This topic is aimed to regular players wanting to know more about smurfs, and more specifically to those having issues facing them. With one exception, all the profiles linked here have been anonymized, as it is not the purpose of this thread to have those users individually targeted nor denounced.
SMURFS - The Good, the Bad & the Ugly - Q & A What it is, and how to deal with it
Table of contents
I - Definitions - What is a smurf
II - How to recognize a smurf
III - How many smurfs are there
IV - Why is it bad - What are the consequences
V - Teamgames smurfing
VI - How to deal with it
VII - TL;DR (summary)
This topic was intended as a reference resource about smurfing in SC2, and as such became quite long. However I perfectly understand that most may not want to know everything about it, so the table of content is here to help you navigate directly to the parts you want to read :
Just CTRL+F twice, copy-paste the name of the chapter you want to read, and press enter.
I - DEFINITIONS
What is a smurf ?
To smurf is to play on an account which isn’t your main one.
- High level players (M1 -> pros) use this not to be spied upon (builds, moves) while laddering.
- Regular users (M2 -> gold) use this in conjunction to freelosing, deliberately lowering their MMR in order to play lower level users.
Why ? - What are the uses of smurfing, and why some of them are good, and others bad.
There is some confusion going there : some players do smurf to play anonymously, and some do freelose in order to play lesser users. Most people will call both smurfs, but the aim and the consequences are very different. Anyway, let’s break down the “why” behind those.
- We’ll start with those to illustrate that there are legit uses of smurfing. Having a secondary account, which can’t easily be linked to you, is useful for high level users, as they do, like most of us, use the ladder as training grounds. Yet, as strategies are the heart of the game, they might wish to create new builds, or refine new strategies without their competitors spying on them. This, so they can’t be copied/countered until they have revealed their strategies in a tournament. Leading to the main legit use of smurfing : avoiding being spied/copied by others. That’s what high level/progamers do, and that’s also the primary use of barcodes. Those are what I would call, the good smurfs.
The major difference with the other types of smurfs, is that they DO NOT freelose. They conceal their identity and that’s all. They’re not trolls, nor immature, nor frustrated. Or at least, not as often as the others. XD
- The second legit use of smurf is much less common, it being the casters/tutorial makers doing Bronze to Masters series. If we put aside that executing a strategy with a master’s execution somewhat biases the results, those are used to demonstrate strategies applicable to the different levels of play, and will be useful to beginners searching for builds adapted to their developing skills. This requires a one time series of freeloses in order to reach bronze MMR, but once they’ve reached it they have no reason to continue doing so, and will not repeatedly play users of the same level. One might also add this is a very infrequent use, that most will never ever encounter.
Freelosers (common smurfs)
Those are what people commonly call smurfs, as their numbers are way higher than those of the high level players. Their main characteristic is that they freelose, in other words they do leave a game they didn’t lose, most commonly in the first 10 seconds. Hence lowering their MMR, and effectively playing users below their real skill. There a variety of reasons behind this :
- Can’t stand frustration/defeat/ladder anxiety/stress. SC2 can be a frustrating game, and due to its hypercompetitive design, a great deal of effort can be required in order to effectively improve beyond one’s limits. Some users have gotten so invested that they can’t stand that frustration anymore. And so they want to win in a relaxed manner, or just don’t have the courage to play a fair game anymore.
And so, instead of accepting that a defeat is most of the time due to one’s own mistakes, they refuse it altogether. And so they do massively freelose, in order to reach a MMR where they’ll almost always win the games they invest themselves into.
Hence, reporting the frustration over the players they meet. Those are the majority of the smurfs that players of low to intermediate levels will encounter.
- Can’t stand a specific MU : a variant freelosing only some match-ups, because they can’t stand the other race’s possibilities in that MU or have a style that does applies particularly badly in that case. They start by cheesing, and then, when they are irremediably convinced the MU is imbalanced, do freelose it altogether. Most end up only a few 100-300 under their real MMR however ; and will often be willing to admit it. Though minority, it’s more frequent than one would know. One could partly understand them, but for one they won’t improve if they don’t confront those difficulties, and for two the consequences for the gamers they do play against are the same than against other freelosers.
