This question already has an answer here:
A few classmates and I are worried about the time lag between submitting a research paper to a conference or journal and the further time lag between rejections and re-submissions during which someone else has the chance of scooping our idea.
Based on the pointers to using the Dataverse network to prove that a research idea was created by a certain author at a certain point of time, I wrote to few conference organizers if they would accept the Dataverse network as a reliable reference. I received no reply.
Given that a paper like this with spelling and grammar errors did not get rejected, I conclude that lower tier journals are more eager to publish anything that comes their way. For higher tier journals/conferences, I read the rejection rate is 98%.
So one strategy I can think of is:
1. Publish an initial idea/approach in a lower tier journal, where revealing details are not presented, but the overview and results are mentioned like in this paper.
2. Once the paper is accepted and published, cite the published paper in a newly written paper where all details are revealed, and the concepts/results are expanded/improved on a bit more and send for publication to a top tier journal.
Is this a reasonable strategy that would work or would you suggest a better approach?
UPDATE: I wrote to a number of conference organizers and journals. So far KDD and IEEE responded saying that putting up a document on arXiv is ok, before submitting it to an IEEE or KDD conference. Officially, IEEE mentions it here and there's also their code of peer review ethics here.