Can we get a 'not quite ARAM' mode?

iHerring··5 votes·465 views

As the title, I want some kind of NQRAM (Not Quite Random All Mid) I'm imagining something like

  • The game picks 5 random class types and each team gets 1 random champion that fits each class. (i.e. each team has 1 fighter, 1 tank, 1 supp etc)
  • The game picks 5 random class types for each team, each team gets 1 random champion that fits each class. (each team has 1 champion of each class, but the 2 teams don't have to have the same 5 classes)

Side note - there are currently 6 classes in the game. Marksman, tank, support, mage, fighter, assassin. Without thinking too hard about it, most champions do have a secondary class that could be used to refine things if needed.

I like the random nature of ARAM. It's a more casual and fun game mode and you often get to try out champions you might not have even looked at before.

At the same time it's kind of annoying when you get things like 5 supports and it turns out the enemy rolled 3 tanks and 2 adcs... I like random but I also like at least an illusion of possibly winning.

4 Comments

PIMPSMOOTH2 votes

ARAM itself should have an algorithm similar to this.

UtilityPunk2 votes

I'd say that it would be fair just rolling a 6 sided die 5 times. 1 = support, 2 = fighter, 3 = tank, 4 = mage, 5 = assassin, 6 = marksman. So for example, if 45221 was rolled, each team would be 2 fighters, 1 support, 1 assassin, 1 mage. Voila, both teams have the same classes, in the same amounts. And from time to time, you'll also get matches like 10 marksmen or 10 tanks.

If you want to add another layer to it, bias the dice. Ordinarily, there's just above a 16% chance of any role being chosen. One option I would recommend is progressively reduce the probability of rolling the same class by, say 10%. The chance of rolling the same class for the fifth time would be 1/6*(0.9^4)=~11%, or ~6.5% less likely. Meanwhile, the chance of any other class goes up by ~1.3%, bringing them each up to ~18%. This also has the side effect of making teams more likely to be meta, so I'd say pumping up the reduction to 15% or 20% would be even better. You'd still have the possibility of crazy teams that are balanced on both sides, but making them the occasional treat they are.

The two issues I see with your choices: Choosing 5 random class-types (but only one of each) removes the possibility of fun, non-meta teams. Choosing different class types for each team creates the possibility of a game skewed in favour of one team (the problem you are trying to solve).

Once the roles have been selected, you could also skew a champ's rate of selection by 25% if they are secondary in that role; e.g, selecting morgana as support less often than soraka.

iHerring2 votes

That's a fair point about only 1 of each class. Those solutions were just examples and I think you've made a good adjustment by having a bias against multiples, instead of a hard rule.

I guess my main point was as you said, it's possible right now to have vastly unbalanced teams where one group of 5 people might as well have not bothered to show up... in the "fun, casual" mode...

UtilityPunk1 votes

Absolutely! You had your heart in the right place and I understood that. I'm glad I could help you hone your idea, and who knows, maybe you'll think of an even better way ^_^

Keep coming up with good solutions to important issues. Really, you're one of the minority here that doesn't simply whine about the problem. You certainly put forward a good effort, and that's what matters. The more solutions you present, the more your solutions will improve (you'll google what other people know, critique your own solutions, and help others improve theirs).

It takes a while to get good at anything, but just the act of trying is very important!

More from Gameplay

More by iHerring