All the physicists already know that the n-body problem reveals chaos, so that the planets around the sun should undergo deterministic chaos with a Lyapunov exponent of the order of, say, $\frac{+1}{3.5\,\text{million years}}$, [1]. This implies we cannot predict neither forward nor backward in time what was the orbital configuration of the planets around the Sun, when the prediction time goes beyond some finite time of the order of the Lyapunov exponent, [2]. Obviously when we account for any and all the celestial bodies the prediction time decrease significantly, I will say maybe to less than a few thousands of years if you ask me, [3].
Now a problem arises ... The apparent regular behavior of the celestial bodies in the sky is nothing but a short-period section of the whole chaotic picture, let say about a single point in the whole phase space. What does then, within this viewpoint, experimental evidences like "red shift" tell us? Or better to ask, what can they tell us? Does the idea behind "big bang" persist or it will lose its witnesses immediately?
(Of course you better know the distant to moon is altering, and many many other facts that all of you know better than me, the moon getting farther doesn't have anything to do with big bang and the whole universe expanding, does it?)
Edit:
@RobJeffries conducted me in his comments to an article in which a regular behavior was predicted for the solar system within periods far longer than what I had mentioned according to another reference, [1]. Meanwhile, @DirkBruere mentioned in his answer that the overall behavior of the universe can be studied independent of studying the chaotic behavior of every single celestial body in universe, just as the cases for the kinetic theory of gases. According to these two notices I sought to find some references that will describe the matter more thoroughly. These are what I can say for now:
the followings are some passages collected from [4] regarding the subject:
Celestial Mechanics was the prototype of order. The use of perturbation theories in the solar system had proved quite effective in predicting the motion of the planets, satellites and comets. Despite the fact that the perturbation series were in most cases extremely long, their effectiveness was never doubted. ... Poincare proved that most of the series of celestial mechanics are divergent. Therefore their accuracy is limited. On the other hand, it was made clear that the most important problem of celestial mechanics,the N-body problem, cannot be solved analytically. ... A completely different approach to the theory of dynamical systems was based on Statistical Mechanics. This approach dealt with the N-body problem, but from a quite different point of view. Instead of a dominant central body (the sun) and many small bodies (planets), it dealt with N-bodies of equal masses (or of masses of the same order) with N tending to infinity. Then a number of statistical assumptions were made that gave results that could be tested by experiment. ... In this approach the notion of individual orbits of particles became not only secondary, but even irrelevant. ... The most important result of ergodic theory was the proof by Sinai (1968,1970) of the basic theorem of gas dynamics for a hard sphere gas. Namely,Sinai proved that a gas composed of hard spheres moving along straight lines and colliding elastically is ergodic, and has positive metric entropy. A particularly simple case of this theorem is the motion of a point inside a square box containing a circular body. The particle moves along a straight line until it hits the boundaries of the box, or the circle. Such a motion is ergodic. After this theorem everyone was satisfied that statistical mechanics had found a sound foundation. It was thought that cases where orderly motion appeared, as in celestial mechanics, were in some sense exceptional. Yet, as we will see, the real situation is quite different. The ergodic cases of statistical mechanics and the ordered cases of celestial mechanics are both exceptions and the most general case is in between.
What I understand according to this quotation is that the case for gases (ideal gases with such critical properties like having their molecules as rigid spheres and etc.) is in the opposing position regarding the order found in the celestial dynamics and that the general case of the many body dynamics is neither totally ergodic as in the kinetic theory of gases, nor totally ordered as perhaps for planets in the solar system. However, in the rest of the book, if I have caught the idea correctly, it is said that the case for celestial dynamics is not totally ordered or integrable, but case to being integrable with probable windows of chaos. Then solar system and some galaxies are studied for stability and things like that and most of the celestial bodies are found to be stable within their ages or within the whole age of the universe, if I am right. However, I didn't find anywhere in the book the author addressing the overall behavior of the whole universe (maybe my mistake). It is only mentioned here and there that the age of universe is the Hubble time which is about 10 billion years, like the age which is usually accounted for the universe assuming Big Bang has happened. So I conclude, that Big Bang has happened is presumed by the author and he had no doubt in it, while he doesn't strictly address the predictability of the overall behavior of the universe, specially if we know the whole universe has no ergodic behavior and thus is not like the cases for the ideal gases for which we have the kinetic theory. Now, please clarify the situation, what are the reasons for us to assume the behavior of the whole universe is predictable like in the kinetic theory of ideal gases.
Especially have the scientists done a statistical analysis of the the whole universe for them to predict its overall behavior during last 10 billion years? In [5] it is written that Einstein never took the Big Bang theory serious since he thought if one traces the motion of the galaxies backward in time the theory will fail. Do we have any integration of the dynamical equations for the whole universe? I mean something like what is done in the paper introduced by @RobJeffries in the comments, but instead of focusing on the solar system studying the whole universe with all its galaxies and superstructure complexes and etc.
Searching Internet I found only one related study regarding the dynamics of the whole universe and chaos. Check [6] wherein it is stated that chaos came to be dominant in the dynamics of the whole universe at the very beginning of its formation from $10^{-43}\,s$ and lasted only a very brief time, like at least $10^{-36}\,s$ in duration. If this is true, then before that will be unpredictable and I saw, e.g., in [7] that a modern physicist may even use this and say Bib Bang is not necessary at all:
I find it more satisfactory to make a cosmological model where the density and temperature are never infinite. This precludes a Big Bang and replaces it with a different picture of time where time never begins and never ends. This is in contrast with the standard cosmological model where time begins at the Big Bang and never ends during an infinite future expansion. While I cannot prove rigorously that the conventional wisdom is wrong, it does entail the singularities and concomitant breakdown of general relativity that we have mentioned. The existence of plausible alternatives now, however, makes the Big Bang idea less plausible.
And last, but by no means least, a very recent paper [8] claims that if one prefers the Bohmian quantum mechanics over the Copenhagen interpretation of the quantum mechanics, then no singularity arises at all and, hence, no Big Bang singularity is predictable at all:
Ali and Das explain in their paper that their model avoids singularities because of a key difference between classical geodesics and Bohmian trajectories. Classical geodesics eventually cross each other, and the points at which they converge are singularities. In contrast, Bohmian trajectories never cross each other, so singularities do not appear in the equations.
With all these considerations now I still have question regarding the predictability of the dynamics of the universe backward in time (general relativity is not a statistical theory), so that the red shift effect and the cosmic background radiation shouldn't be a strong justification of the Big bang idea. Do they?
[1] as for reference please check this youtube movie of Alex Maloney's McGill course on Classic Mechanics, around the minute 43
[2] as for reference for this check the page 322 of the book "Nonlinear Dynamics and Chaos" by Steve Strogatz, when he is defining and explaining the ``time horizon"
[3] This is my own intuition, so I added the expression that "if you ask me"
[4] Contopoulos, G. "Order and Chaos in Dynamical Astronomy", Springer, 2010.
[5] Javadi, H. & Javadi, A. "Physics from the Beginning upto Now, Including presentation of CPH theory", Etta Publisher, 1386 (in Farsi).
[6] After Big Bang Came Moment of Pure Chaos
[7] Frampton, P.H. "Did time begin? Will time end? Maybe the Big Bang never occurred", World Scientific, 2010
[8] No Big Bang? Quantum equation predicts universe has no beginning