Based on the stats, there doesn't appear to be any causal relationship between promotion and payouts.
The last 2 charts do not represent the entire data set, yet they show no clear correlation between the two, and thus your observation appears to be correct.
Whether promotions do have a positive effect on the payouts, I don't know. One thing that is for certain, is that generally speaking, the market participants are not losing money on promotions as a whole - in fact, again, generally speaking, the promoted posts generate several folds more revenue over the promotion investment.
As you have pointed out, it would be interesting to know how much of the revenue is actually caused by the promotion. Unfortunately, I don't know of any good way of figuring that out at this moment.
Rather than looking at the total payouts for promoted posts, we should be looking at the total number of posts that used the promotion feature, how many of those posts covered the promotion costs after payout, and whether or not the payout was higher than their average payout from prior posts.
That's a lot of data to collect, but that's really the only way to get a more accurate picture of the utility of the feature. Looking at total earnings without acknowledging previous success of the users doesn't give an accurate measure of the potential benefit of the feature.
Like I mentioned in my first comment - if someone averages $500 per post, then promotes a post for $100 and earns another $500 payout, we can say that they likely lost $100 on the promotion. It didn't actually improve earnings.
Anyway - I'm glad to see your stats and it'll be interesting to continue watching how this evolves. Maybe a program can be created to calculate what I mentioned above?
I think the main benefit for promotion is to get more followers. The actual payout is based on the luck of enough whales upvoting you. It depends what you value. If you want pure profit then it is not worth it. If you want people to read your posts then it may be worth it.
But the question then is, are more people actually reading these posts and following the authors? More data to collect! (But the first data set really isn't possible to know, currently.)