Player stats, sales, and reception speculation thread

[Kwami]

We do have sales estimates. The difference between Civ6 and Civ7 is so stark, these estimates are good enough to get a pretty good idea.

Those numbers look wonky to me. If they're accurate, then Civ VI sold 23.4M copies, but there are 34M owners. What?

 

[Leucarum]

The one unambiguously bad metric is the low amount of twitch engagement. Streaming media has been a boon to Civ6. It does look as if Civ7 has a long way to go on that front.

 

[imperfectpeter]

Not going to make any grand claims about what these stats might mean for the future of video game development or whatever, but I do find it interesting that even good old Civ V is getting twice the viewers of Civ VII (well at least at the times I'm checking). Civ VI absolutely dwarfs VII, not even close. Viewers can obviously fluctuate a fair amount, however it's worth noting the amount of followers too.

 

[stealth_nsk]

The one unambiguously bad metric is the low amount of twitch engagement. Streaming media has been a boon to Civ6. It does look as if Civ7 has a long way to go on that front.

Well, steam reviews are also unambiguously bad. The problem with both cases is that they are abstract "bads" in vacuum. There's no question we're seeking answer for, no conclusions we could come from it. It's just a "good-bad" scale floating in nowhere.

 

[itix]

Anyone could look through any glasses. For example through professional ones.

Here's a description of methods used by Gamalytics: https://gamalytic.com/blog/how-to-accurately-estimate-steam-sales
While the aggregated precision is great, it requires all of the data points to be correct. For example, using concurrent player count requires estimating playtime, which is cool for research, but impossible to automate. And, for example, top seller rank doesn't allow estimating sales during period when game doesn't reach top rank, also it requires knowing sales structure in terms of base game/DLC, which is info not available at all.

Which means that there is a quite good chance that Gamalytic overestimates Civ7 sales...

 

[IntelligentDisk]

Civ VI also had less Twitch stream viewers than Civ V did for over 2 years, it mirrors Civ VI taking over 2 years to overtake Civ V on Steam, and currently the Twitch viewership is mirroring the Civ V/VI/VII player counts. It'll take years before VII can overcome.

 

[itix]

Isn't your method flawed? You're assuming Civ VII will not increase its annual rate, when it's clear that Civilization games sell more as years go by with bigger discounts, more updates & DLC. I believe the Steam reviews per year for VI shows this:

• 2016 - 22,765
• 2017 - 14,399
• 2018 - 16,103
• 2019 - 29,350
• 2020 - 59,383
• 2021 - 55,260
• 2022 - 50,807
• 2023 - 47,143
• 2024 - 44,015
• 2025 so far - 24,668

So that's 85% of all Civ VI reviews coming 2 years after its release. 77% coming 3 years after its release. Not to mention the fact that Civ VII is less affordable than VI was at the same point (higher initial price plus less sales & smaller discount), plus it launched on other platforms, VI did not. If anyone wanted to play VI within the first 3 years, they had to buy it on Steam and play it on a computer. Both of these factors will slow down VIIs sales and "annual rate".

Err, please dont use reviews to estimate sales. It is known to be unreliable.

In 2019, Steam began to prompt users to leave reviews, which substantially increased review numbers.

 

[IntelligentDisk]

Something I'm confused about is past estimates from Gamlytic, I tried using the wayback machine to fine find past estimates to gauge how many sales they think Civ VI had each year. The only page saved is from October 6th 2024, which states Copies sold: 15.8m (7.9m - 23.7m), Owners 19.8m.

Yet today, October 18th 2025, it states Copies sold: 23.4m (15.9m - 31m), Owners 34m. Do think they think Civ VI sold almost 8,000,000 copies and gained over 14,000,000 owners in the past 12 months? Is this an error? Did they revise their figures?

 

[Siptah]

Those numbers look wonky to me. If they're accurate, then Civ VI sold 23.4M copies, but there are 34M owners. What?

Sold copies are on Steam. Owners include people that bought Steam keys on other markets. I think the distinction is made because the page also tries to estimate the revenue, and a steam copy is usually more expensive than a steam key from another source.

 

[Smegger213]

Reviews for month 9 were the second best reviewed month after the first. A 3%-6% improvement from the previous 3 months plus the 4th best reviewed week and the best reviewed day yet. It shows they're doing something right.

The player count peaked at it's highest in 150 days just a few weeks ago, as well as having it's highest weekly average peak concurrent players in 21 weeks.

You said something about September, you're right September was bad, but it's October now. So far the average for October is 7,426, which is higher than the previous 4 months.



47% positive/53% negative does not show that the vast majority dislike it, record pre-orders does not show the vast majority does not want anything to do with the game.

