I've done a comprehensive study of all the published guessing strategies I could find, evaluating how effective they are. Those guessing strategies that do not require you to understand much of anything about the question are universally either completely ineffective (they are no better than guessing randomly) or are so rarely applicable that they have essentially no value - the time you'd spend learning them would be far better spent actually learning math or verbal skills.
Those 'guessing strategies' that truly work require you to understand something about the problem you're trying to solve. For example, estimation can be a very good way to rule out several wrong answers in many Quant problems. In fact, using estimation and other basic logical principles can almost double your chances at guessing correctly at a random PS question. But if you're looking for strategies like "guess C on inequality DS questions", those kinds of strategies are worthless.