I saw the vid in my daily YT feed, and indeed, it runs brilliantly.

Perhaps that 3% you had to trim it, has to do with fuel viscosity (which is not accounted for in the standard lines for mixture correction based on temperature and pressure) and maybe, if you have recorded data from the summertime you can work out a small tweak in the temperature correction factor to eliminate this?

I can imagine that by taking a different baseline than 0K for temperature would change the rate of correction for temperature. which could compensate a bit better for the increase of viscosity. Just as an example.
Right now we are correcting for temperature by converting Fahrenheit or Celsius to Kelvin, and all else remaining the same, every degree above freezing will reduce fuel by approx 1/273rd
Thi does not take into account that with rising temperature fuel viscosity drops a little, so actual fuel does not reduce by 1/273rd per degree.

So how if we would tweak that formula by pretending 0K is minus, say, 220 (just picking a number here, as example). That would increase the correction rate per degree of temperature to 1/220th.

How to work that into the code, I don't know because I'm not good at writing code. It would (if sufficient data is availlable to work out the deviation) eliminate the need for trimming for winter or summer.

Maybe it is a bit over the top, but maybe some of you guys like to tweak and fiddle.

But yes, it does perform absolutely sweet. Jealouse that I'm at sea and not flying, then again the weather over here is pretty awful at the moment. One storm after another.
Gonna be happy when this voyage is over in about a week, my seafaring days are over, done with the rolling and pitching, done with the stress and the ridiculous paperwork and BS procedures and the pressure from headoffice.