Not quite. The basis of the estimates is known quantities as defined by the hard sciences including atomic physics using reproducible testing methods and reproducible calculations and deductive comparisons. However, there are assumptions that are made based on those finely tuned and EXTREMELY PRECISE measurements, so you are not wrong in all dimensions. The critical flaw is that even if you have a bunch of wild assumptions, you can actually do a set of comparisons that use deductive and inferential logic to test for the verifiability of those assumptions. An example of a real world comparative (and ongoing) test methodology is looking at the deposits of fossil species in various strata. So what you do is make a bunch of guesstimates about how strata are formed and their ages using well known mathematical analysis of atomic behaviour from physics and mary it with slightly more fuzzy but well understood and confirmed geological theory (it's not perfect but it gets better and better over time and is tested in industry all the time, all of the oil exploration is just geology writ large) and you develop a set of claims about the ages of various strata knowing that your guesstimates will be plus or minus a certain amount of tolerence. Then you look at the deposits of fossil species based on that stratification and start to notice that some species seem to be deposited way way earlier than others. Now, in and of it self, that doesn't really tell you anything. BUT, if you use comparative analysis and start to notice that ALL FOSSIL DEPOSITS OF A SPECIES are ALWAYS deposited 40 million years ago and another species of fossils ALWAYS AND ONLY deposited in strata about 200 million years ago, and the two NEVER EVER EVER appear at the same level of strata you start to get an indication of how things start to work. So, no, it isn't a perfect science, there is a lot of guesswork but not as much as you would suggest. The whole thing boils down to this set of opposites: complete bullshit <------> absolute truth All they are doing is collecting as much data and using as much analysis as possible to move along that line toward some value of absolute truth. Science never gets to truth (that is only allowed for our gods) but humans can make a good case for how things look. The keys are reproducability (which it all is) and that the fields that you need to master to fully be able to reproduce any of it are so vast and require so much time as for the field to be effectively impenetrable and making it seem that guesswork is bullshit when it isn't any more than any claim you can make about the universe. What is actually really useful about the bullshit <----> truth diagram is that depending where you are on that line you can actually calculate the probability of two things: * The degree to which you can be wrong about your claims. * The number of possible alternative explenations that you can expect to be valid possible explanations.