I am finishing my PhD which was quite multidisciplinary (in Physics, it was a theoretical PhD in connection with experiments though), and I am now looking for post-docs.

While I would have some idea of future research plan in the same interdisciplinary, I would like to get specialized in some sub-area for which my knowledge is not very high.

On some post-doc application, I have to write some research proposal (either to earn grant or because the advisor would like it).

I find (quite) challenging to write a research proposal on a topic on which you are not very familiar, and my question is the following: how detailed should a research proposal be ?

In order to "classify" the level of detail I will give a scale for which I would like all the answer based upon (to be sure we have a common language for how "detailed" the proposal should be).

I assume for any of the option below that: you have a well motivated "broad" question that you can explain in your proposal. What I mean by broad and well motivated question is that if you explain it to someone in your field he will tell you "it looks very interesting but this is a vast question!".

And in addition to that there are four scenarios depending on the level of detail of the research proposal.

1. Super detailed: You already have a concrete way to answer it. It means that you probably found a more focused question and you already did some "initial calculations". They look very promising to be continued (basically you almost write the proposal in order to have money to continue your already started "concrete" investigations).
2. Half-detailed: You have identified some possible limitations in current litterature about our understanding of this broad question. Other scientists (or yourself) provided a case study on the basis on which you can start understanding the more broad question. You have some vague strategies in order to investigate the case study but it is highly possible that they will be revealed as dead ends soon and then you wouldn't have another strategy ready (you would have to think much more).
3. Not very detailed: You have identified some possible limitations in current litterature about our understanding of this broad question. Other scientists (or yourself) provided a case study on the basis on which you can start understanding the more broad question. However you don't really know "right now" how to start investigating the case study, you need work already to think about how to answer those "simpler" question.
4. Not detailed at all: Apart from the well defined broad question, you do not really have a strategy to attack it neither (which means not even a simpler question that would help you to start)

Currently I am close to 3 (in the middle of 2 and 3 but much closer to 3). I believe this is clearly not satisfying (something at least as good as 2 would be required, 1 is the ideal but usually not possible I feel excepted if you are lucky). I am wondering if indeed, writing something close to 3 is not enough or if it can be considered as good enough (depending on extra factors).

I would be interested to know what is the status of the person answering (to have an idea of the experience in writing proposal): postdoc, professor, etc.