Nope, the current 30 day average is 6515.
Plus the 30 day average since the end of Sept has always been less than 7000.
I don't know where you get your 7426 figure from, but it is plainly wrong.
Regarding Steam reviews. There are STILL around 2000 more negative reviews than positive ones.
Plus there are far more days recently that have more negative than positive reviews.

 

[IntelligentDisk]

Nope, the current 30 day average is 6515.
Plus the 30 day average since the end of Sept has always been less than 7000.
I don't know where you get your 7426 figure from, but it is plainly wrong.
Regarding Steam reviews. There are STILL around 2000 more negative reviews than positive ones.
Plus there are far more days recently that have more negative than positive reviews.

I didn't say anything about the 30 day average. I said for October. I got the figure from SteamDB. Export the data and take a look yourself if you find it so hard to believe.

Yes there are still more negative reviews than positive overall, that doesn't change the fact that month 9 had the second best reviews since launch, and a 3-6% improvement upon the previous 3 months.

 

[Smegger213]

I didn't say anything about the 30 day average. I said for October. I got the figure from SteamDB. Export the data and take a look yourself if you find it so hard to believe.

Yes there are still more negative reviews than positive overall, that doesn't change the fact that month 9 had the second best reviews since launch, and a 3-6% improvement upon the previous 3 months.

Look, there is no average for October yet because October is not over yet.
SteamDB does a rolling 30 day average. That average has stayed at less than 7000 since the end of September.
Why is it that people like you just will not accept that the game is failing?
You do anything to put a positive spin on it.
The figure that counts is the total number of reviews and the difference between positive and negative reviews.
That difference has stayed at around 2000 for months.
The other thing that counts is the number of players. That number is dropping back down, after a slight peak after the last patch.
Sunday 6th Oct saw a peak of 12,594, 12th Oct 11205, and last night the peak was 9634. So, the player count is dropping back down again.
That is fact, if SteamDB is to be believed.

 

[IntelligentDisk]

Look, there is no average for October yet because October is not over yet.
SteamDB does a rolling 30 day average. That average has stayed at less than 7000 since the end of September.
Why is it that people like you just will not accept that the game is failing?
You do anything to put a positive spin on it.
The figure that counts is the total number of reviews and the difference between positive and negative reviews.
That difference has stayed at around 2000 for months.
The other thing that counts is the number of players. That number is dropping back down, after a slight peak after the last patch.
Sunday 6th Oct saw a peak of 12,594, 12th Oct 11205, and last night the peak was 9634. So, the player count is dropping back down again.
That is fact, if SteamDB is to be believed.

Just because October isn't over doesn't mean there isn't an average for October.

Why is it that people like you want VII to fail so bad? In September I was pointing out the average peak concurrent players were the lowest they've ever been, now I'm pointing out we had the best week in 21 weeks as well as the best day in 150+. It's your problem that you think anyone who points out positive things "won't accept that the game is failing". Maybe you need to ask yourself why you are unable to acknowledge anything positive?

 

[Kev1916]

Just because October isn't over doesn't mean there isn't an average for October.

The problem if you try to be too clever, or just get it wrong is figures can be checked . there is no "3-6% improvement "

Steam shows last 30 days of the 795 reviews just 45% is positive
For October we have 529 of which just 44% is positive .
In a week 148 reviews 41% positive
Today 5 reviews 40% positive

Total reviews are still 49% Negative


There is no improvement and as before the player count is on the way down

 

[IntelligentDisk]

The problem if you try to be too clever, or just get it wrong is figures can be checked . there is no "3-6% improvement "

Steam shows last 30 days of the 795 reviews just 45% is positive
For October we have 529 of which just 44% is positive .
In a week 148 reviews 41% positive
Today 5 reviews 40% positive

Total reviews are still 49% Negative


There is no improvement and as before the player count is on the way down

As I said, check SteamDB. Export the data into a spreadsheet and take a look for yourself.

I count a month as 28 days/4 weeks, so we can properly compare to Civ VI. Month 9 (18th September to 15th October) had a rating of 43.1% positive (405 positive, 535 negative). That is a 3-6% increase on the previous 3 individual months. A 4.7% gain over that 3 month period as a whole.

If you don't like that method, and would prefer to use the actual months (August v September v October). Then October currently is 45.5% positive (325 positive, 390 negative). Again, that's a big increase from the previous months, and the second best calendar month so far.

