The gap between an 1100 player and a 1200 player is WAY too big, before the ELO change this last summer, it’s comparing a 2300 ELO player to an old 1200 player.
So it’s really a ####### #### what kind of teammates you get, and usually if you’re ELO is 1100/1200, your teammates would be lower ELO than the opponents, and the skill difference would be significant.
I think either the starting team elo should be 700/800, or the algorithm should be modified again because right now it simply doesn’t work, unless if you’re a premade.
Starting elo will change nothing, it will just mean everyone loses some elo on the long run (because new players bring in less points); the numbers will change, the distribution will not.
This suggestion is posted by someone who dont understand the math behind an elo system. In an elo system you started as average player, which for this game is picked as 1000. Then you move up from there.
If you move the starting elo from 1000 to 800, then all ratings will on average drop by 200. A 1200 rated players will become 1000 and then the discussion is that there is a big gap between 900 elo and 1000 elo and the starting elo should be lower. As result the suggested solution is not a solution.
That being said: There might still be an issue. Let me explain:
These are the distributions of 1v1 and TG. The TG ladder is much more squished together. And with even more players. This pretty much means that 100 elo on 1v1 is much less then 100 elo difference in TGs. Added to that it are most likely 4v4 TGs. This makes makes the difference in average elo even bigger.
This is probably mainly due to TGs being more random in outcome. Mainly if playing solo. What i meant: In 1v1 it only depends on your own skill vs the skill of 1 enemy. In team games (i assume 4v4) it is your skill, with the skill of 3 allies and 4 enemies which determines the outcome of a game.
If a 700 elo 1v1 player starts playing 1v1, then he will drop to 700 elo in the first 10 games. But for team games this works different. Just image 4 new players getting match together with no experience. That is much more noise in converting to your true elo. And you will see this in action in the ratings. In the range of 900 to 1200 you will found noobs being carried by better allies, but also better players that have been holding back noobs. It is much harder to get out this area for anyone because of all this noise.
As result you still have 1100 elo players that are being carried a bit, but in reality are still terrible and 1200 rated players that just escape this middle ground. As result playing at 1200 rating is much more difficult then playing at 1100 rating.
Is there a solution to this problem? I dont really know. Maybe give more points for a longer period for new accounts (only for team games), so they can escape this middle ground more quickly? Nowadays you gain or lose about 50 points for you first 10 games before you convert back to around 16. Maybe change this to your first 25 games? Something like that will help escaping the middle ground more quickly for new accounts.
No, the reason for the lower spread in TG elo is that we had a reset in TG elo a few months ago. Before that, the spread in TG was a lot higher, with some people having like 3.3k (getting there by abusing smurf accounts, has to be noted).
Now the real issue with TG elo atm is that not everyone gains the same amount of points after a win. The weakest player in the team will gain the most (and lose the least in case of a defeat), the highest ranked player will gain the least. That means that if you are pretty high rated and therefor frequently are the highest rated player, you will not climb even with 60%+ winrate. This was introduced as a fix to the elo grinding using smurf accounts but its not a perfect fix.
Doesn’t work like that. Nobody loses elo in the long run, they just get less extra elo, which is exactly what needs to happen for the gap between elos to be more realistic.
I don’t know how it currently works, but it should be very easy to alter the Elo formula so that every game brings the total Elo one point closer to the desired total Elo “1000 x number of players”, in order to not disturbe the total Elo if the eevs decie to change the starting Elo.
We do not have to use the original Elo formula, we can change it as much as we want to fit the needs of the playerbase.
I feel the question is more about deciding who would get noob bashed.
