For strength analysis, the effective laminate is treated as a pseudo three stack laminate. See:
http://hypersizer.com/help/#Materials/EffLam/el-using_struct_analysis.php%3FTocPath%3DFeatures%7CMaterials%7CEffective%2520Laminates%7CUsing%2520Effective%2520Laminates%7C_____1The middle stack is a smeared orthotropic ply that makes up the majority of the laminate thickness. In your case, the effective material stiffness from the 0%/50%/50% material is used with the thickness listed on the sizing form (0.025in). This thickness and stiffness is used to calculate the effective laminate ABD stiffness.
Failure is evaluated at the IML/OML of the effective laminate. The laminate strains are localized into ply-based strains in each ply angle present in the effective laminate. Since you only have 45/90 degree plies in your effective laminate, then HyperSizer will only evaluate ply-based failure for 45/90 plies using the strains at the IML/OML. Since you are using effective (smeared) laminates, HyperSizer treats the IML/OML strains as ply strains and checks the lamina failure criteria for each ply direction because it doesn't know the preferred stacking sequence. This is a conservative approach for strength analysis of effective laminates.
So why is HyperSizer displaying 4 plies (2 plies 45, 2 plies 90)? Since you have a 0%/50%/50% effective laminate and a 3 ply laminate, HyperSizer is showing an additional ply in the sizing form to achieve the desired ply percentage. **The ply counts in the sizing form do not affect the analysis of the effective laminate. This is just a reporting issue for thin laminates. The true laminate effective thickness displayed on the sizing form is used for strength analysis of the effective laminate.
There does seem to be an inconsistency with the effective analysis for thin laminates where HyperSizer is using a 0%/50%/50% effective laminate but choosing 3 plies thickness. We need to figure out the best way to handle this in future versions. Perhaps it should not be allowed to choose a 3 ply thickness and select the 0%/50%/50% effective laminate since it is not possible to meet this ply percentage at that thickness. Keep in mind that the purpose of the effective laminate sizing is to help HyperSizer obtain ply count targets to use to create discrete laminates. The inconsistency will be resolved when the discrete laminates are automatically created and the part is sized with the actual laminate stacking sequences.