I'm not sure 100% if this is how it works, but you'd presume that if say marquis goes purple then it's cause the sample is predominantly MDxx, and that if it was say a weak purple with hints of yellow you'd maybe assume that that sample was more likely to be mdxx+methylone mixed than say a sample that went very deep black. It's not an exact science unfortunately and this area is where test kits fail somewhat, but in my experience I can say that when a purer material was used (I subsequently have had some samples analysed at a lab), that a darker, faster reaction occurred with more fizzing/smoke. Though bare in mind that samples that reacted slower and less violently still contained only MDMA, just at a lower purity
I think this is why it's recommended to use marquis, Mecke, Mandelin and Simons in conjunction to give a more accurate sign of what the sample may contain