Length of publication cycle for peer-reviewed journals Is there any data of the expected length of publication cycle with respect to different peer-reviewed journals in theoretical physics?
I know that the length depends on many things, including the referees, the author and the submitted paper itself. However, clearly some journals have it shorter or longer. While a long response time is not necessary a disadvantage (the referees may actually care to understand paper thoroughly (as it happens commonly in mathematics), while the preprint is on the arXiv anyway), it may influence decision of sending the paper to one journal or another (along with the Impact Factor, the Eigenfactor, etc).
Of course always it is possible to perform a reconnaissance in force. However, with little or no prior experience with a particular journal (1 data point is not enough to estimate the variance) it could be difficult to estimate if it is going to take a few month or 2 years (with all its consequences).
So is there any list of times (mean, variance, median... or the full data) for any of the steps in the following sequence
$$\text{submitted} \rightarrow \text{reviews}\ (\rightarrow \text{response}) \rightarrow \text{accepted} \rightarrow \text{available}$$
for different peer-reviewed journals?
EDIT:
To clarify: I am not interested in mean(/variance of) acceptance time averaging all journals in TP. Rather, I am asking for a list of time vs journal.
 A: For the APS journals (Phys. Rev.), you can find for each paper that is published when it was submitted and when it was published.
It usually takes between a couple of months for "easy accepts" and maybe one year for papers with a lot of back and forth between the authors and the referees.
In fact, here is a document where you will find all those data (and even more) for APS journals
http://www.phys.nthu.edu.tw/~colloquium/2009F/T2.pdf
Spoiler alert for PRL for 2008 (from submission to acceptance):


*

*median: 118 days 

*mean: 139 days 

*min: 7 days

*max: 992 days

A: This is indeed a random variable due to the different procedures adopted by different journals. Just to make an example, Editors of Physical Reviews often decides for a single referee, with a possible option for a further one, while IOP Editors are obliged to select always two referees introducing often delays due to availability. So, due to such different practices, it is really difficult to have a homogeneous sample to do statistics. Besides, one should also consider that letter journals are aimed to speed up these procedures asking for a report in a relatively short time.
On a personal ground, I avoid to use a speed criterion to choice a journal to publish but I prefer other aspects that could be a lot more important than this.
