SMURFS REPORT W30-32
GOLD LEAGUE
Alright folks ! So, contrary to the usual, this is a fully randomized report about smurfing within gold league. It may not seem like it, but it’s a tiny revolution compared to previously, as the results are now completely independent from me as a player, and also completely randomized (MMR sampled, players sampled) to ensure maximum reduction of selection bias.
Anyway, here are the results :
How to read : cf. first post (Part II).
Method used : cf. above post.
Quickie :
Over 105 users at gold MMR :
-
23.81% of confirmed smurfs
(11.43% of freelosers + 12.38% of non freelosing).
-
76.19% to 42.86% of regular players.
Several remarks about this data :
-
There is a certain degree of uncertainty due to the small size of the sample. It could be appreciated by providing a confidence interval, but I haven’t learned yet how to calculate it and, hey, we’re just talking about an RTS game. However it’s quite close to what I’ve reported within my league, which suggest both reports provide acceptably accurate results.
. -
Though certainly theoretically less biased than my player-experience reports, this method has a number of shortcomings, as its wayyy less sensitive when it comes to freeloses detection : I’ve got only 24 games to work with, minus the last game played (can’t deduce duration), minus any custom/co-op/teamgame/arcade map. And if the games are played apart from each other, they will look like normal games into the API. Hopefully, if a player is in a clan, I can switch back to his ingame history which is much more detailed. And… well I’ve got an easier access to their league progression, which is in fact also a solid hint about non freelosing smurfs. That’s why this method probably underestimates the amount of freelosers, but compensates by better detection of non freelosing smurfs (which reflect the original use of smurfs accounts).
. -
If we sum-up, that’s about one fourth (25%) of smurfs at gold MMR, compared to roughly one fifth (20%) at platinum. That’s quite a lot, and as I expected, the proportion of smurfs seems to increase while the level lowers. The next reports will confirm or not.
PLATINUM 1
The usual report at P1-D3 MMR :
Quickie : 20,90% of smurfs (including 14.93% of freelosers), 79.10% of regular players, over a 67 users sample.
With this report, we’ve reached a 500 players sample, which I consider is enough, so I will stop my platinum player-experience reports there.
I’d like to thank all the smurfs who unwillingly participated to this study by trying to steamroll (a less and less) unsuspecting player ; to those who almost fell off their chair discovering that I had guessed right away ; and to those who allowed me to report them by stooping to lowly insults.
I’ve met a few smurfs that were actually pretty chill guys, and I’m sorry if I didn’t extend the conversation for overly long, but you guys should understand that what you’re doing is far from chill/good. It would be a pleasure to meet again the two or three of you, in another league, once you’ve stopped smurfing. 
FUN FACTS
- The girly pseudos being associated to smurf accounts seem to be a thing even in gold league. But in fact, it also applies to more generally lousy, dumb, or somewhat trolling names that are linked to smurfing, as some of them can’t resist the idea of starting to troll right away. It verified more than once through this gold report. If you see someone with a joke-like name (for example OmegaPoop, Rosie, or IreallyLoveU (I’m making those up, not quoting real users)), or a barcode, start considering the possibility that someone is trying to smurf you. I may make a best of of all the hints you can get in the future.
Once upon a time, there was a platinum player, that we will call Obiwan. He was matched against an unranked barcode, that we’ll refer as Anakin. He thought to himself « Damn, I’m 80% sure this is another one. I’m gonna cheese him, at least the game will be short ». Obiwan was wise, but he hadn’t foreseen at what point he would be true, until this happened :
We both had the same idea. Only, Anakin the Barcode committed to his plan despite being scouted right away, while the gentle platinum decided to cancel and build back rax and bunkers at home.
And though minute, the defender’s advantage of having the high ground was decisive, as though losing a lot of SCVs, it allowed the platinum player to snipe the enemy’s repeatedly sent SCVs, and then to finish his own bunker. And that is how the highground, once again, ended up being fatal to Anakin. 
That’s all folks, see you in three weaks for silver league ! 