Did you take account of the polarisation being equivalent at 180/360 (or 090/270)? That would bring the 'cell width' down to 71m at 40km.

There will be differences in maxima with polarisation between different pulses ~ simple interpolation/analogue summation/differences may be sufficient to offer improvements to effective 'position' resolution to the "part cell" level. (eg if cell 3 and 4 have equal strengths, then the target location is at 3.5: if 4 is a maxima and 5 is less than 3 ~ then an approximation to the location would be 3.75 etc) Digitial systems can often be more 'precise', but they may lose the subtleties that an analogue system can offer ~ similar arguments can be heard from audiophiles comparing the "better" CD quality to the subtle variations that can be captured in vinyl (or indeed tape)...

That said... I do sometimes feel that SAMSIM tends to be optimistic about tracking performance and missile accuracy performance** ~ and added to that the target avoidance is minimal/non-existant. The lethality also seems optimistic ~ there *is* a reason that multiple rounds were expended per target from almost all systems, while that seldom feels "required" within the program where a valid launch parameter can be observed before the Weasels take their bite - a single missile is usually sufficient for a successful engagement.

** this might be confirmed/refuted once we can observe the target/missile flight from the operator's POV using Karat (especially if the GPX could be viewed from within the 'cabin' using Karat rather than the radar screens if they were used during the engagement).