Final shape refers to the version that, you feel, is truly representing your work and is fit for submission to a journal. Here I do not consider the formatting issues.

i think the iterations depends on the package you are using and what you intend to do with it

Not sure the final shape is after blind review by reviewers or before submitting for blind review.  If the final shape refers to after blind review, then think number of iterations necessary for a research paper before reaching the final shape depends on several factors inclusive:

1) How critical the blind reviewers - if they are very critical, this will take many iterations e.g. 5-6 times in order to address all their critiques.  Some of their critiques come later & not surfaced during initial round(s).  Sometimes we also need to harmonize between different reviewers as different reviewers have different opinions on the same issue.

2) How ready is the research paper before submission for review - if it is well prepared or less errors i.e. self-reviewed several times after some intervals / getting other scholars to review, this might take fewer iterations e.g. 1-2 times.

3) How reputable is the journal one submit his or her article - for more reputable journals, normally their reviewers are more critical, hence might need more iterations of reviews.

4) How new to knowledge contribution is the article - if the article is relatively new to knowledge contribution, there is possibility of more iterations especially for submission to more reputable journals with more critical reviewers.

Final shape refers to the version that, you feel, is truly representing your work and is fit for submission to a journal. Here I do not consider the formatting issues.

This is an interesting question and at the same time very difficult to come out with a quantitative answer. The number will essentially depend on the writer and how self-critical he is for his own product and efficiency / capability of the person. 

There are, in general, three types of errors / scope of improvement in a write-up: (a) Language related including typographical (b) Subject related mater and (c) Way of presentation (Table / Graph / Bar Chart  etc. ). 

I have observed that each time I made a review or revision of my draft, I discovered new ways of improving the same. Probably, this is true for most of us. But the ultimate question is how long and how many times we can afford to doing this ?

Hi, I think final shape should represent your work only. Therefore, in generally speaking they should in formats like (1) format of paper according to publisher(2) language standard along with equations, pictures, charts, dataset should be proper international standard. 

Furthermore, The quality of work. and new research method, quality of work and future scope followed by result module. 

Final shape is defined only by author - the researcher. I think that shaping the research paper differs in different scientific fields. In engineering, I return to research sometimes to try getting some more info that could be good for research paper. It is an iteration also. Improvement and shaping the research paper to its final version before submission for publishing, is process which differs from case to case.

I think the attached article is relevant in connection with the present question.

The write-up is taken from the following book

ESSAYS IN COMPUTING SCIENCE

C. A. R. Hoare

Oxford University Computing Laboratory

C. B. Jones

(editor)

Department of Computer Science,

Manchester University

PRENTICE HALL

NEW YORK

The first two pages describe how much critical was Prof. Hoare about his presentation.

Well it certainly depends upon the researcher and his knowledge. The more you revise the more perfect the document become. Here we have to adopt a self consistency approach. If there is   little difference between two consecutive revision, then it is ready for publication!!

To this question is no simple answer. Sure you can seek out some general principles. Certainly important is how the writer deal with the problem professionally, whether the problem is thoughtful and intellectually closed.

If not, a writing of every article is difficult, and the number of attempts can be considerable and without much success.

Every problem must first ripen professionally - before any writing.

I know the answer can not be a unique number. However the discussion may give us some idea about the method of academic writing normally followed by the researchers. It can also tell us, I believe, the considerations that has to be made during the preparation of a research paper.

It depends on researcher to researcher. One makes the paper fine in two readings; another may take five readings to male it fine. 

I would like to highlight the point made by Mandumpal. If the difference between two successive revisions is little or negligible, then indeed the draft is very near to submission. 

Well 

Well it truly depends on the researchers , when they feel their work is ready for publication. In my view, the final version of paper is not dependent on the number of iterative revisions  made to it, rather I believe that the paper is ready when the author and his peers have made substantial corrections and deem it fit for submission.  I know friends who write the paper, revise it as per feedback from coauthors and supervisors just 2 or 3 times, and submit the paper, letting the remaining  corrections to be identified by the reviewers, as long as the technical stuff in the paper is not flawed