He assumed that the speed of gravity (the speed at which gravity waves travel, and can thus affect other bodies) was the same as the speed of light, but until recently, it hadn’t been actually measured. Now it has.

I was bothered by the following grafs, though.

Fomalout and Kopeikin said their results are accurate within about 20 percent.

Knowing the precise speed of gravity is important to physicists testing such modern ideas as the superstring, which holds that fundamental particles in the universe are made up of small vibrating loops or strings. It also affects some basic space-time theories.

There is a difference between accuracy and precision. Accuracy means that it’s in the ballpark, while precision indicates that we can measure the value to a large number of decimal places. It’s possible to be accurate, and not precise, and to be precise, and uttely inaccurate (i.e., precisely wrong). As an example, if I said that the speed of light was 1.3578456457746456456 meters per second, I would be giving an extremely precise answer, that wasn’t within six orders of magnitude of the correct answer.

Results that are accurate within only twenty percent are hardly precise, and in fact, don’t provide full confidence that the speed of light is indeed the speed of gravity.

We can never know that for sure, of course, no matter how precisely we measure. It might be that the speed of gravity is exactly 99.99999999999% of the speed of light, rather than equal to it, and we may never know, for sure. Similarly, the exponent in the universal law of gravitation might actually be 2.00000000000000000000000000001 rather than exactly two (so it’s not exactly a function of the distance between the bodies squared), but Occam’s Razor indicates that that’s pretty unlikely, particularly since there are sound theoretical and geometric reasons to think that it’s truly a squared law (having to do with the nature of spherical fields).

Similarly, if the two speeds are measured to be within twenty percent, it’s likely that they’re truly the same (particularly since there’s no reason to think they’re not), and the likelihood will increase as we get better measurements, with less error.