How the most popular game on Steam isn’t popular enough

Anyone that plays PUBG in OCE regularly has probably heard of the recent petition that called for the addition of FPP servers to the game. If you haven’t, here’s the lowdown: 3 other regions (NA, EU and Asia) have FPP servers, with OCE, South America, and South East Asia missing out. Recently Bluehole responded to the petition, effectively saying “Your playerbase is too low for FPP”, which has caused a bit of an uproar with some people thinking “How on earth can the most popular game on Steam not have a high enough playerbase”.

Let’s see if that’s actually the case.

First, we need to have an understanding of exactly where people are playing PUBG from around the world. Thanks to Steam Spy, we can get more stats on a game than Steam itself can show a developer, so let’s have a look:

Important takeaways here: Australia makes up 3% of the games’ ownership and active playerbase, and localization for PUBG has really paid off, with China taking the top spot in terms of active players with 23%, and Korea in #4 with 6%.

Compared to my own Battle:Royale style game without localization, you can see that while western countries are roughly in the same positions (Australia with 3.1% ownership here), there’s no sight of China or Korea near the top:

Down To One playerbase — what a real ‘ded gaem’ looks like

Now that we’ve got those figures, and compared them to a similar game in the genre, we can break down exactly how much % population each server group is getting.

After breaking it down into countries and adding everything up, we see the 3 servers without FPP do have the lowest population by quite a significant margin. Because we’re talking about OCE though, we need to get the # of players that OCE will have online in a ‘best-case’ scenario. So, looking at the hourly concurrent player graph,

PUBG is basically at 1 million PCCU. Sounds great right? It is, but that’s not the number when OCE players are playing. Assuming OCE primetime is approximately 8pm at night, then being generous, we can estimate there’s 300,000 players on PUBG in a best-case scenario for OCE. Now comes the fun part. We have to try and work out what % of those players are actually from OCE.

First, let’s multiply the amount of players on at that time by the amount of players we know play the game in OCE — 3.55%. That gives us 10,500 players as a minimum, but that’s not right, is it? There will be a higher % of players on during OCE primetime vs other regions. Using this graph (from a research paper) we can get a clearer idea of how many people we gain during primetime:

Looks like off-peak is about 80, average is ~250 & primetime looks to be about 600, so there’s a 2.4x player increase from average to peak times.

Multiplying that by our original OCE figure of 10.5k, we get 25,200. Having a think about that, it seems possible, if on the high end (already beating out what the whole of BF1 gets on PC sometimes IIRC).

Quick break.

Now, out of those 25,000 people in primetime, we need to split them up. If we want to add FPP servers to OCE, we’ll have 6 potential queues that people can join, so let’s split that into 6 (as we don’t know the ratios of Solo/Squad/Duo): we’re back to 4,160 per queue. What about third person/FP split though?

A survey that I did for Down To One showing that 50% of people can do without TP in their BR games shows that we’re pretty safe in assuming 50/50 split, maybe leaning towards 1PP.

But — and worth remembering here — not every one of those CCU are going to be actually in a game. Steam Spy counts anyone idling in a lobby, which at a very conservative estimate would be 15% of active users (if anyone has any hard data on this let me know, I couldn’t find any). So, we’re down to about 3,530 again.

Now, let’s work out average wait time. For simplification purposes, we’ll assume that servers run for 40 minutes total, with 50% of players (50) out of the server by 20 minutes in. Taking a nice round number of 500 people to work out our first calculation, we come to a number of 4 minutes average wait time per person. Sounds about right.

Now, with 7x more people, we’ll have 7x less wait, so that will come down to 34 seconds wait per person on average, at primetime, assuming 100% of OCE players queue for OCE, assuming only 15% of people are lobby idling.

In an off-peak best case (7.5x less people on non-peak hours, everything else the same) that’s a ~4.5 minute wait time. After speaking with some OCE members who have tried queuing for 3P in off-peak, this number seems too short for squad + duo queues, which already struggle to find a match at that time.

The Conclusion

Even after being generous with our assumptions, the case for OCE FPP servers is still strong, but not as clear-cut as some people think.