This keeps happening way too often. Keep getting matched with someone whose highest ELO is ridiculously high compared to mine. Check the aoe2 companion app and you can see so many of these smurfs have ridiculously massive win streaks followed by equally enormous losing streaks. With literally no wins during their huge losing streak. Whereas legit / fair players have a much more evenly distributed mix of wins and losses, up and down.
Ranked matchmaking should be based on each players’ highest ELO, not current ELO.
@ferchalurch Thank you for bringing up the comparison to players of other games. For example, here is the ELO of a top 30 chess player who is not a smurf but had a fairly significant losing stream followed by a plateau and rise.
At least for 1v1 taking highest elo will not work:
Elo is “relative”, one elo alone means nothing
Therefore you will make the life of people who come back after a break hard as they loes skills
Also consider that the absolute skill of playerbase keeps increasing, the highest elo in one day may mean nothing in other day
Also consider variance in game skills - I am 150 elo off from highest, I feel like I am playing titanic these days and sometimes I feel like everything is smooth
@NastyHigh brings up a good point about correcting ELO for people that intentionally lose a series of games. I think there is a solution that would also compensate for disconnects, newly created smurf accounts, and laming.
The maximum exchanged ELO win/loss points should not exceed game time in minutes, rounded up. For example, if a game only lasts 1 minute or less, then the winner can only get 1 point and the loser only loses 1 point. A 16 minute game could have up to 16 points at stake. And so on.
A smurf that intentionally throws a series of games would not see their ELO decline as much unless they idle their computer/game for some period of time.
A newly created smurf account of someone that just stomps opponents would still rise quickly enough, because even a constant drush or going feudal with scouts is going to take 10 minutes or more to finish a game. So, enough points would be exchange to rise the smurf account.
Lastly, on team games, if someone disconnects or crashes within the first minute, the team is only out 1 ELO each if the remaining players resign.
Many players consider laming a breach of etiquette, even though permitted by the game rules. So, if this caps the amount a points a lamer will get (because the lamed player quits within 3-4 minutes), I would consider this a benefit.
True, but at least that player is blocked from playing AOE2 during that time. Yes, they could run the bot at off-hours, but the scheme is getting more bothersome for the smurf.
Many players consider laming a breach of etiquette, even though permitted by the game rules. So, if this caps the amount a points a lamer will get (because the lamed player quits within 3-4 minutes), I would consider this a benefit.
I mean if I can quit freely while lamed, I can always be absolutely greedy and push all deers instead lol in the end everyone will just be busy pushing their deers
Also lith 3min drush and/or vill fights will become next to useless
It sounds like the problem is that some players Alt+F4 to avoid maps and keep elo high, while others instantly resign to drop elo and play against weaker players. I don’t really see a way to solve both of these, when players have opposite motives.
Now that’s a terrible idea 11 I would lose 95% oof my games and as my highest elo obviously will never decrease I would have to get new account to have fair matchups again.
Btw that’s what a smurf account is. What you mean is elo just not representing skill properly. Can happen sometimes, I guess we all have good and bad days. I managed to lose almost 100 elo today so am I smurf now?
Maybe there are some people who do that on purpose but that’s just bad for luck today then. There won’t be many people doing this.
what on earth does this even mean? do you even know what elo is? its not a ranking, it would serve zero purpose
if i were to continuously have to face players at 150+ my elo, just because i happened to go on a winning streak with an imba civ, so the rest of my games need to be matched at some ridiculous elo??
and if im on a losing streak (people have bad days, they change the way they play etc) they will continuously face the same elo opponents? makes zero sense
I think some of you guys misunderstood my original point. You wouldn’t be matched with players whose current ELO is the same as your highest ELO. You would only play players with the same highest ELO as you. This would prevent smurfs from intentionally quitting 20 games in a row and then stomping noobs. You might think this isn’t really a problem, but the eco-system of new players is important for the overall health and popularity of the game. If new players get matched with people who are pretending to be noobs, it’s simply not enjoyable for them and they probably won’t stick around. If their interest wanes, then the overall interest in aoe2 events and streamers could suffer. The aoe2 community needs new players for it to succeed long term, just like poker needs recreational players. Obviously it’s hard to imagine that now that the game is doing well, but there were times in the past when it was underground and only the die hards were playing, maybe about 10 years ago. And it’s not impossible that it could happen again, especially if fairness issues in the game like these are not addressed.
