Just finished modeling a 6-element DK7ZB design in EZNEC and comparing the predicted bandwidth to actual measurements. The modeled SWR curve showed flat response from 144-148 MHz, but real-world testing revealed some interesting differences.

Using a NanoVNA to sweep the entire 2m band, I found the antenna's resonant point had shifted about 500 kHz higher than predicted. SWR changed from around 1.15 down to 1.08 and showed almost constant response across the entire band. Adding a single ferrite choke made a noticeable improvement.

For anyone building yagis, don't just check SWR at one frequency. It's far better to look at the SWR at multiple frequencies across the amateur band and check it closely across the band. The bandwidth characteristics can surprise you.

Great point about measuring across the band. I've seen too many hams tune for 146.52 and wonder why their repeater contacts on 147.9 are poor. It's a good idea to choose an aerial that has wide bandwidth of low SWR, even at the expense of small decrease in gain.

What ferrite material did you use for the choke? I was advised to use Grade 31 ferrite for VHF and found that one was sufficient during testing. The SWR improvement was modest but noticeable on my 5-element beam.

The software is pretty good, but it's unable to help you in turning theory into reality. You might fall foul of cutting errors, or find that at UHF getting a driven element spot on is very difficult. I always cut elements slightly long and trim to resonance with an analyzer.