The problem is is that your original statement was not put into context and was erroneous on at least 2 points.
First of all, context makes all the difference. On a car, I'm not going to be messing with the electrics myself, on a boat, it is likely I'll have to. My regular A/C power I can turn off before I touch it, banks of batteries in series present a different problem. Managable I'm sure but the risks are different. When making statements like which is better, context MUST be considered. To simple choose one aspect (ie efficiency) and make an argument based on just that doesn't cut it.
Secondly, your two errors.
1) Lower voltage results in larger wires OR lower efficiency. If I halve my voltage and double the width of my wires I get no loss in efficiency (I've been reading)
2) In your second example you stated the absolute number (when refering to resistance) does not matter. Well, in the interest in putting this statement into context, I did some reading and came up with this. If your statement is true then putting different numbers into the equation shouldn't make a qualatative difference to the result.
Googling I found this site : American Wire Gauge table and AWG Electrical Current Load Limits
Based on my recollection, the wires I saw on the Lagoon were somewhere in the region of 6mm, maybe fatter And I'm going to guess they were no more than a meter long.
Using this table, I get a resistance of approx .5 milli ohm. No idea what the controller resistance is but let's double this to make this 1 milli ohm.
Putting this into your original equations (you may need to correct me as this is the first time I've done this), I get an efficiency of 99.97% for 240V and 99.77% for 72V. Now, if I use 10kW instead of your 20kW example (which is more realistic for the Lagoon), I now get an efficiency of 99.98% for 240V and 99.81% for 72V.
So let's at least get this straight. The resistance does not matter if all you want to do is compare efficiency only for two identical systems. 240V will always be better than 72V and 1000V will always be better than 240V.
But in context, using real numbers (subject to my doubling assumtion), the actual efficiency trade
of is 99.98% -> 99.81% (not 96.53% -> 61.42% as your example impied) for a system where I'm less likely to harm myself during maintenance
. There are no doubt other factors but when comparing systems, I prefer to deal with all the aspects (not just efficiency) and to use numbers which are closer to reality.
Seriously, I appreciate this discussion as it's making me research
and learn things that I am benefitting from.