Point, though that's obviously something that is actively transmitting in the radio band right at you, and not taking any real steps to avoid detection at all (in fact it's actively trying to get your attention). There's plenty of countermeasures an attacker could take that would make detecting them more difficult, such as any combinations of:
1) Only burning far away, and then "coasting" the rest of the distance.
2) Using vacuum insulation between an inner and outer hull, and then coating the outer hull with an extremely cold substance after any burns to reduce thermal profiles.
3) Approaching from the other side or next to a warmer thermal body, such as a planet, thus letting its heat signature drown out your own.
4) Approaching from the direction of a more active background area, such as along the milky way, thus lowering the difference between your own thermal output and theirs.
5) If you do have to do a burn that is close enough to be easily detected, launch a handful of decoys in different directions that are hotter but smaller, then have both you and the decoys "go dark" by cooling their outer hull. The end result would make it very difficult to tell which direction you went in.
To build off the black plane, I'd say it's more of a matter of two people in a black plane with a background that is constantly twinkling slightly, with a few very bright lights scattered about, and with lamps that are only really visible when they are "pushing" and the rest of the time are basically invisible; still possible to track someone down in, but much more difficult.
Hee. That is a point, but bear in mind that said 20W radio signal is even smaller than the thermal radiation that would be output by the ship itself, much less any modern or probable near-future maneuvering system. As for the proposed countermeasures:
1. Burn-and-cruise is viable if your operational planning is on the level of months or years, and your target has no "meteor watch" instituted to deal with more mundane unpowered threats like incoming asteroids. It will need to be implemented at tremendous distances - a significant part of the problem is that any burns, either to maneuver or accelerate/decelerate, will tend to be seen from tremendous distances, on the order of hundreds of astronomical units. Burns powerful enough to cut that time down will also be much easier to detect. Even if you ignore the detection issue, this is potentially highly dangerous if the situation changes at all over the course of those months or years, however, say if military forces get redeployed or if there was even a minute imperfection in one of your ship's thrust nozzles - remember, at these distances, differences of a fraction of a percent point between vectors will translate to at least tens of thousands of kilometers. If you're targeting a mobile force (say, an enemy fleet), that won't work at all, unless it's willing to politely wait there for a tremendous period of time for your arrival.
2. The entire problem with thermal detection at a distance is the thermal energy being radiated through the radiation from the ship to the detector - the detection problem for a ship with no constructed shell can also be treated literally as a ship with a shell the distance between the detector and the ship. Your shell will catch all of that thermal radiation (plus whatever is conducted through the support beams, since that will be non-zero), certainly, but that will just heat it up to the same temperature over time. If you build your shell large enough so that it won't heat up quickly (since the rate of thermal exchange is proportional to the ratio between the surface area of the outer and inner shells), you start to risk occluding entire stars - someone's going to notice if Polaris suddenly blinks in and out. If you have a magic coolant that you can apply without generating more thermal energy, why not apply it directly to the ship? How do you keep it from heating up as well just from regular shipboard operations, even ignoring the additional thermal pressure of burns?
3. That requires a warm thermal body that will take you from your launch point to your target. At interplanetary distances, planets don't typically move that way, and you'll have to be very, very lucky with asteroids. That's a viable tactic, but only at very, very close scales - say, if you're fighting in LEO, where you're using the Earth itself as an occluding object.
4. Also possible, but you need a backdrop that's literally hundreds of Kelvin hotter than the cosmic background radiation, which means it's going to need to be hot enough and/or close enough that the inverse-square law hasn't frittered all that away - say, heading from the Sun to Mercury, but not so much from the Sun to Jupiter. It also runs into the same problem as the (unmentioned, but related) idea of radiating your heat away from the target - it only works if your target is the only detecting source. If they have even two detection platforms separated by a significant distance, you'll need a backdrop that can hide you from both. (
EDIT: Also, don't forget (like I almost did) that energy is conserved. If your backdrop is around as hot as you are or hotter, your ship is also going to be absorbing and reemitting thermal energy from the backdrop as well, driving its own temperature up even further. You can reduce the difference, but it's questionable whether you can reduce it enough.)
5. If the decoys are smaller and hotter, they'll immediately be able to tell the difference from the thermal difference between the source and the decoys. If they're smaller and just as hot, they'll see the acceleration curves are different (remember acceleration is force over mass). If they're the same mass and the same thrust, why not just use more ships?
No twinkling lights here, sorry, and the only bright lights that won't be well-known are likely to be other people. Which, mind you, is a fair notion of subterfuge, if you can make your warships look like freighters from their mass/acceleration profiles.
I had a crazy idea and I want to know if it's crazy like a fox or just crazy. I'm thinking about airbraking spaceships without exiting them from orbit. The breaking length of quality fishing line is 350 km while LEO "starts" around 120 km. So I'm thinking, would it be possible to equip a spacecraft with a "kite" that would reach down to the upper atmosphere on a 300 km tether? That way you could slow down for a stable orbit or an orbital rendezvous without expending fuel.
Suppose the kite was at 50km in height. The atmosphere at that level is 10^-6 g/cm^3. If the kite was moving at 10km/s (the speed of the Apollo 8 Trans-lunar injection departure), it would be creating 100 kN of drag per square meter of kite area, at least at first. That is the same as the rocket thrust for the Apollo 8 injection burn. But it would actually be a pretty small burden for the tether, most of the tether burden is the tether weight itself.
A complication would be that the kite moving at supersonic speeds would have all kinds of crazy aerodynamics to worry about. But maybe those crazy aerodynamics could be put to some use. The kite would be creating a pocket of pressurized air so maybe some of that air could be siphoned up the tether. This way there would be a source of volatiles for the spacecraft. With an ion engine it might even be possible to refull your tanks with nitrogen at earth by braking, go fly off to a different planet and then slingshot back to earth again to repeat the process.
So... crazy like a fox or just crazy?
Not crazy at all, and you don't need a kite. If designed properly, the ship itself can be used to aerobrake, as several probes we've launched (Magellan, multiple Mars probes) have done. ^_^
EDIT:
Here, if you're curious. What you're describing is likely a variation on the trailing ballute design they outline. ^_^