Is it possible to know a random choice in the past?

[Gjmdp]

There is a perfect software program that can answer "everything" you ask. Can this software program predict a posible non-existent random choice (of 2 objects) in the past chosen by another perfect software program that answers competely random? The choice is completely random but it never occurred in the past, so its a prediction about how randomness would behave in the past.

I think that if this software program know this answer, the choice wouldn't be random, because the choice is predetermined, known. BUT this would occur if the prediction is before the choice: Is this also right if the prediction is after the choice (that, remind it was never produced)?
Would this program predict right?

 

[WWGD]

I imagine the answer is no for the being able to answer everything part, otherwise the program would know how to make any program halt in a Turing machine.

 

[h6ss]

There is a perfect software program that can answer "everything" you ask. Can this software program predict a posible non-existent random choice (of 2 objects) in the past chosen by another perfect software program that answers competely random? The choice is completely random but it never occurred in the past, so its a prediction about how randomness would behave in the past.

I think that if this software program know this answer, the choice wouldn't be random, because the choice is predetermined, known. BUT this would occur if the prediction is before the choice: Is this also right if the prediction is after the choice (that, remind it was never produced)?
Would this program predict right?

What's particular with ideal randomness is that it is perfectly independent of everything, in particular knowledge and time. If the first machine did randomly choose between two objects, then I would assume the second machine would rightfully predict the first machine's choice assuming it knows everything that happened, and the choice did happen. However, if the second machine would attempt to predict the first machine's random choice prior to it choosing, then it would have to choose randomly.

Your best bet at predicting ideal randomness is ideal randomness (sorry, astrologists).

 

[Stephen Tashi]

You could ask the question without have a sophisticated program that predicts the past. You could simply have a "program" that records what happens in the past and gives that result as its "prediction" of the past.

From that point of view, you are asking if a random choice remains "random" after the choice is made. Mathematical probability theory does not deal with this question - which involves the distinction between probable events and events that "actually happen". Applications of probability theory deal with this question by using "conditional probability" to distinguish the probability of an event before definite information is given and the probability of the event after definite information is given.

 

[_PJ_]

So essentially, computer A is 'predicting' a (truly random) result from B

Well, the issue I think is within the definitions more than the software, and therefore, this is more philosophical.

ruly Random is impossible for current computer computational process*

For B to make a truly random choice, it would require some form of algorithm to do so, and this algorithm would (by nexcessity) be entirely logical - there may be some possibility in Quantum Computing*? Otherwise, the idea of a Truly Random choice becomes something like that idiom of "An omnipotent Being (God) Creating A Stone so heavy, they cannot lift it". The problem at heat becomes that of a paradox in the definition of "omnipotent" = capable of everything with the word "cannot" = incapable.
Therefore, the rationale of a COMPUTER PROCESS leading to a RANDOM result is the contradiction.

However, given that somehow*, B is able to produce such a result, then A with similar programming technique* to overcome the "sequential incapability to result in a randoim output."

Aside from the potentials of Qubits* I have used, really this whole problem is largely circular. The ONLY means for B to create a random result is through a mysterious and 'impossible' technique C - Given this situation, the ONLY way for A to make any prediction of such, would necessitate A also using at least a similar conceptual technique, also as impossible, as C, let's call it 'D' even though it may be extremely similar to C.

This then destroys the principle determination of the randomnisity, since C is a sequence of instructions resulting in the choice B makes, and D is a set of instructions arriving at a result to equal that of C but for A as the correct prediction.

If it's predictable, it cannot be random. If it's random, it cannot be predicted any better than, as h6ss implies, the actual probability.*So far as current technology and software development goes, For B to arrive at a truly random choice, would necessitate some form of quantum element to computation - by the same token, Quantum computers could identify results from all possible outcomes simultaneously, and produce the most likely such results, so A could use this technique to at least examine the possible results from B.