Would you tell me what Bayesian methodology is used for?
There are many reasonable ways to answer this question. JayUtah has answered it one way. Let me answer it another.
Bayesian statistics is used for exactly the same types of problems as non-Bayesian, or classical/frequentist statistics. There are exactly three types of such problems: (1) ascertaining the plausibility of a hypothesis (hypothesis testing); (2) estimating an unknown value (estimation); and (3) predicting a future event (prediction). Jabba is attempting to use Bayes for #1: ascertaining the plausibility of a hypothesis about immortality (or some such thing). I will explain Bayesian statistics by using an example of #2: estimating an unknown value.
It's easiest to explain Bayesian statistical inference by contrasting it with classical statistical inference. To that end, suppose we choose a coin at random---we just pull one out of our pocket---and we want to estimate the long-run probability that when flipped it will land heads. We decide to flip the coin 5 times, and we find that it lands heads every time. In classical statistics our best estimate of the long-run probability that our coin will land heads is 100%, because that is what we observed. Is that a reasonable estimate? In one sense it is: if the coin were fair (ie, had a 50% chance of landing heads on each flip), the probability that it would land heads in 5 out of 5 tosses would be 1-in-32. Thus our results mildly suggest that the coin might be biased in favor of landing heads, or maybe even 2-headed. So, on the basis of this little experiment, the classical statistician estimates that the long-run probability that this coin lands heads is 100%.
Although the classical estimate is reasonable in one sense, it is unreasonable in another: it doesn't take into account our prior knowledge of coins. Bayesian statistics does. The Bayesian estimate of the long-run heads probability is a weighted average our prior knowledge about this probability and the results of our experiment, and the weights are determined in a mathematically optimal manner: the more prio knowledge we have, the more weight our prior knowledge is given. Sometimes we don't know very much before we do an experiment. For example, if we are testing the effectiveness of two new drugs against each other, we may have no good prior information about their relative effectiveness, and so the Bayesian solution would be heavily weighted toward the results of the experiment over our prior knowledge.
But sometimes we know a great deal about the subject, like with coins. We're pretty sure before doing the experiment that the long-run probability that a randomly selected coin will land heads is close to 1/2. The fact that we observed 5 out of 5 flips land heads should have
some effect our estimate, since, after all, there is
some chance that we have picked a defective coin. But the odds of that are slim. So our Bayesian estimate will weight our prior information much more heavily than the observation of the outcome of just 5 flips, and the Bayesian estimate will consequently be close to our prior estimate of 1/2, shaded upward just a bit to take into account the somewhat unusual results of our experiment.