I agree re more sensors. I may ask for more explicit guidance on updating that firmware (perhaps in an exchange of DMs), because my efforts a while back seemed to fail at the last hurdle. But, I’m a bit preoccupied with other things for a couple of days, so not quite straight away, please 
(and only after I’ve come up with that diagram!)
I wouldn’t call the Viessmann behaviour “hunting” exactly. It’s just being stopped from getting very far towards its own goal (flow temp 60degC) by the evohome, which makes its own calculation of how much heat is needed and how long the boiler needs to burn for in order to deliver that heat. That’s how the boiler gets shut down long before the set flow temp is reached. The evohome “learns” what the boiler output is and adapts its TPI parameters accordingly. Or so I assume.
Of course, the boiler modulates its output, which could confuse evohome but, if the demand is only for a minute or two at a time, tbh, the output is going to be whatever is typical of the “stabilisation” period. It sounds (literally - the boiler makes a sound that varies with power output) like it’s more than 25%, perhaps between 30% and 40% (based on gas meter readings which are available at 10 s intervals, if I can be bothered to sit there pressing buttons to wake up the meter’s display…)
There’s definitely a bit of a thump every time the boiler fires up, so I’m sure you’re right that the gas isn’t being burned quite as efficiently right at the start. But if much methane actually escapes, then that’s not great for CO2 equivalent emissions, so I’m assuming that someone at Viessmann has looked at that fairly carefully …