Interesting info on B3702 Injection timing
The boundary coincides with IVC event, it's the last point at which fuel can enter the cylinder for the current cycle, and the earliest point for SOI on the next cycle. EFILive doesn't provide the calibration to adjust the boundary, it doesn't really matter as long as you know where it is, your IVC is 580ATDC, 5 degrees earlier than the PCM boundary value. If you adjusted to your cam, it would give you 5 more degrees before EOI would start retarding finishing later after IVO.
The PCM calculates the IPW for airflow from MAF or VE table, it calculates the number of crank degrees required for that IPW, and calculates SOI back from EOI, B1205 and B1206.
The complete SOI to EOI happens before IVO, as IPW and RPM increase, SOI eventually meets the boundary where IVC from the last cycle. EOI has to start extending past IVO still injecting at the beginning of induction into the air stream, at these high loads I imagine cylinder temps are high enough to maintain vaporisation but may not be mixed with the air as efficiently.
Another way to look at it, there are 720 from IVC to next IVC, at above 85% duty cycle, SOI to EOI takes over 612 crank degrees, there is a minimum off time for the injector in between. It's flat on all the time. The purpose to maintain less than 80~85% Duty cycle may well be to ensure EOI before IVO..?
I see where you going injecting after the EVC, IVO, but If you look at your inlet duration 227 minus the overlap 11 degrees, minus the minimum injector off time, you could be down to say 190, it could be OK at idle and low rpm, but how efficient is mixing and vaporisation of homogenous charge?
Injection happens on a closed inlet valve long enough before it opens to allow vaporisation and cooling of the IV. During overlap some vaporised fuel may be lost to the EV, but not raw fuel, if it still smells raw fuel I would experiment and try increase B1205 values to allow more time for vaporisation.