I create this thread upon noticing that there is a lack of a suggestion thread on this very much important matter, that needs discussion. The premise is simple, should a skill-based matchmaking system be added to War Thunder?
For simplicity’s sake, from here on out, I will here from out refer to skill-based matchmaking as SBMM for short.
These polls will be split into a few categories, the reason there is so many is to gather as much data/opinions on the subject as possible to possibly help Gaijin with implementation.
IMPLEMENTATION
- Yes
- No
- Like in other games, where we have a “casual” mode and a “competitive” mode
- The entire matchmaking system should only use SBMM (no new “casual mode”)
- Other method (explain in comments)
- I voted no on the first question!
- Yes
- No
- I voted no on the first question!
METHODOLOGY
- Using match outcome data (so win/loss data, placement on team, match activity & etc)
- Individual player skill/impact (think raw performance; kills, score, assists, kills per life & etc)
- Contextual factors (did the player drastically outperform the rest of the team? did the player win against mounting odds? what vehicle was used? was the player up-tiered? and more)
- Temporal conditions (basically, monitoring player performance in the long term; are there spikes in player performance? or sudden drops? etc)
- Other (elaborate in comments)
- I voted no on the first question!
- It should be taken into account, because gameplay can be vastly different at different BRs
- Only when gameplay is vastly different (Modern BVR/ARH gameplay is much different to that of WW2 propeller combat, for example)
- No, because player skill is gained over time and is reflected by the skill rating
- I voted no on the first question!
GAME-MODE/DIFFICULTY DIFFERENCES
- Players should have a different skill ranking for the 3 game-modes
- Player skill ranking should be shared between game-modes
- It should be shared, but with some caveats taken into account with how players perform in different vehicle types (especially ships, planes and tanks)
- Other (elaborate in comments)
- I voted no on the first question!
- Arcade Battles
- Realistic Battles
- Simulator Battles
- I voted no on the first question!
A preface to some of the upcoming points I’m about to make is that they are my opinions, based on my own roughly 3,3k hour experience playing the game. Which has been spread around to air and ground. I have not played navy to a major extent, so keep that in mind.
I aim for this thread to become a discourse point regarding the implementation of a SBMM system in War Thunder, much like my other big thread has become regarding the utilization of SPAA; and how player skill obviously plays a part in it’s effectiveness.
I will use the spoiler feature to split my points up into easily accessible text boxes that you can close and open as you please to avoid cluttering this thread. These will be my own core points to why a SBMM system should be implemented to War Thunder.
Another caveat is that, by the nature of this post, and the scale of the suggestion, a implementation of a SBMM system would most likely require some level of overhaul of the current matchmaking system. As such, I will touch upon the matchmaker as a whole, but focus on the implementation of a SBMM system.
1. The state of the current matchmaker, and how it is detrimental to gameplay
War Thunder’s current matchmaker is a system that many of you are undoubtedly familiar with, and it is in my opinion one of the game’s most flawed systems. Yes, it gets the job done, but it does so in my opinion, very poorly. Players are placed into matches with little account for their skill, line up (outside of maximum battle rating) and intentions that match.
This is especially important when it comes to game-modes like air realistic battles where players only have one life; where the impact of a single player is much greater than that in ground realistic battles. The issue this creates is simple. Teams are often awfully balanced, and often lead to full steam rolls where one team dominates a match. This is a very common experience in Air RB, and especially in ground battles at the higher battle ratings.
Yes, there are matches when by sheer luck, equally skilled players are put against each other, and we have good matches that can last the full duration. These are some of my personally favorite matches of all time. But they are rare.
It is obvious that this sort of lack of balance leads to a poor experience, where players are put against variably different opponents in terms of skill. This happens at lower BRs and at higher BRs. Leading to short, unfulfilling matches.
As such, the simple solution to the problem is to implement measures to measure player skill, and place players against opponents of equal skill. In this manner, newer players can have an easing into gameplay, rather than being thrown against veterans with far more skill than they would ever be able to counter. A SBMM system would go far in ensuring that players fight opponents of equal skill, especially when no such system is in place at the moment at all.
2. Possible issues that could be spawned from implementation of SBMM
One of the biggest concerns I have is if Gaijin fails to properly implement the system. As such, I push forward the notion that Gaijin should be very transparent about the implementation of such a system, if it ever comes to it in terms of implementation.
Secondly, queue times could take a hit when it comes to implementation of such a system. I have no idea by how much, but I’m sure Gaijin could possibly figure something out.
Gaijin has set a good trend as of late when it comes to broadcasting their next moves regarding updates to the came. As such, I feel they could extend the hand out and properly coordinate with the wider community via feedback prompts and other forms of communication to slowly craft a system that works extremely well.
Another one is that of smurfing/seal clubbing being made “official” which is an issue that already exists. In my opinion, it would be made less prevalent by the implementation of a SBMM system. If Gaijin creates the proper metrics/systems for measuring player skill, smurfs would be very easily noticed as abnormalities in the system (spikes in performance for just created accounts are statistical abnormalities.) General exploitation of the system should be something that Gaijin takes into account in it’s implementation.
Cheaters would also have a vested interested in a system like this. Especially if a reward incentive is created by rewarding players reaching higher ranks of play. But this is unavoidable. At the same time, I feel a SBMM system would also give Gaijin the ability to track player performance in a way that could be used to counteract cheating. (Think how VACnet works at VALVE.)
3. Caveats to the different game-modes & difficulty levels of War Thunder
Another thing that should be is addressed is how would player skill transfer between game-modes? Especially when there is a mixing of what vehicles can be used in what game-modes? This is a balance that should be found, and will take some time to figure out how it should be done.