We could call those two types, The Bad.
- Enjoying in trolling, preying on weaker than you. Those are the ugly ones : experimented users who get bored/stagnate in their own leagues, and who kill time by winning their opponents in the most ridiculous/humiliating manners possible. That’s certainly an immature behavior (some even compared that to bullying colleagues at school), but hopefully they are minority.
We recently had a gold player from our community who boasted how easily he had trolled into complete panic a bronze player. Let me tell you our reaction wasn’t the one he expected.
But yeah, there are some users who not only feel contempt at destroying beginners, but also brag about it. One of them, a master 3 Z/P went as far as sending one of his replays vs a gold player to a well known caster, Lowko.
Here is the list of his accounts :
To users with a keen eye, there are multiple clues hinting the smurfing in this video. More about it on another thread. Since he wished to have his performance publicly displayed on Youtube, he will be the only user that I did not anonymize on that post.
Non freelosing smurfs
With SC2 going free to play, rose another type of smurfs, who don’t freelose regularly, but recreate new accounts, which are initially deliberately placed under their true level. And that gives you users with abnormally low career games for their leagues, and often abnormally high winrates as well.
Some of them then switch to unranked (thus ending up playing at their real level but with a displayed league way below it) ; others simply make sure to never get out of the provisional MMR state, or just start another account again if they do. A good proportion of those just transition to regular play afterwards though, specially if they made new contacts in teamgames with that account.
II - HOW TO RECOGNIZE A SMURF ?
There are several clues that could hint that your opponent is indeed a smurf. Ingame skill put aside, a league past reached higher than the current one, unusually high APM with matching skill, an uncustomized account playing random, a high level of play with a low level of games played (for example D3 with 10 games played), etc.
In fact those are only hints. The most reliable way to confirm someone as a freelosing smurf, is to spot multiple freeloses in his profile’s history.
- If you see several of those, then you can be sure your opponent is without doubt a smurf/freeloser.
If you don’t see them :
they could’ve happened in the past (beyond recent profile history), or you could be facing a non freelosing smurf. Others criteria are then required to affirm that : a total amount of career games abnormally low compared to the league (for example platinum with a 10 games career), a sustained abnormally high winrate, a displayed league abnormally low outside of provisional state (for example if a master 1 is being played by a silver league user), or if the players’ career summary informs you that he reached much higher leagues several times in the past with the same race (≥ 2 leagues at least).
Or your opponent after all isn’t a smurf. Facing a player that has only been one league above current, or more but with another race isn’t enough to consider him a smurf. A player with seemingly high APM could just be spamming. It’s time to look at your mistakes rather than the opponent’s feats.
III - HOW MANY SMURFS ARE THERE ?
It’s quite difficult to assess precisely. I hadn’t the means to do a randomized sampling at first, so I began by simply reporting the proportion of smurfs I met at my MMR (P1-D3). But with good advice, and a bit of help, that changed, so I will separate my purely observational player-experience report, and the fully randomized sampling I’m currently doing.
1 - OBSERVATIONAL REPORTS - PLATINUM
That’s how many smurfs I met directly, as a player. It has the double bias of being applied only at one limited MMR range, and of not being independent from my gaming habits. However it’s quite precise regarding freelosing detection, as I have access to an up to 80 games history, with their exact duration and early events description.
With those limits in mind, here are the proportions of smurfs I encountered in my own region, at P2-D3 MMR :
Weeks 18 & 19 2020 : 15.56% of confirmed smurfs, 84.44% of regular players (including 12.22 undefined).
Weeks 20 & 21-2020 : 17.50% of confirmed smurfs (freelosers), 82.5% of regular players (including 18.75% undefined).
Weeks 22 & 23-2020 : 22.39% of confirmed smurfs (19.40% of freelosing + 2.99% of non freelosing) ; 77.61% of regular players.
