Discussion Life-Time Steamspy Owner Estimates for STEAM Titles released in 2019 - 2020, having minimum 500,000 Owners

[AshenOne]

What is Steamspy?

As per the ABOUT page, here's what it says..

Steam Spy is Steam stats service based on Web API provided by Valve and cool idea of Kyle Orland from Ars Technica. Steam Spy automatically gathers data from Steam user profiles, analyzes it and presents in simple, yet beautiful, manner.

For more information on its working mechanism please refer to the ABOUT page i linked above!

DISCLAIMER:

Since Steam implemented its GDPR update in 2018, Steamspy's inaccuracy has grown and so has its margin of error but it still it seems like it has some credibility and reliability to it with its new algorithm but obviously that means that its range at which it estimates the owners will go wider and not as specific as its previous algorithm which had a 10% margin of error but using the 10x multiplier of concurrent player number peaks for these games along with this new algorithm does help it! So PLEASE take these owner estimates just... as estimates, neither as factual nor as official sales numbers of these games and with a grain of salt ofcourse! This thread is ONLY for preservation purposes and fun comparisons with actual game sales when announced

Also... As to how do I get this data...its thanks to a fellow ResetEra (and Metacouncil?) member who is a Patreon of Steamspy and that person kindly shares the data with me!

Further proof that Steamspy STILL can be reliable sometimes..recently at least:

• It was pretty damn close or had reached 500, 000 estimated owners by the time, Atlus West made the 500, 000 Sales on PC announcement
• It had 2 Million Owners around the time Fall Guys Devs announced 2 Million Sales for the game
• It had around 1.6M Sea of Thieves estimated owners around the time Rare announced that Steam version of the game had sold over 1m copies on the platform

Anyways now on towards the main part which is about...

Life Time Steamspy Owner Estimates for STEAM Titles released in 2019 - 2020, having min. 500,000 Owners:

2020:

2019:

Note: I deleted the FREE entries from the list so there are no FREE or F2P Titles in the list. Also sometimes games like the Halo: MCC for example, will count DLCs owners or different edition owners as one game so please keep note of that.

 

[warp_]

holy shit that fall guys number

out of curiosity what was the top f2p game?

 

[Ex-User (119)]

I appreciate info like this but so many are so inncurate yes fall guys has sold alot but no where near 15 million.

Devil may cry is probs at like not even 800k on steam considering shipped (not even sold through) number we have for all platforms is 3.6 million

But here it's saying dmc 5 is at 4.5

It really is telling how false this is

 

[AshenOne]

I appreciate info like this but so many are so inncurate yes fall guys has sold alot but no where near 15 million.

Devil may cry is probs at like not even 800k on steam considering shipped (not even sold through) number we have for all platforms is 3.6 million

But here it's saying dmc 5 is at 4.5

It really is telling how false this is

Yeah, this is why i've stated that it can be really accurate too sometimes like with the examples i gave but if the game sells really well and above 2m or something that things start to skew off according to Steamspy's creator and other stuff like counting expansion sales along with the game as ONE like how steamspy counts Halo: MCC DLCs aka the individual games as a single entity when showing estimates.

 

[Maniac]

Using SteamSpy for any kind of serious data analysis or assumptions is bad form. Please don't post something as misleading as SteamSpy's numbers. The error margin isn't just big, it's astronomical. We're talking a game being able to have sold 5 copies but be reported as selling a hundred thousand.

Just... Don't bother wasting your time on bad data.

 

[Ex-User (119)]

Yeah, this is why i've stated that it can be really accurate too sometimes like with the examples i gave but if the game sells really well and above 2m or something that things start to skew off according to Steamspy's creator and other stuff like counting expansion sales along with the game as ONE like how steamspy counts Halo: MCC DLCs aka the individual games as a single entity when showing estimates.

Its just a pointless thing like the numbers are far to wrong there's nothing to gain when the numbers are far too inncurate it's just a complete disservice to make a thread talking bout said sales lol

 

[Maniac]

IF there was ANY credibility to the algorithm in use it would EITHER A: Be very accurate at lower volumes or B: Be very accurate at high volumes. SteamSpy has proven itself to be neither with the exception of lucky guesses that are close enough that one can justify it as just being margin of error.

Surely you've heard the metaphor about the broken clock, yes?

You're basing statistical analysis off of a broken clock.

 

[Prodigy]

Yeah, lots of these seem completely off. Fall guys is probably closer to 0copies sold than 15Million, that's how bad it is. There are some semi accurate numbers but it is too inconsistent to really ever know.

 

[Alexandros]

I wish the industry was more open with stuff like that.

 

[warp_]

dang too bad...that was a hell of a number. woulda been great for a devolver game to hit that big.

 

[Wok]

Fall Guys has 120,000 Steam reviews. It could be close to the 5-10 million range of copies sold.

You need to reply to this thread in order to see this content.

Actually, similar to M&B II. These games should be closer.

You need to reply to this thread in order to see this content.

 

[Wok]

That is a factor of ~ 54.

 

[Durante]

It seems like reviews * 50 is still a better ballpark estimate than either Steamspy or Playtracker

 

[Prodigy]

It seems like reviews * 50 is still a better ballpark estimate than either Steamspy or Playtracker

Yeah that actually seems true.

 

[Mor]

It seems like reviews * 50 is still a better ballpark estimate than either Steamspy or Playtracker

At the very least it gives you the area where the real number is, for sure, but if you want something more specific, we still need those services to have, at the very least a rough idea of where the true point is.

 

[Alextended]

Some times it's accurate in retrospect, except until real data is given you can't know when it is and when it isn't so it's useless until the real data come, which also make it useless. And it's an op pimping steamspy and showing off support for it (pr and monetary) forgetting if not avoiding to mention its relation to EGS?

 

[Wok]

Yes, we have to be wary of the data which was used to get these estimators. For instance, will PlayTracker fiddle with its algorithm so that it gets the numbers right for Fall Guys? Then PlayTracker would be spot on... a posteriori. But that would not mean that the algorithm is able to generalize well.

---

As for the ratio between the numbers of copies sold and reviews, it looks good, but I wish there would be more transparency. I am sure that we could get more information from the anonymized data (if it were shared) than just a range [20 ; 60]. Sadly, the guy with the data does not seem to be inclined to show more stuff, like a scatter plot (sales, reviews), a scatter plot of log-values (to make it more readable), or the distribution of log-ratios.

Anyway, it is interesting to know that the chance to write a review is somewhere between 1.5% to 5% for most of the tested games released in 2020.

Edit: I realize that we don't even know the number of such tested games. We know there are 237 games in total, but we don't know how many of them were used released in 2020 and thus used to get that [20 ; 60 ] range. sigh