This is why I have left poll options for that matter to be voted upon, because it’s implementation will be important. In my opinion; air, navy and ground should have their own skill levels as there is quite a difference in gameplay between the game-modes.
There also some differences to how different modes play based on difficulty, especially apparent is the difference between simulator gameplay and arcade gameplay. This is another important matter that needs to be properly discussed. In my opinion, each gamemode naturally should deserve it’s own skill rating, maybe with the exception of simulator battles, but who knows. It feels like Gaijin may be moving on from SB with the coming of Aces of Thunder.
4. Possible concerns for the different ranks/eras of gameplay
Anyone who has played War Thunder for an extended period knows that the game is very different at the different ranks and battle-ratings, the game becomes very different in terms of it’s gameplay. Nowhere else is this more apparent than in air battles. Sure, some core concepts still apply even at higher BRs but for one the complete absence of missiles at lower ranks means that players who for example, buy their way to the top, will be completely thrown into a different world of gameplay.
I also say this as someone who did the same thing to some extent, preferring to skip the propeller era of gameplay by a long shot to get to the jets. I have since then returned to try out many propeller aircraft. But the point still stands that at different ranks, the gameplay is different. As such, it should be discussed and taken into account.
In my opinion there are two camps here. One could be that player skill, as we learn and get better, will ultimately mean we will understand more about the game over time – meaning that generally speaking, players at high skill levels would know how to play at each rank. However, in my own time of playing, I have met those who only play at certain ranks. I have no idea how big this group of players actually is, but they exist, and thats what matters.
What do you think regarding this?
To conclude this section, I will list the general upsides and downsides a SBMM system could have for War Thunder, in my opinion.
Possible Positives
- Matches would be far more balanced, and most likely last longer on equal terms – improving the gameplay experience for everyone
- Through the SBMM system, players could gradually improve by fighting opponents of equal skill
- SBMM would highlight War Thunder’s potential as a competitive game, giving players easier access to the competitive scene
- A new SBMM system could give Gaijin the ability to monitor new statistics, which could be used to balance the game and possibly even counteract cheaters
- Creates a incentive for trying your hardest each match to win, rather than focusing on increasing RP count on next vehicle
Possible Negatives
- Poor, rushed and miscommunicated implementation could result in more issues than it fixes
- Possibly longer queue times based on how the system would be implemented
- Exploiting/engineering of the skill ranking system by methods such as stad padding, could become prevalent if the system isn’t built to be smart
- Up tick/increased interest in cheating software, botting or other third party programs being used to exploit the system
I think one of the most important things establish early on is that Gaijin should communicate with the community when it comes to the possible future implementation of a feature like SBMM.
A proper rollout/testing plan would ensure that Gaijin is given enough time to iron out any kinks with issues regarding the system, and giving players opportunity to test it would be ideal. Similar to how we have had events like the SAM site destruction one recently, large scale tests like that with feedback prompts in-game rather than outside would give Gaijin the opportunity to collect valuable data over time from large parts of the community.
Thus, I push for the notion that Gaijin if they choose to implement such a feature, would properly communicate and test their implementation of SBMM rather than throwing in into our face out of no where.
Another thing that is to be made clear is the way metrics are used, and in what sort of algorithm. While I am personally not the most technically knowledgeable when it comes to this sort of subject; I have done some quick reading into how this stuff works to at least get a surface level understanding of this sort of thing is usually done. Which is why, I urge anyone more technically knowledgeable to give their feedback when it comes to proper implementation of a SBMM algorithm.
Anyway, there are a few systems used in the industry, many of you probably have come across the term “Elo rating (system)” maybe once before in your time on the internet. Such a system was largely designed for rating 1v1 chess style matches, and it works for that. Obviously, such a system only works for that, because it’s designed for that. It works largely by just giving players points, and then subtracting them based from the loser, to the winner.
This doesn’t work for WT’s format, or really any other format, which has lead to systems like Microsoft’s TrueSkill being created. These use more complex mathematics to represent players as differing variables, taking into account the differing nuances of player performance; which I listed at the start in the polls. In general, we would want a system that is smart, flexible and robust. A system that takes into account practically everything the player does; because it gives Gaijin the ability to rank us better. We can combine these methods with other ones, such as purely ranking player performance; which Gaijin does seem to do to some extent already to base vehicle BR balance on.
In general, the core focus in my eyes should be on player skill on the individual scale. So kills, score, battle activity, deaths and etc. Of course these would also have to be tuned for different game-modes. The reason I personally think it should focus on individual player skill is that most people do queue in alone.
For the actual implementation, many differing factors should be taken into account, and from my quick and brief understanding it seems a Bayesian style formula is the one utilized the most. Of course, I think the developers themselves could figure something out regarding this if it ever comes to that. This is just my suggestion.
Other things that would help are the incorporation of placement matches, also rank decay to promote active gameplay and avoid skill drift/rust. In general, Gaijin does have many tools they could use to make this really easy for themselves and I believe the development team could implement this very easily…
Wrapping this up, I want to spawn discourse regarding this subject. Much of what I’ve said here is my opinion, so is anecdotal as it gets. But I feel there is a large part of the player base that agrees with the premise that we should finally get a system akin to SBMM. Which is why I open this thread to begin with. To spawn that discourse, in a way where Gaijin may actually notice it and we may finally do something about it.
Please, vote, and comment. Make this thread visible and relevent for staff, so it may be forwarded to developers some day; and we may finally get a modern, robust and great matchmaker for War Thunder. Thank you.