By Radu Muresan
September 18, 2019
Matchmaking is a critical part of Dota 2. The millions of people who play the game each week depend upon Valve’s system to get fair and balanced matches. So the better the matchmaking system, the more enjoyable the game will become for all of us. To that end, while the new Dota 2 matchmaking update seems to have a few flaws, it’s clear that Valve’s heart is in the right place.
On September 17th, a new ranked season began and Valve has announced an entire list of changes and potential future changes with regard to its matchmaking system. I’ll take a look at all of them below.
Smurfing is one of the most annoying things that players have to deal with every day. It’s the equivalent of playing football with your friends and realizing that Lionel Messi is playing on the opposing team in disguise. In such a situation, it doesn’t really matter what you do. You’ll probably still lose the game.
Detecting smurfs has now become even more of a priority for Valve than it was before. In order to better deal with this issue, they’ve announced a series of changes along with potential plans for the future.
The first thing they’re considering is increasing a player’s MMR by more than 25 points per match if they consistently perform at a much higher level than usual (for that division) in order to bring the account to its real MMR much faster. However, the obvious problem with this is the following: account boosters will be encouraged to boost even more accounts than before and the whole account boosting industry will boom as a result of this change. Instead of having to win 10 games in a row in order to gain 250 MMR points, you might have to win just seven or eight. This is an unbelievable incentive for those invested in boosting as a profession.
The main issue with smurfing is that it works both ways. In one direction, you have highly skilled players who ruin games of lower skill. In the other direction, you have account buyers who have been boosted to unnaturally high divisions and are now ruining a lot of people’s game with their lack of skill. By providing a bigger reward for those who are clearly smurfing at the lower levels, you incentivize the boosting behavior and will likely end up with even more smurfs as a result. It’s more than fair to say that both smurfs and account buyers will thrive under this new rule.
With this measure, Valve is trying to do the same thing I’ve already mentioned above: bring accounts to their real MMR division much faster. Well, what could go wrong here? If you thought “account sellers might have a better time”, you’ve guessed correctly. Here’s the problem: in Dota 2, nobody wants to be in the lower ranks. Everybody wants to be higher ranked. Each player, no matter how unskilled, thinks of himself as being at least twice as good as his division colleagues and at least 80% as good as Miracle-. And based on that logic, a lot of people are trying to get to a higher division as fast as possible, by any means necessary.
This is the basic psychology of casual Dota 2 players. And nobody can change that. As a result, anything that makes it easier for people to undeservingly reach high MMRs is generally a bad idea. In order to eliminate smurfs, the first step is to recognize the cause of the problem: people paying other people to boost their accounts. Or people boosting accounts in order to sell them.
I suspect that most of the smurfing activity results from these kinds of motivations. So in order to eliminate this side of the problem, you need to make it harder for people to reach high MMRs, not easier. Instead of increased rewards, offer decreased rewards: the first victory gives you 25 points, the second one 24 and so on. So that in the case of those who win 20 games in a row and are clearly smurfing, there is less incentive to do it.
This is exploitable to the nth degree. I mean, just think about it: you’re playing with unranked players against unranked players. What if everyone agrees to just stay in base and do something else on their computer while the clock is ticking? Playing 10 games of 10 hours each or even leaving your computer open for four days in a row is an easy way of getting ranked accounts with ease, which you can then boost very quickly for money. And even better, it now pays off to have multiple computers.
Pretty soon, you’ll have people account farming without even knowing it. Keep in mind that the server options are still there. One could very easily think of a way in which you set up everything so that the matchmaking system places you in a game-setting that you yourself control. The possibilities for scamming the system appear numerous for those who want to find loopholes.
Ok, this is where I fully agree with Valve’s decision. But how will the assessment be made? And is there a mechanism in place to inform players that if they play support instead of carry or mid, they’re not losing any points? Because otherwise everyone will want to play core, knowing that gpm, xpm, kda and other factors contribute to their initial MMR and that the best way to score high numbers is to play a core role in every match.
This is another great idea but it can be tricky to implement in a way that isn’t prone to risk. On the one hand, you want to catch the bad guys. But on the other hand, you don’t want to punish players who are genuinely good. Especially new players. Otherwise, how do you grow your community. When new people start to play Dota 2, the last thing you want to do is ban them by mistake and convince them that this community is not for them.
Dota 2 is a harsh game anyway, with hundreds of notions to assimilate, a steep learning curve and a relatively toxic community. Having the owner of the game ban you for no good reason is the last thing you need as a new player. But if that’s the case, you need proof beyond any reasonable doubt that someone’s account is being boosted/sold. And that makes it very difficult to actually catch and ban the actual bad guys.
Overall, I feel like the smurfing problem should be dealt with by removing the incentive to smurf in the first place. In a lot of cases, this incentive is a practical one. Eliminate the cause and the effect will decrease on its own. Making it easier for people to get high MMRs might be well intended, but it’s quite naive.
This is a new and powerful concept that has a big impact on every game, deciding to a large extent who wins and who loses. This new role symmetry should play out like this: in a “position versus position” matchup, for each of the five positions, which of the two teams has the higher ranked player?
In other words, balance is not just a matter of overall team rank. It’s also a matter of role symmetry. And in fact, role symmetry may play a much bigger part than the team rank (calculated by taking everyone’s MMR into account). That’s because a big imbalance between the two midlaners can quickly lead to one of them crushing the other and creating a big advantage for his team in the process.
This applies to every role, but the cores will probably take precedence. If you have SumaiL as your mid and he has to play against someone who’s less than half his skill level in the middle lane, winning the game will be much more likely for you than for the enemy team, even though the two teams might be balanced overall.
For the new ranked season, Valve has announced that they’re using this notion of role symmetry in order to create more balanced matches and avoid situations where the cores on one team are much higher skilled than the cores of the other.
As you probably know, since the introduction of role-based MMR (core or support), players are now ranked based on the roles they play. If they play a core position, their core MMR increases or decreases. If they play support, the same thing happens to their support MMR. In their latest matchmaking update, Valve announced a slight change to this rule.
Apparently, a good player still has a lot to offer even when he plays an unfamiliar role. So in order to have a more accurate reflection of everyone’s core and support MMR, the following change will be introduced: every time you win a match playing a core position, your support MMR also increases but only by a percentage of your core MMR increase. The same thing happens when you lose MMR by playing a support position.
Thanks for joining us for my impressions on the Dota 2 matchmaking update! It’s a lot to digest, but I hope you’re leaving with a clear impression of what to expect from these new changes. Stay tuned for more Dota 2 news and updates as we near the beginning of the Dota Pro Circuit.