Who’s in favor of lowering the starting Elo of new accounts? I can’t imagine it would require very skilled players more than a handful of games to get beyond, lousy players would cease to spoil the decent games of half-lousy players (like me) , and it might discourage “smurf” accounts because it would increase the time investment of their despicable treachery, and possibly make them feel foolish. I think this would also increase the starting enjoyment of new players.
For me its too late but when i First started playing online and losing most of the Matches it was discouraging to me and it still is to other new players now.
What is the starting ELO Atm? I feel like it should be 1000.
It is 1000 right now, after a fashion. You play some placement matches first, which have an increased effect on your elo. Only after that you get to see an elo, which for actual new players should be lower than 1000, assuming that the opponents on several of their placement matches didn’t drop or something.
The problem with lowering the starting elo without any other adjustments is that over time the average elo is just going to drop down to the new starting elo. That’s what the system is balanced around. That’s because play doesn’t generate elo points, it just redistributes it between players. The only source of fresh points is new accounts, so all accounts together (sort of roughly, I’m skipping over one or two effects) have the same average elo as a new account.
There are elo and elo-like systems that work around this. You could for instance start people at 500 elo but make it impossible to drop below 500. Beating someone at or close to 500 elo now generates new elo points, and the average is going to be significantly higher than the starting number. But every system has downsides, and you might soon start seeing elo inflation like on the old teamgames ladder, with the average and maximum scores just going up and up forever. This would be bad because it doesn’t give people a consistent idea of what being a certain elo means and because returning players would restart well below their weight class (although that wouldn’t be too bad, returning players could use a few warmup rounds).
In short: just lowering the starting elo is only temporarily going to change things. The current 1000 is fine.
I dont know how its calculated on lichess ( Seconds biggest Chess Platform, after Chessdotcom) over there you have a question Mark behind your Elo when you’re new and lose significantly more Points until you are at your “real” ELO, so you dont have to lose as many Games If your “real” ELO is for example 500 as you would have to lose “normally” and i think thats a good way to Not discourage new players.
Note: normal Players dont get more points If they Beat newbies, they get the regular amount, newer Players Just lose more Points.
That could maype Work for AoEII aswell but depending on how much/less Work it IS IT should have the highest priority
Edit: couldnt write Chessdotcom as you would normally
But then everyone just congregates at a lower starting elo. You aren’t fixing any problems, just overcomplicating stuff, and moving the average elo down 200 points. Doesn’t make a meaningful positive change.
On the flip side of that, couldn’t you just cap off the highest ELO to prevent the scores from just going up and up? e.g. 3000 max. If you hit that ceiling then you don’t get any more points, the other player just loses points.
That’s a very technical answer, so thanks. That’s kind of what I was looking for. But I’m not sure I understood it. (Math is not my strong suit.) Wouldn’t shifting the starting Elo toward the lower-middle range make for a more competitive game experience for the lower ranges, while hardly affecting the upper-ranges?
What he says is that if we lower the starting elo, the people who now have about 1k elo will over time lose elo and get closer to that new starting elo.
It’s how the elo system works.
Maybe there is potential for a different system.
But you could also just make a tweak to the matchmaking system and make the newer players more often face players with lower elo. So if someone starts new he will be matched with people with about 800-900 elo instead of the 1k elo.
I’m less concerned with higher levels losing elo (because they’ll lose in comparison to higher elo players). What bothers me is the idea that a nine year old kid, who doesn’t know anything costs me a game because he can’t even signal. That is echelons apart from what I think is now the bottom elo in the game. It’s very annoying, because I have arthritis and my eyes are failing, yet I get paired with these players who constantly drag me back into the doldrums. I UNDERSTAND this is a team issue. I wouldn’t have this problem 1V1, but I LIKE team games. I like the chaos. I just wish there were fewer tire-flats (i.e. better matching.)
How fast the average would shift depends on how stable the player base is. If there are a lot of new people entering ranked (especially if there are also a lot of players leaving) the average will shift down quickly, if all the same players stay with the game with few people entering or leaving the average will only drop slowly. But it would move eventually.
My first instinct is to guess this wouldn’t work well because at the extreme high end elo numbers are more important than at the lower end. You don’t want a big tournament for 100,000 dollars to be populated entirely by pseudo-random people who played until they hit the ceiling and then stopped playing until the selection date. Unless you want to start using a secondary “competitive elo” system for those purposes and make things super complicated this means you probably don’t want to have more than a handful of people bumping into the ceiling. And even with how many matches pro-players play, the 10 top pros not gaining more elo after a certain point are not going to be enough to offset all the new players creating more elo points by bumping into the floor.
Another alternative is higher elo players losing points over time. I know it was used on Voobly in the past, I don’t know if it’s in use on DE. That might work better in avoiding the above competitive problem.
However, that’s just my guess. One place where I know the “you can’t drop below starting elo” method is used is the Smogon pokémon battling community. It may seem a pretty random example, but because of how short their matches are (compared to say most RTSs or non-speed chess) their data sets of all matches played and elo movements are probably enormous and a pretty decent way to predict what a system like that would actually do to average elo scores over years of playing AoE2. So that could be one place the developers could look to if they were going to research a move like this…
I understand the concern, but you are going about this the wrong way. The correct way is to change the rate at which ELO changes. So, let’s say you lose 100 ELO for each loss till you get a win, then the rate goes down to 50, and eventually down to 10-15. This way, you get to 500 ELO just by loosing 5 games. If you think even that’s too much, make it 200 elo change at the start, then 100, and so on.
This rate of modification is what needs to be optimized for this to work.
It actually took me a while also. Especially with the TGs it’s indeed a problem when “noobs” start with the same elo as established players. As having an overrated newbie in your team will definetely have a negative impact in your whole team performance.
It’s not good for anybody including the newer player, cause it leads to a lot of toxicity and drives away newer players from playing TGs.
Maybe we need to restrict the Ranked TGs to only people who have a minimal Lobby TG experience to make sure it’s not that big of a difference then?
This would also reduce the amount of TG smurfs and throwers as it would need some time for these people to get “eligible” for ranked TGs.
But this is mostly only viable for 1v1s, TGs is more complicated. As there it is indeed the elo that determines the gains or losses of the team members.
Team games are fundamentally more complicated. You might be able to figure out someone’s 1v1 ELO pretty quickly with 20-30 games. You would need maybe 300-400 games to figure out someone’s 4v4 ELO. This is inherently different.
TG ELOs are also way more variable. Take 4 people with 1000 ELO in 1v1. They might have a lower TG elo if they don’t communicate and work against each other, while their TG elo might be 1500 if they have excellent communication and teamwork. Since people are getting into TGs with random people, this introduces a high degree of variance. I can’t think of a good solution for this issue.
Pretty sure ELO slowly drops if you stop playing, I might be wrong though. It should, if it doesn’t already. Personally I think it should also reset back to 1000 if you don’t open the game for 1 year. Another way to do it would be to set the total number of ELO points based on how many multiplayers are actively playing on the ladder within the last 6 months. E.g. Every player starts with 1000 ELO where the total number of possible ELO points in the game = Number of players on ranked x 1000 or whatever.