I’d like to start by saying that your idea for measuring dominance is very interesting and I appreciate the work you put into it. It did make me think. However, I also think that the question of points systems is _very_ legitimate.
no revisionist history
I understand what you mean, replacing Senna with Prost would be a bit revisionist, but in my opinion, keeping the “official” results has even worse problems.
Here’s an example: Imagine two drivers, A getting all first places and B getting all second places. Obviously these two are going to be #1 and #2 in the standings. If this happens in 2002, the winner appears in your chart with domination percentage 40%. A year later the very same situation yields 20%.
Now granted, this was an artificial example, so a good question is what difference we can expect in real results. I do not have machine-workable data for drivers, but I can relatively easily check on teams. I tried two cases, what follows is year, gap using official points, and gap using the 9-6-4-3-2-1 system:
2003 9% 14%
2004 55% 66%
So it seems that the difference caused by changing scoring system can easily be about 5%, even exceed 10%. This begs a question whether the comparison using mixed scoring systems is even meaningful, given that a 5% change influences standings in your chart by quite a bit. I think converting to one points system would make the results much more relevant, despite other problems that this would bring.
Incidentally, note that the top 5 most dominant seasons in your chart happen to be seasons where the winner was getting almost the biggest share. On the other hand, 2011 rewarded the winner much less. I took the time to recalculate that season (by hand, I might have made a mistake) and came up with the figure 41% instead of 31%, tying Vettel with Fangio.