Actually, no, that’s not true. Your highest ELO is a score based on an accumulation of wins. So your highest ELO does mean something. It’s the peak of your ability thus far. Your highest share / stock price. Equally skilled players should have pretty similar highest ELO stats. And it’s a much fairer way of determining matchmaking than your current ELO, which is so open to abuse by smurfs.
That’s not a losing streak. Didn’t lose more than 4 games in a row. I’m talking 15, sometimes 20+ games, all losses, with not a single win during the lose streak.
elo is designed to be relative. Single elo number NEVER means to represent skills. It only represents the win rate between players within same system. Please read the original academic paper for elo.
If highest elo is used, that means the opponent I get matched may not be of same skill of me: I can get matched to much weaker opponent (if my highest elo is current elo), or much stronger. Do you seriously think it is a good idea?
Yeah obviously. Still the system would be terrible for a bunch of reasons. Firstly, you assume that the difference between current elo and highest elo is somewhat constant. Maybe that applies to top 100 or so but certainly not for the majority of players. People have different habits of playing, some are playing regularly while others may play a lot over a couple of days or weeks and then have some break in between. You’d expect the first category to have more stable elo than the latter. Alread a big problem for your model. Then there is a lot of other reasons why elo varies. Take me for instance. My highest elo is around 200+ points above my current one. Do I play worse than 2 or 3 months ago? I don’t think so (maybe a tiny bit since I play less than at that time but should be pretty minimal effect). The actual reason for that is at that point I had 3 days where I somehow managed to virtually exclusively play arena (and I usually have very high winrate there) so my elo got inflated hugely. So in your system I would basically any match that isn’t on arena or maybe hideout. And there will be the other extreme: People without performance difference between maps (or people that play arabia 95% of the time) and those whose current elo is super close to their highest one. Also I bet there is more reasons why elo difference isn’t constant like people grinding the ladder vs trying out whacky strats on so on.
Second argument is elo would get inflated constantly. Since highest elo basically replaces elo. For the points given have to based on highest elo difference and not based on elo difference. So your new score (highest elo) can only increase and not decrease (like your regular elo) and everything get inflated over time. One more reasons why it’s bad for people that didn’t play for some time because elo shifted in the meantime (even if it’s in their favor this time). Unless points given are based on current elo. But that makes zero sense. You wanna give points to the score that lead up to the match up in the first place.
ELO wasn’t designed to account for aoe2 smurfs. Live chess tournaments etc. don’t have such a thing as smurfs. And chess is a single game on a standard board, always 8x8 black and white squares. Not multiple different games with multiple players on different maps. So the academic paper might be somewhat useful but it can’t cover all the specific situations relating to ELO in aoe2.
That’s true, but you could counter that with a system where you slowly lose ELO points if you stop playing. And if you guys think my idea is that terrible, how would you solve the issue of smurfs?
ELO wasn’t designed to account for aoe2 smurfs. Live chess tournaments etc. don’t have such a thing as smurfs. And chess is a single game on a standard board, always 8x8 black and white squares. Not multiple different games with multiple players on different maps. So the academic paper might be somewhat useful but it can’t cover all the specific situations relating to ELO in aoe2.
We agree that it is far from perfect at AOE2 but it cannot be simply understood as raw skill level, that is much more wrong than anything else.
Smurfing is not that serious in 1v1 to be solved with extreme measures, possibly because I am beyond beginner elo (1850). Also it remains unsolved in all other games like LOL, SC2 (to less extent) etc.
Well this actually would make it even worse. Because if you don’t play and everyone else does their elo will get higher due to general massive inflation so relatively speaking you’ll be stronger compared to people with same elo. You could argue that equals out what I mentioned before but there is so many factors and the matchmaking will be quite likely a total mess while top players’ elo will get to 10k sooner or later.
You are talking about a very specific phanomen that I wouldnt even call smurfing. Because your solution wouldnt even address how most people do it (that is creating second accounts). And in the end the only solution imo is to have players’ reporting and people being paid to investigate these or generally monitor stuff. Anything else will likely not work or be bad for other reasons. Or you could use some AI software I doubt MS will put that much effort in it.
What percentage of players are above 1600, let alone 1850? The fact that’s it’s not really a problem at high ELO doesn’t mean it doesn’t need fixing. I’d be interested to see the ELO figures. I’d imagine it’s something like 90% of active players are below 1600, possibly more. So smurfing is a problem for the vast majority of the player base.
