Can someone enlighten me in which cases or in what kind of Systems a Trade Synthesize selection in monte carlo would be more appropriate to use? Could you suggest any reference to take a look regarding the appropriateness of each method with respect to the trading system specifications.

It's probably easier to indicate the type of system that Trade Scramble and Synthesis are not for. For example, a dip-buyer type system that may enter a few trades here and there, and then a large number of trades on a "shock" event, only an equity-curve scramble seems appropriate in order to model those price dip events. Otherwise, randomizing the trades throughout the entire history would probably just serve to create a much smoother and gradual equity curve that would ever be achieved by trading that type of system. (You can try it and see if I'm right.)

For strategies that have a more even or random trade distribution over time, the Trade Scramble and Synthesis methods are appropriate; each method and option providing increasing degrees of randomness. See the descriptions in the User Guide. For example, you might start with a Trade Scramble. But if your Position sizing method is based on the equity curve (% of equity, max % risk, etc.), you'd probably want to see if Trade Scramble and Randomize makes a significant difference in the average result.

In any case, there is no right and wrong, just investigation. The goal is not to rely on the results of a single simulation, which is just a stochastic series of events that are unlikely to occur in the same order ever again." The question Monte Carlo answers is "What if... the [base] simulation had occurred in this sequence...?"; and it gives you thousands of answers to the same question, providing probabilities that you can use to judge what outcomes to expect in the future - best case, worse case, and average.