@YCor I wasn't really sure what statistics exactly you're looking for, anyway I tried to do two SEDE queries similar to the description in your comment.

Posts with score at least 50 per month, and the same query for non-CW posts. (Of course, you can change the parameter from 50 to a different number.)

If you are interested in various changes on MO over time, some SEDE queries were posted here: Does the number of answers really drop over years? (Although the focus there was mainly in changes related to answers.)

@MartinSleziak Oh, this is exactly this. It confirms that there were huge scores in the first months, and decreased until somewhat normalization after, say 18 months.

I'd be curious about the reasons of this effect. I suspect that there were less restrictions for voting, but I don't know.

Well, there were certainly more soft-questions, big-list questions in the beginning. (The ones for which the CW is usually misused on MO.)

@YCor I don't think so. It was simply less questions, so a question got more views. Also many interested questions were posted in the beginning, partly in the attempt to populate the site and get it started. (Among the very first questions, I have seen quite a few where the OP knew the answer and posted it mainly to make the question available on the site.)

Moreover, some difference in score over time is to be expected. It is difference whether a question got upvotes from users who saw it during 6 years or during 6 months.

Old questions are still gaining some upvotes. Of course, the majority of upvotes probably comes during the first days.

Here are the queries showing how number of votes is distributed over time (since the post was made) for all posts and only for questions.

I understand your arguments about the flow of questions; I guess it could be confirmed by statistics about the flow of questions, namely if increased in the first 18 months until roughly stabilizing. I'm less convinced about the "old question gaining upvotes" part; I guess it's a negligible contribution since otherwise it would make posts from, say 2012-2014, much higher in the curve.

Not completely negligible, but definitely most votes come soon after the question/answer was posted - see the graph I posted above.

If by "flow of questions" you mean how many number of questions were posted, we can simply look at the graphs I posted here: Does the number of answers really drop over years?.

> Here is the number of questions/answers/all posts per month (and a corresponding query which includes deleted posts).

This is the number of posts (including deleted posts):

Even without digging into various statistics, you can confirm from your own experience that occasionally you get an upvote on some of your older posts.

In fact, there are users who no longer need to make new post and still reach daily reputation cap: The Stack Exchange pension, or, reaching the daily reputation limit entirely with votes on old posts.

BTW traffic on the site is also a factor. For example, if a site has 100 questions per day, many users will see your question. (In fact, at that rate some users will at least check the titles of all questions on the site.)

If a site has 1000 questions per day, smaller percentage of users will notice your question. (And at this rated, it's very unlikely that somebody looks at all questions.)

I have posted something about how votes are distributed soon after a post was made and then later - it is probably easier to understand if we look at percentages.

Here is such query for all posts.

And the same for questions.

Average score of questions per month

Average score of answers per month