1. Improved Matchmaker
Two months after the initial reveal at WG Fest, we are excited to talk more about major fixes to the matchmaking system. This time, we’ll talk about it in greater detail.
If you look at the current system, it can create 8/5/2 teams, where the two vehicles at the bottom have little to offer against tanks in the middle and at the top. And they rarely get to engage with their brethren from the opposing team: there are just two of them on each side. Pretty vulnerable to all the other vehicles, they’re likely to get destroyed rather fast.
The improved matchmaker puts an end to this imbalance and a few other upsetting scenarios, with a system of patterns and rules. At the heart of the system are three setups: three-, two-, and single-level battles. As for the other rules, they tackle some of the most common queries popping up in the community:
- Eliminate frustrating scenarios where vehicles in the middle and at the bottom of the list are outnumbered by those at the top (just like the one we described above), as well as minimize instances of getting matched at the bottom of the list way too often during a series of battles.
- Ensure a balanced vehicle spread for each team, including the restriction of no more than three arty per side.
- Provide an enjoyable gaming experience for platoons by letting only one-tier Platoons into Random Battles.
- Reduce the instances of map dupes.
The matchmaker analyzes the queue (the number, types, tiers and battle tiers of vehicles in it), and then creates teams based on the speed-vs.-optimal-setup principle. Simply put, it balances getting you into action quickly and creating a perfect match. If the search for the “perfect” match will leave you queueing too long, these rules will be relaxed in order to get you into a battle faster. The longer the wait, the bigger the step-away the system takes from the “perfect” patterns: 3/5/7 and 5/10. This way, you don’t get tired of waiting and always end up with a setup optimal for the current situation on the server. Now, let’s take a closer look at the three scenarios.
Three-Level Battles
If there’s enough variety in the queue and you aren’t playing in a Tier I–II vehicle, you’ll get into the most frequent setup: a three-level battle with a +/-2 Tier matchmaking spread. Having analyzed vehicles waiting to hit the battlefield, the matchmaker will use the 3/5/7 pattern or one of its variations to build a team.
Position in the list | Most frequent vehicle rotation | ||||||
---|---|---|---|---|---|---|---|
Top | 3 | 3 | 3 | 2 | 2 | 2 | 2 |
Middle | 5 | 4 | 3 | 5 | 4 | 3 | 2 |
Bottom | 7 | 8 | 9 | 8 | 9 | 10 | 11 |
Two-Level Battles
You’ve most likely been in a two-level battle with a +/-1 Tier matchmaking spread before. With the improved matchmaker, we are formalizing this setup and, most importantly, imposing clear restrictions on when it can occur.
- Firstly, this rule applies to low-tier vehicles (Tier I–II). In this particular case, it’s designed to create favourable conditions for newcomers.
- The second case is for situations when the vehicle variety in the queue doesn’t allow for a three-level battle, and we wouldn’t want to keep you waiting long to get into action.
Here, the matchmaker will use the 5/10 pattern or one of its variations to assemble teams. Just as with three-level engagements, the longer the wait time, the more the pattern will differ from 5/10. And there will be no battles where vehicles in the middle of the list are outnumbered by those at the top.
Position in the list | Most frequent vehicle rotation | ||
---|---|---|---|
Top | 5 | 4 | 3 |
Middle | 10 | 11 | 12 |
Single-Level Battle
Finally, when a large amount of same-tier vehicles are queueing at a given moment, it can result in an excruciatingly long wait time. To get you into battle quicker, the revised matchmaker will create single-level battles where all vehicles are of the same Tier.
Other Mechanics
Along with the three setups we just described, the matchmaker will receive several new mechanics, all aimed to elevate the playing experience by resolving issues that you reported to us:
Getting into the top/middle/bottom of the list. If you are at the bottom for a few battles in a row, you are likely to get in the middle/top for the next one, regardless of whether you play in the same vehicle or choose a new one during this series.
Vehicle spread. The difference in the number of light tanks, arty and TDs between the team can influence the battle outcome before it even begins. That’s why, when building a selection for a battle, the matchmaker will begin with these three. Teams will be assembled starting with light tanks, then SPGs and finally TDs so that each includes an equal number of these three classes. One note: the team setup also depends on the queue sizes and how each class is represented at a given time. So, the two teams won’t always have all 5 types.
Arty. To prevent arty from spoiling the gaming experience, we’ll limit the number of SPGs per team to a maximum of three per side. Recently revealed in the Sandbox, this limitation was well received by the tester community and proved to have a positive effect on gameplay. It’ll make its way to live servers soon.
Platoons. Another frustrating scenario that has to stop is getting matched against tanks that are 3+ tiers higher than yours. Currently, it often happens with multilevel Platoons. Let’s say, you are playing in a Tier V tank, join forces with a comrade of yours who chose a Tier VIII vehicle, and end up facing off with Tier IX–X vehicles. That doesn’t sound like much fun. The improved matchmaker will allow only one-tier Platoons into Random Battles, solving this issue for good.
Map rotation. Currently, you end up on some maps way too often, and playing the same, let’s say, three maps over and over again is frustrating. It can even get to the point where it discourages you from playing on. The improved matchmaker reduces the number of unfortunate events like that with new logic. It analyzes maps that tankers from the two newly created teams played over the last 10 battles. And then, it places them on a map none of them have seen in the last few sessions.