If you put proper error bars you'll notice no trend. The number of daily reviews is far too low for the graph to mean anything. Remember that if you see 25 reviews in a day, the associated standard deviation is sqrt(25)=5, or ±20%.
The graph isn't measuring day periods. It's measuring 2 week periods. So the standard deviation of a day doesn't mean anything, because that isn't what it is. It's a rolling 2-week average.
Please don't subject us to your pointless graph that proves nothing.
We can all see that you are basing your so called 2 weeks of positive reviews on a week that starts on a Thursday.
As others have pointed out, you can get totally different numbers depending on what day you start from.
The release date according to Steam was 10th Feb. That was a Monday.
So, its more accurate to count weekly numbers starting on a Monday.
Which is what I did. I proved that for the last 3 full weeks starting on a Monday, only 1 week had more positive reviews. But it was only positive by 6. The next week was negative by 6. So that positive week got cancelled out by the very next week. When you count from a Monday.
But, hey, lets move the weekly start date on our counting, until we get 2 positive weeks eh?
You getting caught up on what day of the week the graph starts has me gobsmacked. The full release was on February 11th. The early access launch was on the 6th. I don't know why it would say it launched on the 10th
What about the issue with the fact that you can get a totally different graph, by starting the counted week on a different day?
No you do not. You do not understand what the graph is showing. Perhaps that's my fault for not explaining it clearly enough. It's a 2-week rolling average. Every single day starting from the 19th of February up until when I made it which is the 22nd (I think) is plotted showing the 2-week rolling average. Starting on a different day - which is already plotted - would not change the graph. Unless you mean if you start on a random day in April? Then yes the graph would look different because you're not including all the data we have. That graph includes all the data we had up until the 22nd.
It's simple - the data is junk in the first place. I mean you could make qualitative conclusion that Civ7 has not the best reception so far. But any attempts to get quantitative results, like trends, fail because we have so limited data.
BTW, a piece of information to mull over - we usually look at reviews for the base game, but there are separate reviews on, for example, DLC content. They are really few and thus have even less value, but since this thread is about speculation it could be interesting to speculate about them and their difference with the base game reviews.
The data surrounding the Steam reviews isn't limited though? It's plain to see for everyone?
Game was released on 10th Feb. Which was a Monday. That is why it makes sense to count weeks starting on a Monday. If you want to coubt each week. Personally I prefer to just add up the figures each month.
The official release was on the 11th February. The early access started on the 6th. As I've already stated, 64% of the reviews came before the official release on the 11th. Why would you be happy to omit the majority of reviews?
Does it really matter?
This is one reason why it's not really feasible to count reviews on a weekly basis. Far better to count them at the end of each month. Plus keeping an eye on the total review count and the average number of players each month.
It was on a 2-week basis. Why is a 4-week basis fine but a 2-week basis unacceptable to you? A month is basically just 4 weeks. Here is a 30-day rolling average as an attempt to satisfy you. Once again, it shows recent reviews have been the highest since launch. Every graph will show this as that's what the data shows...
I'm kinda checked out of this discussion because it's so stupid, but you can ask yourselves how granular you want to go and whether it makes sense. It's like looking through a microscope at an asphalt road - it's going to be bumpy and rocky. You can find places where it appears to go up or down, but if you zoom out, it's a flat road.
In the case of Civ 7, it's a very slightly downhill road that's flattening out. There's no proof of an incline coming. The only reasonable way to measure with this low amount of reviews is over months.
How far do we have to zoom out to see this "flat road". The above graph is now a 30-day rolling average. It's still clear to see that recent reviews have ticked upwards - the best since launch.
You literally posted a graph where it appears that there are just a few days in the entire time where there were more positive reviews than negative. Can we at least agree that if there are more negative reviews than positive, that the turnaround isn’t underway?
We’re over 6 months post-release and virtually every metric for this game reveals underwhelming performance, at best. You’re going to need to show more than the reviews being slightly less negative for a few weeks.
Have you been reading a single thing I've said? The reviews are the best they've been since launch. It's a fact. It's too early to say if a turnaround is underway, I've never claimed a turnaround is happening, but if one is in progress then it has already started and it's clear to see in any graph that is made because that is what the data shows. All I've stated are facts based on data. The best 2 week period since launch, as well as the best 30 day period since launch. The August peak is close to this peak, but this October peak slightly edges it and had better build-up.
