The way I understand it you are talking about two different points. If not let me know, but I will answer them separately.
Game time: I did not include game time, because Damage% and Gold% are always 100% each for the team. So we have a measure that tells us how much damage (for example) they do compared to their teammates. And since very teammate obv. has the same game length, game time doesnt matter. Does this method have weaknesses? Definitely. I could have used damage per minute. But as I mention in my article, this has other problems, which make it difficult to compare champions. A fight heavy game for Kalista could mean a higher dpm than a 0 kills win for Ezreal, but that doesn't mean Kalista is in general better at dealing damage. When we use the Damage%/Gold% we at least take that factor out, because a fight heavy game means Kalista's teammates also had the chance to deal a lot of damage. And since she is compared to them, it is mostly taken out of the equation.
One dot vs multiple: Having more data points was important for me to see a more holistic picture. It also allows a more detailed understanding of the damage trade off. I wanted to see if there were any observations where, for example, a top laner dealt a lot of his team's damage while still taking a lot of damage in teamfights.
Hope this was helpful, if not, let me know :)
EDIT: I am not saying my approach is perfect or the right one, but I think that is okay as long as I tell you guys what it is that I am doing so you can say "maaan this makes no sense" if you disagree :).