Weeks 24-> 26 2020 : 22.81% of confirmed smurfs (21.05% of freelosers + 1.66% of non freelosing) ; 77.19% of regular players.
Weeks 27-> 29 2020 : 24.73% of confirmed smurfs (22.58% of freelosers + 2.15% of non freelosing) ; 75.27% of regular players.
Weeks 30-> 32 2020 : 20.90% of smurfs (14.93% of freelosers + 5.97 of non freelosing) ; 79.10% of regular players.
- 20.64% of confirmed smurfs (all categories) over a 511 players sample.
About the method used :
- Matchmaking used as randomizer.
- Users with ≥ 3 recent freeloses considered as confirmed freelosing smurfs (≥ 2 freeloses prior to week 24-2020).
- Users with ≤ 3 freeloses in recent games history, but with major profile inconsistencies
◦ Total career games too low for MMR ( < 50 for platinum, < 100 for diamond)
◦ ≥ 2 leagues ≥ 2 times with the same race.
◦ ≤ 2 leagues of his opponent outside of provisional MMR.
◦ ≥ 75% global 1v1 winrate with ≥ 20 games played.
considered as confirmed non-freelosing smurfs.
- Users without freeloses nor significant inconsistencies considered as regular players.
- Users who couldn’t be confirmed as smurfs, but with significant profile inconsistencies considered as undefined.
- MU smurf : % of confirmed smurfs freelosing only specific match-ups.
2 - RANDOMIZED REPORTS
Those are fully randomized reports, where I do randomize the MMRs of EU 1v1 ladder players within the range of the league studied, and then pick at random players at that same MMR to find the one I will inspect. From there, I look at their profile and games history, via SC2 API or via the ingame menus. The detailed procedure is described here.
This method provides the advantages of :
— being nearly free of selection biases, as the players selected are now completely independent from me as a player.
— Being applicable to the whole MMR range (and not only to my league).
However it’s less sensitive when it comes to freeloses detection ; and the size of my samples is currently quite limited. With those limits in mind, here are the results, sorted by league :
Gold league — Weeks 31-32 2020 : 23.81% of confirmed smurfs (11.43% of freelosers + 12.38% of non freelosing) ; 76.19% of regular players, over a 105 users sample.
Silver league – Weeks 33->35 2020 : 20.95% of confirmed smurfs (13.33% of freelosers + 7.62% of non freelosing) ; 79.05% of regular players over a 105 users sample.
Bronze league – Weeks 36->38 2020 : 28.57% of confirmed smurfs (22,86% of freelosers + 5.71% of non freelosing) ; 71.43% of regular players over a 105 users sample. Corrected estimates for population : 19.74% of confirmed smurfs ; 80.26% of regular players
- 21.5% of confirmed smurfs (all categories) over a 315 players sample (corrected estimate).
About the method used :
Randomized MMR sampling within one league tier.
Randomized player selection in rankedFTW at the MMR sampled.
Players with ≥ 3 freeloses in their games history considered as freelosing smurfs.
Users with ≤ 3 freeloses but major profile inconsistencies considered as non freelosing smurfs.
Career games < 40 gold, < 80 plat, < 160 diam (MMR reached)
More than 70,0% ratio with ≥ 20 games played
≥ 2 leagues compared to current ≥ 2 times with same race
≤ 2 leagues of opponents with ≥ 25 games played (≈ equivalent of provisional placement)
Users with minor profile inconsistencies or unexplained freeloses considered as undefined.
Users without significant freeloses nor profile inconsistencies considered as regular players.
IV - WHY IS IT BAD FOR THE GAME ?
What are the consequences ?
Increased frustration : SC2, by its competitive and demanding nature, can be a frustrating game. In order to keep players motivated to improve, the matchmaking does balance reward and frustration by placing players at a MMR where they will win approximately one game for each match they lost. Common smurfs do bypass that : to decrease the frustration of a single user, they will increase the one of all the players they did actually play. That frustration :
◦ at intermediate levels will sometimes get erroneously directed towards the balance ; and at a higher level towards unranked.
