Community Care Package - April 2022

While we try our best to keep fans updated with important information, we’ve always wanted a place to talk about community issues that may not have been addressed through our regular announcements and Patch Notes. Even though our great team of Community Managers tackle many of these issues on various channels, that info can be easy to miss. To remedy this, we’ve decided to start the Community Care Package, a new, non-regular announcement to talk about some of the most desired community issues.

Why non-regular? We plan to put these announcements out as we have updates on community issues instead of keeping it to a forced schedule. This way we can be the most efficient with the time spent putting them together.

Without further ado, let’s jump right in!

Map Selection

We’re already a few months into PUBG: BATTLEGROUNDS free-to-play and we’re thrilled to see so many new players joining our game. We know that with the influx, many of you in regions without map selection have continued to ask for separation between Battle Royale and Mini Royale maps. This is something we’ve spent a long time trying to figure out, considering many possible ways to make this happen. The unfortunate reality is that even with the increase in players, we still don’t have the numbers necessary in these regions to provide more choice in map rotation without bogging down matchmaking.

The following is a more detailed explanation of why it is so difficult to change the current map rotation.  

  • How we decide the map rotation

As mentioned above, it is true that we saw an influx of new players after the transition to F2P, on top of a dramatic increase in the active user population. We know many of you have noticed this as well. 

However, not every new player is making it as far as the matchmaking pool. Many of the new users remain in the tutorial stage, drastically lowering the number of these players who contribute to matchmaking compared to the increased number of active users as a whole. For example, one user may play 30 matches a day, while another joins the game and exits after playing a couple of tutorial matches. In this case, only one is really contributing to the overall matchmaking pool. 

The list of playable maps can also result in different simulation results depending on what factors we consider, including the number of active users, the number of matches requested, the number of game sessions, etc. Therefore, we decide on the map rotation by factoring in comprehensive metrics such as per-session details and traffic differences by time and mode instead of simply looking at the number of players increased due to the F2P transition.

  • Perspective 

Many of you have noticed that matchmaking speed differs depending on what perspective and mode you play in. Solo Mode in the AS, KR KAKAO, and SEA servers on PC, for example, are seeing games with many bots due to not enough players in the pool. In the most popular mode (mostly Squad) of the same servers, matchmaking is completed relatively easily without bots.

You might think that in these cases it would be okay to allow for map selection in Squad Mode and not for Solo, however, dividing the matchmaking pool of both modes with map selection would only worsen the matchmaking issue, making it very likely that those who prefer Solo Mode would always end up playing matches full of bots. This is because dividing the queues into large maps and small maps means a smaller overall matchmaking pool. 

In addition, if we apply different options for different modes, some users may have to play the mode they don’t want to get the map they prefer. For example, players who prefer FPP would be forced to play in TPP or vice versa.

Last but not least, many of the new players who joined our game after the F2P transition are still enjoying TPP Mode, while the most favored mode on the EU server is FPP. Therefore, the increase in TPP Mode players doesn’t significantly affect matchmaking in the most popular FPP Mode in the region.

  • Bot ratio and MMR

Along with requests for map selection from our community, we also saw many comments regarding bots. As requested, the bot ratio was recently reduced with various adjustment measures. If the bot ratio is increased back again, it would only cause inconvenience to our players.

We understand that players would prefer a zero bot ratio, but we need to consider various possibilities in matchmaking. More bots mean more sessions that can be generated per player. We need 100 players to generate one session, so we can say each player can bring about 0.01 sessions if we assume there is no bot. In other words, the number of sessions where players can be distributed increases because of bots, allowing players to be placed in different sessions based on their skill gap. Therefore, reducing the bot ratio would increase player density, resulting in a wider skill gap among players within the same session.

If we introduce map selection and divide our matchmaking pool into two, it would be inevitable to see either a higher bot ratio or a wider MMR gap. 

The number of players split by half means more bots to make up for the half. Or, if we don’t want to add more bots, we have to match players with a wide gap of MMR together. This would ultimately make new players who meet enemies with way better skills lose interest in our game and leave, which in turn would lead to the return of a high bot ratio as before. 

Based on the considerations above, while we know it isn’t the answer you hoped for, we’re quite concerned that providing map selection just because the number of active users increased would force us to sacrifice many other values and lead to an overall worse player experience for many.

  • Server status

We currently consider matchmaking works properly as long as the average session count per hour exceeds about 300. The Squad/FPP queue of the EU server is an example of the hourly average sessions of 300, which is the biggest server among various other random-map-based servers. It is natural to experience more frequent matchmaking during the peak hours when there are a lot of players, whereas matchmaking is less frequent at midnight with less player concentration. Therefore, we say the matchmaking status is pleasant only when frequent enough matchmaking is ensured, at least 5-6 hours of peak time within a day. Under this criterion, we can say matchmaking is pleasant when the average session count per hour exceeds 300.

If we divide the matchmaking pool by half to provide map selection, the number would also be reduced by half, and the criterion would double. In this case, the bar of per-hour sessions should be about 600 to provide map selection while maintaining the current bot ratio and ensuring pleasant matchmaking after splitting the queue into two.

The EU server, which has the biggest number of players, reached the figure right after the F2P transition. However, the number soon dropped to the extent that made it difficult to provide map selection. It would be a more dangerous move to divide the pool whose traffic is not on an upward trend. Also, please be advised that the situation is not much different, or even worse, in other servers in terms of player count.

Ranked – 10 Division Restriction 

In our current Ranked Mode, players with a 10+ tier gap, cannot team up as a squad. The idea was to ensure that squads of similar tiers were matched together. However, we found that such a restriction serves as a barrier that blocks players from playing Ranked Mode while teaming up with their friends with a 10+ tiger gap in some servers where the skill-based matchmaking system doesn’t work as intended. Moreover, there have been requests to lift the restriction from some server communities since a while ago, and we are fully aware of them.

