This large red shift means, or is thought to mean, that the object is receding from us at a rapid rate (forgetting for the moment any Compton energy loss). It also means that the signal originated very early in time, since distance in light years equates to time in years. There are no stars within a distance of less than a billion light years that recede at a rate of one third the velocity of light or more. The relation between observed red shift and age is derived from the empirical relation between magnitude and z, the inferred relation between magnitude and distance (as well as age), and the inferred relationship between z and recession rate as predicted by SRT.
Hubble observed a linear relation between the red shift, denoted by z, and the magnitude or brightness (or better faintness), and by inference, therefore the distance, and therefore also the age, of stars. Except for the enormous energy in supernovae, we can't see individual stars, or even individual galaxies, at these great distances, so it is a matter of the brightness of clusters, of clusters, of galaxies - except in the case of type 1A supernovae. Type 1A refers to those supernovae which lack hydrogen in their spectra. This differentiates them from other stars and supernovae, and suggests that they originated near the very beginning of time. These supernovae give us the only view into the very distant past based on individual stars. All other, similar, astronomical observations have been limited to a relatively short time, in the past, far less than a billion years, and correspondingly, distances of less than a billion light years.
In a Euclidean space recession and expansion are different concepts. The solar system may be moving away from the origin of the universe, giving rise to an expansion velocity that was largest early in the life of the universe. But a recession velocity is based on the distance between a star and our present position, and would depend on our two positions relative to the origin on the universe. That distinction cannot be made under relativity theory, but is revived here. Roughly speaking recession relates to two points on the periphery of a sphere, whereas expansion relates to the distance between a point on the periphery and the center.
Since the type 1A supernovae are very ancient phenomena they are very faint and can easily be missed. There are teams of observers scanning the sky and collecting as much data as can possibly be identified as belonging to these supernovae. It is a heroic effort, and the data continue to dribble in - a few dozen, or a few hundred can be harvested each year.
Until recently it was thought that supernovae, and in particular type 1A supernovae, were relatively rare. They are very faint, since the signals originate well over two billion years ago. But new techniques have succeeded in harvesting a dozen or more per year. Many are missed, and we are nearing the end of the line of these early supernovae. But we can conclude that if we had started to collect data, say three billion years ago, we would by now have observed about 10^12 such type 1A supernovae (not to mention all the other ones). It would appear that they are the source, not only of the heavier elements, but of the background radiation that is still permeating the universe. It turns out that the smaller the star, (but at least 1.4 solar masses), the longer it can burn before exploding as a supernova. Those with a mass near 1.4 solar masses have a life-time of about three billion years. Stars which have a smaller mass do not end as supernovae.
The small stars are those that lasted the longest and whose supernovae are still visible at present, the larger ones would have exploded and been seen much earlier. The time it takes for the light to reach us from the explosion, as can be inferred from the red shift ,is upwards of about three billion years. (The relation of time to red shift is approximately linear out to about a value of the red shift, z, of 0.3 under any view.)
One thing is clear. If the type 1A supernovae come from stars created near the beginning of time then the arithmetic which combines their life time (about three billion years) with the time until the light from the explosion reaches us (also about three billion years) gives us an estimate of the age of the universe of about 6 billion years. That is less than one-half of the 12 or 13 billion years that is currently believed. There is no way around this. Current estimates of the age based on Hubble's Law, based in the end on relativity theory, cannot be reconciled with these numbers.
We should also note that if we postulate a center of expansion and decreasing velocity of expansion as a function of time, we have the problem that we can never look back to the origin of time. In a model proposed (See Appendix 2, ref 4), we are never more than 3 billion light years from the assumed origin, so the light we see from this region would have originated 3 billion years after the beginning of time in order to reach us now, (that is at the age of 6 billion years). Looking back to the beginning of time is an impossible task.
The values of z observed from type 1A supernovae point to a smaller and denser universe. An observed value of z = 1, that is a Doppler factor z + 1 = 2 would be achieved by the recession of the star at the speed c/2, and no movement of the receiver. Recession velocities greater than one-half the speed of light would result in values of z greater than one.
The chance of finding such a large value of z is slim. If a line is drawn from the present position of the solar system through the presumed origin of the universe and extended, that would be the region from which such a value of z greater than one could perhaps come. The z values of type 1A supernovae, in the data presented in 1998, lie between about 0.36 and 0.85, which is consistent with any of these views. The universe, and the Doppler shifts resulting from 1A data, can be shown to be consistent with a Newtonian model with a present age of about six billion years, and a diameter of five to six billion light years. The age is, in any case, constrained by the time line - the lifetime of the star + the time until we see the supernova. The time for the star formation is a minute fraction of a billion years and can be ignored.
If part of the shift is the result of movement, which results in an increase in wavelength, and a significant part is due to the loss of (Compton) energy in the signal, then the "sum" of these shifts can be very red indeed - and we should not infer distances or times based on merely observing a red shift. The large values of z, that are being discovered in radio astronomy are probably as much the result of energy loss as of recession (wavelength increase) and thus cannot be used to infer the age or size of the universe as well as signals from supernovae, that is to say individual stars, in or near the visible region.