◦ at lower levels will possibly discourage users to go further.
Increased MMR shifts
◦ Directly : One could say “hey Trias, if freelosers do freelose at every game won, they would give back the points taken, so it’s a zero sum game, right ?”.
That is true only once they have reached the MMR they wanted, and if they do freelose regularly. Yet, to reach that MMR, they will give points to the players above that level, and then take those from players at the level they do smurf at.
◦ Indirectly : more smurfs means more accounts into intermediate to low leagues, inducing more MMR league tiers adjustments in order to maintain the desired proportions of players.
Increased matchmaking randomness :
◦ Directly : if you do stumble against a smurf which freeloses right away, your MMR will be overestimated. If he does play, your MMR will likely be underestimated. Since the primary aim of a matchmaking is to be matched against players of comparable level, and since their numbers increased since F2P, the problem is growing. There are even some hours (late at night) where the proportion of smurfs can create temporary shifts in effective MMR, with smurfs ending up playing other smurfs.
◦ Indirectly : if a significant amount of players of some high league are smurfing, while they are doing so there will be an equal amount of players unavailable for matchmaking in said league. And so the system will widen its matchmaking range, ending up matching players of farther respective MMRs than if not.
Longer queue times :
◦ At intermediate to low levels : even if you’re past being mad at users bypassing the matchmaking, it’s still somewhat irritating to requeue a game after 10s + loading time just because your opponent decided to freelose in your face.
◦ At higher levels : while the user is smurfing, that will make one less player to be matched against in his original league. And in leagues where there already are fewer players, effectively decreasing the available pool will induce longer waiting times before a game can be launched.
- Can this be solved ? Easily, provided the system detects when one user freeloses (which happens mostly by leaving in the first seconds of a game). Some players have suggested pretty good fixes I think, at least better than the ones I first came with :
V - TEAMGAMES SMURFING
It has been brought to my attention that there also were smurfs issues in teamgames. But contrary to 1v1, smurfing in teamgames could take different forms.
1 - Freelosing
- Similar to 1v1 freelosing smurfs, the user simply leaves a game in its first seconds.
- Easiest form to spot.
- However it’s not as efficient in teamgames than it is in 1v1, as leaving a game does not ensures defeat. Even more since the system balances random teams with both users MMR, and so the lower your MMR in random team, the more likely your partner will be to have a higher one. And so if you’ve been matched with a master as a 2v2 partner, and that the opponents are two golds, it is entirely possible for him to win the game despite being left alone. And so your MMR wouldn’t decrease…
2 - AFKing
- This method is more original : it also consists in not playing a game, but by not doing any input during a certain time instead of leaving it right away. This is commonly referred to as being AFK « away from keyboard ».
- The aim is in fact to make sure not only that your partners are being left alone, but also that your resources won’t be shared with them, contrary to what would happen if you had freelost. In other words, it is a way to increase your chances of lowering your MMR with each game queued.
- However, contrary to the freeloses, this also affects all the partners of the team ; who would not only have to deal with outnumbering opponents, but also quite likely see their own random team MMR forcefully lowered.
- Since the main drawback is the time consumption, the most desperate smurfs do use bots set to queue the games in their stead, and to then automatically leave at a set timer. They can be recognized as the games are almost always left at the same duration (at ± 5 minerals mined each), and as abnormally large quantities of games are played in a row (often ≥ 24H long sessions ). This is a form of hack referred to as botting, more about it here. It is of course strictly forbidden by the code of conduct.
3 - Teamkills
- This one is of course specific to teamgames : the user simply attacks his teammates (=teamkilling), most often by worker-rushing them in the first seconds of a game.
- This is not always related to smurfing, as it is not infrequent that this is done as an act of hate or frustration, after being for example matched with a partner they dislike, or having trouble setting for a common gameplan. This can also happen if an indelicate or thoughtless partner happens to have full-walled the base, and leaves it closed, effectively locking-in his teammate (though in that case, the walling player is at fault).
