Anyway, why is quality based on a straight average of all inputs? That rings quite false to me. For example, in a wine grape quality is crucial. Also important is the process, and I suppose how well it's sweetened matters somewhat. But the glass? A wine made from quality 0 grapes and 25 glass is just as good as one made with 0 glass and 25 grapes? That's not how I buy my wine!
Wine ingredients are:
20 kg Grapes
0.3 kg Sugar
0.5 Liter Water
0.2 kg Glass
And of course the wine tech itself.
So if I have q5 grapes, q2 Sugar, q0 Water, q0 glass, and q3 Wine tech that's a 2 product on the straight average.
But I get the same from q0 grapes, q0 sugar, q0 water, q10 glass, and q0 Wine. This makes no sense to me. That'd be terrible wine in a pretty bottle. I drank that stuff in college, and it does not pass for quality wine I assure you!
So why not just assign some value to the tech and then do a weighted average? For example, tech weight = 10
20 * 5
.3 * 2
.5 * 0
.2 * 0
10 * 3
-------
130.6 / 31 = q4 product from good grapes and tech. Meanwhile ...
20 * 0
.3 * 0
.5 * 0
.2 * 10
10 * 0
-------
2 / 31 = q0 wine. Which makes more sense.
And of course there are other products where glass is more important.
In the second iteration, we could assign values to each component separate from their input qualities. For example, E-components are more important to a Television than plastic despite 1 of them to 5 plastic kg. But in the meantime, inputs are an easy way to weight. Just to open the door.
So I ask, why not weighted averages for quality?
