What is the power of prediction?

What is a scientific prediction? In another thread, discussion about what makes a prediction cropped up. You can see the origins of this thread here.Part of science is prediction. A scientific theory should make predictions on the nature of the evidence that will be found. Here is an example of a prediction that is relevant to the debate: Emile Zuckerkandl and Linus Pauling, 1965To translate. If ToE is true and if Evolution happened then we should see congruence. The theory predicts this congruence, before the congruence is discovered. This is a prediction, we know that classifying organsisms based on cladistics produces basically the same tree as one done using the molecular evidence; there is no reason outside of the ToE why the data is congruent like this.Faith's response to this, in order to present a contra-opinion was I'd like to explore this. I guess the question is, what is the power of prediction? When is a prediction the result of real science, and when is it not? Are some predictions better than others? I suspect the specificness of the prediction, and the number of possible ways the data could be go towards some kind of metric to judge how confirming a prediction is of a theory.Naturally, the focus is specifically in comparing 'real' scientific predictions with ToEs predictions to see if ToE is 'real' science.Is it Science?(?)/Bio Evo? Probably the former I'd think.This message has been edited by Modulous, Thu, 09-March-2006 07:12 PM

OK, so if I was able to predict a specific pattern would emerge out of a possible 100,000 patterns using five different methods, would that be a convincing prediction?What about if I was to predict a very close pattern (within error bounds) out of a possible 10⁷⁴ patterns, would that be a good prediction? All I would use to do this is the ToE.What if the possible number of patterns was massively larger than this? Of all the possible outcomes, I can use a theory to predict what general pattern would emerge before even understanding how to employ the method for testing it?What if I was to predict where in a massive pattern of organisms a given organism would be found? Would that be a prediction? This organism has never been tested before. If predicting where an object will be in space and time is a good prediction, why is being able to predict where an object will appear in a certain 'phase-space' and 'phase-time' not a prediction?This message has been edited by Modulous, Thu, 09-March-2006 08:54 PM

I think Message 13 is an important one with regards to prediction. You say that predicting where an object will be in space and time is a strong prediction. My question: Another important point is predictive power. A theory doesn't have to predict something that has yet to come about, it merely needs to have been able to have predicted it. Another way of looking at it is that if we give someone some science training and the details of the theory, yet keep them blind to certain data...will they be able to predict what that data should be?

It really doesn't matter. Consider them 'states' if you will. It could be just a single number. Being able to predict a specific number or even just a small range of numbers 1 out of 10⁷⁴ would suggest some underlying knowledge about how the number was generated, or incredible luck.

I almost brought Einstein's prediction up the OP, but it looked a bit cluttered so I deleted it. The question now is, how does this compare with being able to predict the structure of the relative genetic similarity of all organisms that exist...ten years before the first realistic methods of sequence comparisons came about?Is Einstein's work 'real' science and the prediction above not 'real' science? What seperates the two?