- This method is an even more efficient way than AFKing to make sure the MMR indeed decreases, as it leaves the regular players not only alone but also severely impaired.
- It’s also quite time consuming, and so, as in other toxicity examples, there has been reports of users probably botting to do so in an automated way. Similar to the AFK-botting hackers, they can be recognized by their abnormally long sessions (for example ≥ 24H).
Can this be solved ?
If there was a will from the devs, freeloses could be detected and punished in a similar way the solutions proposed for 1v1 in part IV.
AFKing and teamkills could be strongly reduced with a simple interface option :
I’d like to advocate for a variant where blocking a player prevents him from being in your team.
VI - HOW TO DEAL WITH IT
I have guessed/confirmed in the past that my opponent is a smurf/freeloser. What should I do ?
- Well, play the game ! Specially if this is ranked, don’t let your points get stolen without resistance ! ^^ Take this as some kind of benchmark, a chance to test your strategies against a higher league metagame. If you get defeated (and if you weren’t against the Ugly/troll types =p), analyze why, and you’ll have gained valuable information. Though if you have guessed/remembered your opponent is a smurf, you have the advantage of him ignoring that you know. So you do know you will have to take strategic risks in order to win : cheese him, mindgame him, allin-him, deny him the information. Though the odds are against you, you can win if you play smart. And believe me, if they couldn’t handle the frustration of losing against their peers, losing against someone 400+ MMR below theirs will be infinitely more humiliating to them than to you.
- DO NOT insult them, specially after the game, even if you confirmed that they were smurfs. At one hand, Battlenet’s charts disallow that, and at another hand it will show that you have been overrun by your rage, and as we said some of them do play precisely to trigger their opponents’ rage. Plus it serves no purpose (and more than half of them will have already disabled communications with users not on their friendlist anyway). If you have confirmed that your opponent was a smurf, keep that information for you : it’ll be useful if you face him again in the future.
- They call me a noob/L2P/Gitgud/insult me, etc. Well as we have said, if they were mature they wouldn’t be smurfing/freelosing in the first place. Hence, some trashtalk is to be expected. As a matter of fact, most smurfs aren’t even that good (D1-M3 and below for most of them), as high level players (M1+) are busy actually trying to improve. Ask them their league, and remind them that compared to the league above, they also are noobs. Yet unlike them, you do have the balls to play against opponents of similar (or higher) skills. If you have been really insulted though, you have a valid motive of reporting them through the interface. Smurfing is currently allowed by the system, but insulting isn’t.
- In teamgames, just don’t go for random teams. Play in fixed teams with a partner you know : it will spare you from having both AFK smurfs and team-killers in your team, and will allow you to work on refining your sync, communication and synergy (with prepared gameplans) within your team in a way impossible to achieve in random teams.
- Talk about it. Provided you do anonymize your shots (so that the users do not get targeted), you will get more data about it. This game is built upon players. If there were none but the very best player, he wouldn’t have anyone to play against. Hence, everyone, even those who don’t play to surpass themselves, even those who play casually or who are just discovering the game have the right to have their ladder as it was intended to. If enough people are talking of it, the devs might do something about it.
- Do “GG” at the end of the game, even if they are smurfs, even if they played a match-up you consider favored. And that because it will show that you have accepted that defeat as result of your own mistakes. And by that, you will : at one hand show them that you are not raging (and this will save you lots of trashtalk from those who want to trigger you) ; and at another hand, admit that there was a mistake of yours. Hence, something that you could correct or work upon. Hence, a way to improve.
More GG, more skill - White-Ra
VII - TL;DR
- To smurf is to play under another account than your main one.
- There are some good consequences/uses, and a lot of bad ones.
- You can recognize a smurf by searching for freeloses or inconsistencies into his profile.
- There are about 20% of smurfs in the metal-leagues.
- The devs could easily reduce smurfing, would they be willing to.
You can deal with it :
- by playing smart in 1v1 against opponents you think are smurfs ;
- and by playing teamgames only in fixed teams.