After putting our thoughts together for a long time and making various analyses, however, we decided to keep the current skill-based matchmaking system. We understand this decision may disappoint the players on the server who made the requests for a long time, and we would like to explain why we made the decision. It might not be the decision you want to see, but we hope the following explanation can help you understand our intention. 

  • Rank Point (RP) difference in Ranked sessions
    • The following are the average and maximum RP differences on the weekend of March 26th and 27th.
Server Mode/Perspective Agv. RP difference  Max. RP difference
AS Squad 864.5 2263
SA Squad-FPP 2199.2 2765
NA Squad-FPP 2063 2063

We can see the AS server has very healthy skill-based matchmaking with an average RP difference of 864. Its max RP difference stood at 2,263, but this came from a match during non-peak time, not a usual case. We can also tell that the average RP difference is around 1,000 in a server with frequent Ranked sessions, whereas servers like NA or SA with smaller player pools have bigger average RP differences. It was an extreme case, but in the NA server, the max and average differences were the same because only one Ranked session was played during the same period. 

As you can see, in many servers, except for some, the skill-based matchmaking system is working properly.

  • RP difference ratio in Servers with healthy Ranked matchmaking

The following indicates the rate of unbalanced matchmaking according to RP differences.

RP range AS KAKAO EU
0~500 1.5% 0.1% 0.1%
500~1000 81.5% 43.0% 31.8%
1000~1500 17.0% 46.0% 53.8%
1500~2000 0.1% 6.8% 7.9%
2000~2500 0.0% 3.1% 4.8%
Above 2500 0.0% 1.0% 1.6%

In the AS server, 83% of the entire sessions showed less than 1,000 RP differences, which means skill-based matchmaking is working properly. In other servers except for KR, matchmaking was made with less than 1,500 RP gaps for more than 84% of the sessions.

Therefore, skill-based matchmaking is properly working for Ranked Mode, with 1,000 RP gaps in most servers.

We are aware that players in some servers where skill-based matchmaking doesn’t work due to not enough player pool have asked us to partly or completely lift the 10 division restriction so they can play Ranked Mode in a team with their friends with a wide tier gap as skill-based matchmaking isn’t working anyway. 

We did review the option of easing the division restriction from 10 to 15 or even 20 through strenuous data analyses and discussions.

Our conclusion, however, is to go back to the basics: Skill-based matchmaking is the foremost precondition for Ranked matches to ensure fair competition. Some players might still ask us to make an exception for some servers where skill-based matchmaking isn’t working properly, but it is technically difficult to differentiate regions. There is no perfect alternative available, but we hope Normal Matches can be the place for free play with friends. 

We will continuously monitor player feedback on our final decision from various servers.

Console – Enhanced Graphics Options

It has been quite a while since we first discussed enhanced graphic options on consoles.  We’re aware that many of you are disappointed that after almost 2 years, you’re still unable to take advantage of your higher-spec console hardware.

It is still early to announce a specific release timing, but since we are getting close to finishing development on enhanced graphic options, we wanted to share as much information as possible through this letter.

  • Stability issues and delays in development

Initially designed for the PC platform, PUBG: BATTLEGROUNDS has experienced stability issues since its console launch. Until recently, we saw some periods of increased crashing, hitches, desync, and graphic loading issues that would sound familiar to those who played PUBG: BATTLEGROUNDS on the console.

Due to a recent surge in console crash cases, we had to put a significant amount of dev resources into further addressing performance issues. Stability issues are usually more severe on low-spec consoles because they have less memory available. When we release new content and features through updates, memory use increases. This also applies when introducing enhanced graphic options that would require higher device performance, potentially worsening stability issues.  

Another reason for the delay in providing enhanced graphic options was that we needed support from platform holders to help us distinguish between the older (PlayStation®4, Xbox One/One S) and newer consoles in our game (PlayStation®5, Xbox Series S/X. Unfortunately, despite our collaboration with platform holders, we faced several additional technical issues that caused more delays in development. However, with the help of the console platform partners, we successfully solved the problems and were able to advance the development.

We are committed to introducing graphics enhancements options without compromising stability.  But please understand that the options may not be supported on all devices as there will be potential stability risks.

  • Targeting level of graphics

Our target is to create an environment where users with higher-spec console devices can play PUBG with a steady performance of 60 FPS at a higher resolution. We’re also planning to improve Anti-Aliasing from the current FXAA (Fast approximate anti-aliasing) to TAA (Temporal anti-aliasing), which will soften the sharp edges you see on objects when playing. That being said, differences in visual performance will exist on various consoles, and some devices might not see any difference compared to their previous generation. In addition, we have difficulties making the best out of the capabilities of the current generation consoles since our development aims to preserve backward compatibility. Thus, newer consoles will share the same system builds as the older ones. So we’re going in the direction of offering the best possible graphic options optimized for individual console models rather than providing the same level of graphical improvements for all current-generation devices.

Anti-cheat

We haven’t forgotten about anti-cheat, but are still compiling that particular announcement. We will be back with an anti-cheat announcement soon.

Conclusion

Once again, this letter is designed to be an update on pressing community issues but we also very much want it to be a place where we can continue dialogue. By nature, that unfortunately means we’ll sometimes have to deliver bad news. It’s never easy for a developer to deliver news that we know players don’t want to hear, but open communication means doing our best to explain why those decisions, while they may sometimes feel bad, are good for the health of the game overall. That doesn’t mean the conversation has to end, though! We’re always looking for ways to improve the game and your experiences playing it. If an opportunity arises to readdress these topics, we will use letters like this to keep you updated on those developments.