I have had my ecometer installed for something over 6000 miles now and I've picked an efficiency setting of 83%. Each ecometer efficiency setting has been measured against actual average mileage for about 1100-1300miles. Cumulative fuel used vs. total miles traveled provided the data for actual mpg; the ecometer was not zeroed out for the duration of the test segment. Individual fill-up variations/calculation vs. incremental ecometer readings would typically show fairly significant differences in results. The purpose of the extended longer test period was to rule out the effect of those shorter fill-up calculations.
The ecometer can only be adjusted in single percentage steps, e.g. 81%, 82%, 83%. My overall calculations revealed that the optimal setting for my particular vehicle would have been 82.5% but that was not a number that could be input, so I settled on 83%. With a setting of 82% my ecometer read about 0.3-0.5 mpg below actual fuel usage calculations; a setting of 83% yields an ecometer reading about the same amount of 0.3-0.5 mpg above actual fuel usage. Picking 82% vs. 83% was based primarily on impressing random smart looky-loos with a bit better mileage showing on the ecometer

. Keep in mind that changes in rims/tires which result in different rolling diameters could be expected to alter the efficiency number required for some other vehicle. I am running 16” tires on my smart, but the rolling diameter is within 1% of the stock tire and wheel combination. In that the ecometer does all of it’s calculations from information coming through the OBD port, it could not be expected to know about changes to wheel/tire combinations. Your efficiency input might vary from mine.
You should not be expecting the ecometer to show you absolutely precise mpg numbers any more than you could claim 100% accuracy from individual fill-up calculations. I don’t find any problem with readings that are plus or minus 0.5 mpg.