• February: 50.9% positive (32,635 reviews)
• March: 43.2% positive (5,583 reviews)
• April: 41.6% positive (1,912 reviews)
• May: 33.5% positive (1,770 reviews)
• June: 27.5% positive (1,928 reviews)
• July 40.6% positive (1,293 reviews)
• August: 36.9% positive (1,386 reviews)
• September: 38.4% positive (870 reviews)
• October*: 45.5% positive (715 reviews)

*as of 19th October

 

[Lazy sweeper]

I Just spent 120 Eurs for a new game... which is NOT Civ 7....
I would have loved to use this money to buy Civ 7 but no.
Too little, too far.
I loved the new UI projections, and map seed generation.
But the game is too short.
Too childish.
Maybe it's good for some but not for me.
Major changes has to be made.

 

[Smegger213]

Just because October isn't over doesn't mean there isn't an average for October.

Why is it that people like you want VII to fail so bad? In September I was pointing out the average peak concurrent players were the lowest they've ever been, now I'm pointing out we had the best week in 21 weeks as well as the best day in 150+. It's your problem that you think anyone who points out positive things "won't accept that the game is failing". Maybe you need to ask yourself why you are unable to acknowledge anything positive?

Look pal. You will NOT get an true average for October until October is finished.
All we can go on is the 30 day average, which currently sits at 6515.
Which is higher than Septembers low of 5766.
So, its highly likely that Octobers average will be higher than Septembers.
How much higher will that be? Well, we won't find out till October has finished.

Oh and I don't want Civ 7 to fail. I am still playing it, and now have 120 hours play time on it.
I am countering people like you that keep saying "we had some days of more positive reviews" etc etc.
Purely because, the figure that counts IS the total number of reviews.
That figure has hardly changed for months.

 

[Evolena]

Today 5 reviews 40% positive

69% now for today's 13 reviews

 

[GeneralZift]

The same level of scrutiny applied by the players? By us, in these threads? I'm sorry, but it absolutely should.

Are we games developers? No (well, I'm not, and presumably you're not). Do we have time? We have oodles of time, I've seen enough of your posts and you of mine

My point is that it's easy to decree that something is bad, or flawed. It's harder to prove it, or come up with a workable alternative. And when we do, then hey, it's a win-win all round. Maybe the developers stumble on a good community idea, maybe it helps us think deeper about the game's systems? Is that not useful? Even when we recognise the limitations of what we can write in a post compared to a game with a living design document and a team of developers behind it.

I make mods. I've worked on some pretty big mod teams, on some pretty big mods (for other games). It's not the same thing as being a professional developer (which I am, in software), but we can absolutely have a crack at design ourselves. It's a time and practise thing.

And yes, I will always read everything I participate in, and no, you're under no obligation at all to do anything that I ask.

Again, I have to disagree. And yes, partially because we are not developers, but also because our designs aren't finished, even if we did spend the time.
Look, I have worked on mods as well, and after having worked on designs myself, there's a certain realisation that comes that the entire game's design on its own is like a structure whereby each section relies heavily on another to stay standing.

But here it is, I've worked on tons of threads:

Of course these vary in size, complexity and presentation, but I have actually spent the time to write up these concepts, and you'd be unsurprised to find that not many people are interested in reading paragraphs of concepts.
For example 'Alternate Victory Conditions' and 'Growth (etc.)' was quite well planned out and saw very little conversation besides "Oh yea sounds cool".

Climate Policies was one I wrote to target uneven starts without just levelling the playing field.
Unique Governments was written about making the gameplay more varied and personalised throughout the game.
Counter-Yields was written to balance wide/tall and curtail strong players.
Alternate VCs was a way for losing players to win the game.

So you can go back and read those if you like, but I think most people prefer the shorter discussion form where they have the time to grasp the concepts without getting into the details.
Because of this:

A: It's difficult to want to write up full concepts​

B: It's difficult to read all these concepts and try to scrutinise it in your mind, in the same way you would a REAL game in front of you.​

C: Details get lost in translation, and people simply prefer to converse about shorter topics.​

When I do go back and check these, even though they are long, they're STILL unfinished in comparison to a professional designs. So, again, no true comparison can be made between the two.

I'm pretty sure Ideology is going to make a comeback in some form. Maybe when / after religion is reworked? Going from ancient religion (with less structured government and / or an ideological bent to speak of) to modern ideologies (with religion falling back to a more supporting role) seems like a natural push-and-pull throughout a game (and especially across Ages, where the balance and impact of each can be tailored to suit the time period).

But I don't think this by itself would solve lategame tedium. It didn't in CiV. I could guarantee AI behaviour quite comfortably based on which one I picked, and which one they'd picked. As you note, the more predictable (and less competent) the AI is, the less of a challenge things are, which also causes tedium.

