Working on optimizing a 3-element 20m Yagi design and getting different results between EZNEC and MMANA-GAL. Use tools like MMANA-GAL or EZNEC... MMANA-GAL, EZNEC, 4NEC2: Modeling software. EZNEC shows 8.2 dBi gain with 18 dB F/B while MMANA gives 7.8 dBi and 22 dB F/B for identical geometry. Element spacing: DE at 0", reflector at -60", director at +72". Both use NEC-2 engine with medium ground. Advanced, cutting-edge computer modeling and optimization techniques were used to produce every one of the EAntenna premier HF beam antennas. Which results should I trust for real-world construction?

The difference you're seeing is likely in the ground modeling implementation and segment density. EZNEC tends to be more conservative in its gain predictions. I'd trust the EZNEC numbers for construction - better to be pleasantly surprised than disappointed.

The middle curve shows an optimized Yagi. The lower curve represents a poor performance design. Are you using the same wire diameter and height above ground in both programs? Small differences in these parameters can cause the variations you're seeing.

I've built several beams and found EZNEC predictions track closer to real-world measurements, especially for F/B ratio. The gain difference between programs often disappears once you factor in real-world losses. Focus on the pattern shape rather than absolute numbers.