the process to perhaps one million years, but require the level of orchestrated precision that we currently can't even contemplate.


I read through the answers and believe one thing needs to be said re: your “feasibility study”. It is physically possible, therefore “feasible” depends entirely on your level of technology. You are essentially asking if we can sink all seven pool balls & the eight ball in sequence with one shot. Yes, with perfect information, it is possible.

To make this feasible you need:

  • Extremely precise location and momentum information on every rock in the solar system (not just the ones you are using, because everything has gravity).

  • Enough time to “hit the cue ball” ( IOW the first rock)

  • Enough time for all the other rocks to get into position and move each other (completely random number)

  • Extremely precise location and momentum information for the rogue planet.

  • Enough time for this scheme to intercept the planet before it is too late.

The ballpark range for having all this information, forming the plan, then launching the chain reaction will be between hundreds of years (in a perfect scenario) to millions of years (only limited by the time you get perfect information about your planet.)


A somewhat simpler method would be to adapt the "gravity slingshot" method we use to adjust the orbits and velocity of spacecraft. Probes like Voyager "slingshot" around Jupiter to increase their velocity, but Jupiter actually slows down due to the energy exchange. Since Jupiter is orders of magnitude larger than a space probe, the velocity change isn't even measurable by most instruments, but happens nevertheless.

If an asteroid is shifted to fly past a planet and be accelerated by the pass, the planet will lose some velocity and change its orbit slightly. If the asteroid is decelerated, then the planet will speed up and move into a larger orbit.

What is needed is thousands of asteroids (clearing out one of the Jovian "Trojan" asteroid clusters will do), and fitting them out with light sails and sophisticated computer guidance systems. Each asteroid manouevres around the target planet, adding or subtracting velocity as needed, but then deploys the solar sail in order to readjust its own orbit and gain or lose energy for another pass.

Other things will be needed, such as an effective Space Traffic Control system to ensure the asteroids are not striking other spacecraft, planets and so on, and enough fine control to adjust for the varying mass of the asteroids, or alternatively to have enough space manufacturing capability to build uniform masses for the job.

The speed at which these sorts of adjustments can be made will depend on a lot of variables. The closer the pass the greater the momentum exchange, but a planet with an atmosphere will have a clear limit to how close of a pass you could make. The greater the momentum of the asteroid the greater the potential momentum exchange, but this might require long and complex orbital adjustments as the asteroid moves under power of the solar sail, making the overall orbital adjustment time greater. If we suggest a thousand asteroids can be gathered and prepared to move the Earth, and each asteroid is in a solar orbit of about one year's duration, then roughly three asteroids a day will pass the Earth to do momentum exchange. Obviously we would want more interactions to make the process faster (10,000 asteroids would result in about 30 encounters a day, which would look pretty spectacular as all these solar sails furled and unfurled around the planet).

A British researcher named Paul Birch took this idea to the logical extreme, calculating that using solar energy on a grand scale (about 2% of the Sun's luminosity) and hyper accelerating a stream of pellets like machine gun bullets it may be possible to move planets around in a time span measured in decades.



It's probably not feasible, but close.

The issue is that you will have delays (such as to cool down your billiard balls). This process may take millions of years. The solar system, despite its appearances of a bunch of planets on wires, is actually a chaotic system. Our models of the solar system are limited to a few million years because of this.

Which makes the problem more interesting. Not only do you have to play snooker on a planetary scale, but you need to do so in a way which stabilizes the chaotic nature of the solar system long enough to do it. Instead of single chain of impacts, you would have a woven web of impacts designed to retain the ability to build up to the key impact.

And on that scale, you might as well just focus on stabilizing the chaos to do what you want, rather than playing billiards!


This would not work, not because the physics is wrong, but because the math is wrong.

The Bad Guy here is Chaos theory: Each gravitational interaction is non-linear, and non-linear interactions are prone to producing chaos (in the mathematical sense.)

Wikipedia has a decent discussion of Chaos Theory, but I'll try to summarize. Basically, if you do a close pass between two bodies, small changes in the geometry of the pass will produce diverging orbits and, given time, very large changes in motion. So the first interaction will probably go well, but a one foot difference in the closest approach will, in time produce an position which diverges from the desired orbit by miles. If that first close approach is followed later by a second, the mile-sized error in closest approach in the second case will be magnified yet again to produce huge errors after the second encounter, and yet again bigger for each subsequent encounter.

The math -- which is well-understood -- says that except in certain special cases, the result of a series of close approaches will be a final trajectory which is even in principle unpredictable.

(The special cases are what NASA uses: One approach is to do a course-correction after each encounter and correct the error from the first encounter before it has time to grow. Do this between each pair of encounters and your cosmic billiards can go on for quite a while. The second technique is to find and use relatively small gravitational resonances which produce orbits which don't diverge. (This is what keeps the Solar System stable.) But in general, those resonances are not going to be where you need them.)

Ultimately, you don't need any really complicated schemes. Quite small impulses done at the right time and in the right direction can produce quite large changes in orbits given enough time. (If we have to divert an Earth-killing asteroid anytime soon, that's what we'll do -- get to it ten years before it's due to hit and nudge it so it just misses. This is within range of our current technology for quite large asteroids.)


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .