ELO Assistance, Please

Is ELO based off of, let’s say, the last ten matches or your entire career? Because if so, I hope my record is cleared as I’ve been preparing for a real competitive mode, and trying out new characters, and uh… it hasn’t been great. Lately, I’ve hit almost 95% win rate. But only lately.

No clue but I couldn’t resist…

1 Like

It’s a chess system.

However, I see no evidence that Gearbox is actually using an ELO system. If they are, they’re doing it wrong I think.

1 Like

A well-designed ELO system will favor recent performance. That being said, the idea of a non-1v1 ELO system is completely foreign to me. I don’t even know how you would begin to design something like that.

…Google time! (Google time for me. Not OP.)

Most team based games use it. League of Legends has had some success with it. Generally, your ELO is heavily based on your first games. After say, 50 games you have established your ELO and will receive less points (either for wins or losses). Most games will take into account large win sprees, and will basically reset your ELO benchmark.

The biggest gripe many people will say is why should a bad teammates effect your score/ELO. In the long run it balances out. If you play above the “Average” player, you should win more then you lose, as the other team has 5 possibilities of given a bad player, vs your team only having 4 possibilities.

Is this post directed at me? If so, I’m more interested in the specific functions themselves. I wanna know, for example, how they are “averaging” (for lack of a better word) a team’s ELO in order to calculate expected outcome (I’m specifically interested to see if it accounts for range of individual players’ ELO ratings on a single team, and how the change in each player’s ELO ranking is determined when comparing their individual ELO to their team’s and their opponent’s ELO after a match).

From what I can see a lot of games start with an ELO system, then change it after a few seasons, and all their scoring algorithms are proprietary (or at least they’re treated that way, so no public access).

Generally the entire team will lose/gain the same amount of ELO no matter their own personal ELO. Whatever team has the higher average is more favored and will lose more ELO if they lose, but gain less if they win and vice versa for the lower team.

The exact formula is almost never given out. Its proprietary (sp?) information, and something most companies spent a good deal of time/money on and dont want to just give it away.

I havent heard of any games that started with some form of ELO based matches and then changed it to something else. Not that i dont believe you, i just havent seen them. The only thing i can think of is Heroes of the Storm, but they still used ELO, just made some changes to it to make it harder to become top rank.

LoL, WoW, and HotS were three examples I ran across. Though, the “new systems” weren’t specified, so it’s possible all three just went to improved ELO systems.

I don’t even know why they wouldn’t base it on your average score in game.

You already have a system to separate the good players in the game. But hey team based game, solo based ranking system.

A common way for ELO to work in a game such as this would be as follows (using made up numbers).

Everyone starts with a score of 1000.

Killing another player nets you between 1 and 25 points, depending on the level difference and subtracts a similar amount of points from their score. So if you’re high ranking and get killed by a low ranking player, they gain a lot and you lose a lot while if you kill them the gain/loss is minimal.

You get 5 points for achieving goals in game. Such as destroying a sentry position.

You get 25 points for winning the match.

The problem is this game allows pre-made groups. So how do you match them against a group of individual players? There’s really no good method, even using their average score can be exploited.

Your score doesnt reflect how well you did that game. It doesnt take into account assists, objectives, healing/support, tanking, buildables, thralls, minions, and other things.

If you make it based off score everyone is going to go for kills and ignore everything else.

This creates a problem of farming other players to improve your score. Killing the same person over and over doesnt show that your a good player

It works out in the end. If you’re higher ranked, and thus able to farm them, you’re only getting one point per kill where they get twenty five for killing you.

What i meant was that people will start to drag out games so that they can earn more points, instead of playing and finishing the game.

Why end the game with 10 points, when you could farm them for another 15 minutes and have 30 points?

1 Like