Gamstat.com estimates are based on a representative sample of ~6.5 mln accounts retrieved via PSN API. The obtained data is extrapolated to all PSN accounts with a small margin of error.
Every public PSN account includes a list of played games (trophy list). When we collect a somewhat significant number of accounts, it gives an idea of “who played what” by this date: 0.5% of gamers have played Goat Simulator, 7% – Mortal Kombat X, 10% – The Last of Us, 33% – GTA V, etc. Which allows putting a few dots on the graph, arranging all games according to their relative numbers.
What makes the data especially valuable – is the fact that PSN also publishes the date when every trophy in the game was earned. By looking at the earliest trophy we learn when each player started it. (Well, often with a lag of a couple of days, which is a small price to pay for learning what PSN looked like years ago.) Even if a player hasn't earned any trophy, the game still appears in the trophy list with the date it was launched.
After we calculate how many players started playing each game on each day – we can go further and draw the whole graph, although without scale labels yet.
Now there is just one thing missing – the absolute numbers. Knowing that one game has three times more players than another is good, but is it 3 mln to 1 mln? or 7.5 mln to 2.5 mln?
In fact, to scale this graph for all games at once we would need just one strong number – for any game, for any date in the past. Then we would be able to relatively calculate every other number based on it.
Fortunately, we have not just one, but hundreds of those numbers – thanks to Sony and its “MyPS4Life leak”. On 13 December 2018 PlayStation Europe launched “My PS4 Life” promotion videos featuring some of the players' personal stats. And it turned out to be a gateway to calculating the total player counts for most PlayStation 4 games.
Original Reddit post, which, coincidentally, was how gamstat.com started.
Among some personal stats, every video disclosed player's rarest trophy which included game title, trophy name and, crucially, the precise number of PSN users who had earned it (as of 17-19 November 2018, as we found out). Combined with the percent of players who had this trophy (this information is always available in PSN) it gave the total player counts in one arithmetic operation.
Top 100 most popular PlayStation 4 games as of mid-November 2018, based on the MyPS4Life leak.
By combining our graph with the numbers obtained via “MyPS4Life leak” we can scale it and calculate the absolute numbers for any PSN game, no matter the platform, and for any date – even for those released after the leak happened (like game “C” on this graph).
This is basically how gamstat.com works, with some minor adjustments and tricks behind the scenes.
Questions or remarks?
The official sales numbers are much lower than the estimated player counts.
Yes, they should be. The more they proceed in time, the more they diverge. There are many ways to play the game without buying it from the store: rented games, used games, PS Now, PS Plus, EA Access, family accounts, guest accounts, extra accounts, etc.
So your numbers are useless.
No, they are just different (and personally, I think they are more interesting and useful than just primary sales since they include every possible way to obtain the game). It is more of a way to estimate games' popularity rather than counting developers' pockets. Player counts and sales numbers should complement each other, rather than conflict.
I also provide a set of other metrics to give a better picture.
I found some other website's estimates, and they are very different from yours.
So it's their word against mine. It is up to you to decide which numbers to trust. By the way, did they publish their estimating method?
Still, some numbers look impossible.
Since these numbers come directly from PSN and the official MyPS4Life promotion – you should argue with Sony first. I'm just trying to mirror PSN as accurately as possible. But it might be a bug or a faulty logic, and I'm always willing to investigate and improve my algorithms, so feel free to contact me.
twitter.com/GamstatC – I prefer discussing in public.
But how do you know that “MyPS4Life leak” was real? Sony has never denied nor confirmed it.
As you previously learned, at some stage I combine hundreds of “MyPS4Life” numbers with my relative estimates. If “MyPS4Life leak” wasn't real, it just wouldn't be possible – the numbers wouldn't add up. There are also secondary ways to check them – like comparing the official number of PlayStation Plus subscribers with my estimate, derived from the long formula – and I've tried pretty much them all. That leak was real.
This game's servers are closed, but gamstat.com says it has 1000 monthly active users.
Active players are those who launch the game. They are still counted, even if the only gaming experience it has to offer is the “servers are down” message.
What's with Fortnite: Battle Royale? How much did Epic Games pay you to hide its MAU?
Fortnite has two modes: Battle Royale and Save the World. And in Battle Royale mode there are no trophies. When someone hasn't earned a single trophy in the game, normally it would still appear in their trophy list (with zero trophies). In fact, this is also how it worked for Fortnite before the 29th of March 2018. Then in a patch v.3.4.0 they changed it. Since then if someone plays only in Battle Royale mode – Fortnite doesn't appear in a trophy list.
I use another method to calculate its total number of players, with a bigger margin of error and providing only the lowest estimate. Its number is probably higher.
And, alas, Fortnite: Battle Royale's new players and MAU can't be calculated this way with a reasonable margin of error.
If the game is available for several platforms, then how the numbers are combined?
Someone played a game on different platforms from the same account = one player. It brings some counterintuitive side-effects like when a game can have more new PS4 players than overall new players (because they've already played the PS3 version and thus new to the PS4 version only).
Last month your site said this game had 50 mln players, this month it gained another 1 mln, but its total number became 49 mln – even less than before!
All the numbers are recalculated and become more accurate with every update – don't treat them as final. And that's what “10% margin of error” remark is for.
Why do estimates become available with a lag of several days?
PSN API rate limits are very low. It takes several days to collect enough data for reliable estimates.
Why isn't gamstat.com updated every day?
PSN API isn't exactly a bowl of cherries and I haven't yet fully automated the process, so it takes a lot of my personal time. So far it's just a hobby. If you want it to become my job – become a patron.