If you know that a question type is one that you are not likely to get right within 3 minutes then this is a good candidate for guessing on the Quant section. If you can get about 25 questions correct on the Quant section you should be able to get a score that is from the 60 to 80th percentile. So you can afford to miss quite a few questions. It is actually worse in terms of overall strategy to take 4 minutes and get a question right than it would be to have guessed at the question.
You are thinking strategically and that is a good thing! You plan is good for Problem Solving. But understand that strategy dictates that you should not merely randomly guess at any DATA SUFFICIENCY question. It is much easier to see that you have enough information to solve a question (Data Suff) as compared to actually having to solve it (Prob Solving). Think about it, what information do you actually need in order to solve a mixture problem? Just the average of each of the components, such as liquids x and y, the overall average of the new mixture, and the total amount of mixture. If you have those things you can solve. Plan on giving yourself at least a minute to quickly look at the question if it is data suff.
On the verbal side, it is also much easier and more effective to narrow down answer choices when compared to problem solving. So when you see a bold-faced question you might take just a minute or so to see if one of the portions in bold is quite obviously a prediction or the conclusion or evidence or something. You may be able to rapidly narrow down the choices. You might even find that you are down to just one choice!
The problem with saying "I am going to skip this type of question always and randomly guess" is that it may wind up being a very simple version of the question. At least on bold faced questions and data suff mixture problems give yourself a minute to see what you can do.
That is my advice.
