The 15m beam DE is based on a Mosley trapped element originally intended for 10/15/20m operation.
Under this new configuration, the 10m capability is unused. The 15m operation serves as the 15m DE. And the 20m operation is re-tasked for 17m operation.
As 17m operation has no So2r complications, it's use on a stand-alone basis simplifies design.
Performance vs. Dipole
The configuration performs a bit better than a dipole, in that it's pattern is slightly more focused. Below we see a comparison of a dipole element of the same dimensions as the model, however standing alone.
And here, the same element, within the complete model. Notice the lack of the center "peak" at 90 degrees (vertical) - and a more elongated azimuth pattern. In reality, there is also a bit of trap loss not included in these numbers.
Dipole vs. Yagi
A further refinement of the design would be to add a set of directory wires into the 10/15 element set.
The prior 17m antenna implementation also used the trapped DE, but the yagi end elements were configured as either a director or reflector through a series of relays and 3 inductors (17m dir, 17m ref, 15m ref).
While this was workable, the implementation was poorly done - and if I were to move back to a Yagi, it would be of the two element variety for the reasons explained in the 10m and 15m pages.
An example of the added director would be positioned as shown in the wire frame drawing below.
The quick simulation with the element positioned as shown above - as proof of concept - offers a 3 db increase in forward gain, and more importantly, a quite nice F/B - at least on the east-beaming direction the following EZNEC+ plots are drawn from. The west beaming implementation would not be expected to offer as good of a F/B in that the is a lot more antenna for the signal to "swim" through on the west beaming side - including the blessed stucco wall...
Look for this option to be explored further after the So2r and HF antenna work settles down.