Hacker News Reader

It was discussed on HN 4 days ago although it seems to have been flagged off the front page or something after a very short time: https://news.ycombinator.com/item?id=29813079

That happens to be my own submission, FWIW.

DDG is accusing Google of manipulating browser-extension standards for Google's own benefit, FWIW.

(Discussing specific examples in contenxt is often useful.)

That particular issue seems to have more bearing on DDG's attempts at creating its own browser than of Google specifically thwarting DDG's search.

> That happens to be my own submission, FWIW.

Ok, I felt something was odd but didn't connect the dots as I was to quick :-D

Anyways, thanks!

I try to avoid Chrome.. what roadblocks?

See my reply to dredmorbius (sibling comment): https://news.ycombinator.com/edit?id=29861689

(I think there is some automation to prevent duplicate content so I link to it instead.)

NB: for virtually all real-world human-behaviour data, smoothing intervals in multiples of 7 days are the best choice as they smooth out weekly usage patterns.

Try 7, 14, 21, and 28 day smoothing as alternatives to 10.

Note that 7-day averaging results in far less variance than 10-day, despite being a smaller interval.

Could be a plateau, or it could be a temporary peak somewhere around January last year.

28 or 90 day average shows essentially the same thing but with less noise. Showing a 365 day average is actually good for the casual observer to remove holiday/seasonal fluctuations and show YoY growth.

That's how you lie with statistics.

I was looking for a comment like this. To me this is the real news.

This is the first time I’ve seen DuckDuckGo Traffic not look like a steady exponential. With the 365-day rolling average you can see growth leveling off noticeably by the end of 2021. Without the rolling average you can see an unusual spike in Jan 2021 followed by a V for the remainder of the year.

They have lost me for the moment.

Reply on news.ycombinator.com