I also think linking ideologies purely to culture is limiting. I do see the link, but it also affects things like science, production, growth, you name it. There's a lot of flexibility in terms of yields and their application to that kind of thing.

You say it doesn't in Civ5, but you've not played Civ5 have you? At least I recall this being something you said quite often.
Of course the concept itself is not going to solve the issue, but how you design the mechanic w.r.t. everything else that will solve an issue like late-game tedium.

When it comes to Civ7, it's yet to be seen, because they like to wait until later to actually finish their game design --- I mean redesign their game completely --- I mean charge you the price of an expansion what should be in the base game... Oh whoops I've said too much.

I find the AI personally better than in VI or V, though admittedly I'm not a high-level player (I pushed as high as I could go in Beyond Earth, didn't do too badly, and yet that was noted for having easier AI than CiV proper). It's hard to evaluate the AI because the developers should always be seeking to make it better. Which means the quality of Modern (or any Age past Antiquity) should improve over the lifetime of a game.

It's also why I don't rate it much as a suggestion. It's a general thing, it should always be happening, and we have repeated patch notes to suggest that it is happening.

Changing how events apply to the AI is an interesting one, but one I feel creates too much (opaque) RNG. AI works when it feels human. Increasing the arbitrary nature of how it responds to a game's progression is one of those things that feels human, but in practise will just make it seem overly quirky / irrational.

A lot of what people generally suggest when it comes to AI is just sinking more and more resources into it. And that in of itself is a problem. It's not actionable if something costs too much resource to justify. And AI is generally a difficult field. It's not trivial to improve. It's not trivial to make substantial advances in. It's why most, if not all, games use smoke-and-mirrors to achieve the illusion of an effect. So long as the player finds it believable, the illusion holds. The trick is finding the balance between maintaining the illusion and ramping the difficulty. There's a reason "bullet sponge" became a known phrase in other genres of game. It's a boring way of inflating the technical difficulty. But at the same time, creating something that rivals its impact while being more believable takes more resources than most teams have at their disposal.

I once worked with a developer who estimated a project at either 30 or 60 weeks of effort, depending on the approach taken. They were good estimates, they came with a multi-page design doc and implementation outlines. This was close to a decade ago. Neither approach has, to date, been implemented. Or even made the roadmap for implementation. Because there's always something more to be done with those 30 or 60 weeks that has a greater and / or more immediate impact.

This is a big problem for AI in games specifically, because it's one of the most theoretical / effectively compsci parts of games development. And most developers tend to be more engineers than scientists. Proper compsci folk working on AI are working on it as a part of a PhD, or some other postgrad research team. Getting that kind of cross-sector mingling is logistically difficult! I'm not trying to run defense here, I'm trying to explain the difficulty in AI in games in general, and why so many games struggle with it at the high level, especially for longstanding franchises with dedicated communities that have figured out (for the most part) how the game tries to emulate human behaviour.

Quirky and irrational AI can cause Surprises. Surprises are good. Surprises reduce boredom and tedium.

You want players to feel immersed right?
Well in reality, if you know how everyone around you is going to behave with 100% certainty, then it's not immersive at all right?

An AI surprising you with a backstab or a strange move, or simply behaving slightly differently than usual is going to put players more on edge and have them more intrigued as to how the AI actually works.

For this reason, I think all AI needs to have personality scores to determine their gameplay style, and how they react to different events.
Perhaps one leader can really really hate close settling. Another leader could really really love Gold gifts.
This is the type of variety that would have players exploring the game more than they usually would.

That's a weakness everybody shares. Because people view different things as problems, and each have different solutions at times to each and every problem.

Perhaps. And this is why you need more perspectives.
Also: thanks for your reply Gorbles, I only found the time now to reply

 

[IntelligentDisk]

Look pal. You will NOT get an true average for October until October is finished.
All we can go on is the 30 day average, which currently sits at 6515.
Which is higher than Septembers low of 5766.
So, its highly likely that Octobers average will be higher than Septembers.
How much higher will that be? Well, we won't find out till October has finished.

Oh and I don't want Civ 7 to fail. I am still playing it, and now have 120 hours play time on it.
I am countering people like you that keep saying "we had some days of more positive reviews" etc etc.
Purely because, the figure that counts IS the total number of reviews.
That figure has hardly changed for months.

You're right, we won't get a finalised average for October until it's over but we still have the current October average, and so far it's looking good.

I didn't just say we had the best reviewed day yet, but the 2nd best reviewed month since release, and the 4th best reviewed week.

The total number of positive/negative reviews isn't the only thing that matters. The positive % can't rise without new reviews (unless people change their negative reviews), hence why better recent reviews is significant, and worth talking about.