I do not have any statistics but sending a MS for review does not means that the MS has high chance of acceptance. The acceptance depends largely on reviewers comments and then subsequent evaluation of those comments by the Editor
Once the paper has passed the screening of the editor and gone for review, your chances can be up by say 65% (because editors are pretty thorough these days).
If a manuscript scales through editor's review, it surely meets the minimum quality and scope of the journal, hence, we could assume it has 50 % chance of acceptance.
If you are talking about top journals, one cannot estimate the chances. It is always different from journal to journal. The chances of acceptance heavily rely on the scientific content of the manuscript as well as the assigned reviewers. It is always advisable to go through the scope of the journal and its previously published research papers thoroughly so that one can self analyze the appropriateness of their work according to the journal's criteria. This would definitely increase the chances of acceptance. Rest depends on the reviewers.
The fate of the manuscript can not be predicted for a favorable acceptance since the status changed to Under Review. One needs to wait until the review reports are received. If one is a editor, he knows the percentage acceptance of articles passed to review process. However, this percentage acceptance will be a mere number and can never be used to predict the outcome for a specific manuscript. Each manuscript is to be considered case by case and the outcome can not be correlated to each other.
But let me tell you that since the paper already passed the initial editorial review, that means it's within the scope and area of interest. The reviewers will come up with comments and ultimately the editor may approve, reject, , or more commonly ask for revision (major/minor).
Yes, there are other factors to consider also like if you have previously published in that journal, your h-index, etc. I know one editor in chief who showed me his/her first quick screening method.