Grand Arena Algorithm simple solution
I’ve seen a lot of uproar over the matching algorithm put on here. A simple solution would be to let users set the difficulty to low medium or hard. You could scale the rewards to each tier and people would naturally select a tier they have a better chance at winning.
1
Replies
The easiest solution is to make matches based on the GP of a certain number of characters (20 most powerful, 30 most powerful, 40 most powerful, whatever) and then ignore the rest from the calculation. That way you aren't being paired against functionally stronger opponents for having worked on several teams or for having a large middle of your roster, nor are you being paired against functionally weaker opponents for having ignored the rest of your roster.
I'd even go as far as to say ignore fleets for the purposes of matchmaking entirely...
I agree with this, a cut off to not penalize the collectors. Idk how fleets will count towards it yet.
I don't think that will work tbh. The most powerfull toons are the ones with the most abilities maxed. My top40 isn't that much different from a random guildmate who has roughly the same amount of zetas equiped while the "roster strenght" for GA purposes is not rougly the same.
GP will always just be an indication of progression, not a reliable measurement of power/strenght/usefullness.
Regardless, any parameter used to base matchmaking on has it's own issues. Some will be beneficial for player A, others will be beneficial for player B. It's basically always gonna be an issue for atleast part of the playerbase.
No cause if you choose easy the rewards are scaled less. Like follow
Top tier rewards
Hard 200 cuffs
Medium 150 cuffs
Easy 90 cuffs
Middle tier rewards
Hard 100 cuffs
Medium 80 cuffs
Easy 80 cuffs
Bottom tier rewards
Hard 50 cuffs
Medium 60 cuffs
Easy 70 cuffs
So if you know you suck then you would take the 70 easy cuffs. If you are good you would try for the 200. Keep the same gp tiers and algorithms on top of this.
I am 2.8m GP and have to play 6 defence squads. I’ll also need at least 6 squads to fight my opponent with. So that’s a minimum of 12 squads I need; 12x5= 60 toons.
So matchmake me on the basis of the GP of my top 60 toons.
Easy way to get some close matches and keep both the collectors and the meta-chasers happy.
The long and the short of it is that if you allow people to pick their rewards, there's still no way to ensure an incredibly strong roster isn't matched with a mediocre one with roughly the same GP anyway. So it doesn't really solve the problem.
Like, if I knew my roster wasn't good enough to place in the middle of a tier, I'd just drop down to the one below it and stand a better chance of placing in the top of that tier. So my roster, which should be middle, will go up against people who should be lower. See?
This doesn’t sound simple at all.
I actually think OP’s suggestion could work, provided that the reward tiers were very carefully calibrated to reduce the incentive to play down. I’m not SURE it would work, because a lot would depend on mass player behavior, which could be hard to predict. Still, I think it’s foolish to reinvent the wheel when a fair pairing system for pairwise zero-sum games already exists and is well understood.
If I understand correctly, OP is essentially saying you get to choose your bracket: maybe better than easy, medium, hard, you could call the brackets “beginner,” “intermediate,” and “advanced.” If you say you’re a beginner, you get paired with other people who said they were beginners, and play for low rewards. If you say you’re advanced, you get paired with other people who said they were advanced, and play for higher rewards.
And further maybe limit each player to selecting one of the top 2 brackets they’re eligible for.
And maybe add more brackets, say a total of 8, so everybody is able to eventually settle into a level of competition where they have interesting matchups.
It’s kludgy, and it only approximates what you’d get in a true Elo system, and can probably be gamed if you are determined enough. But I’ve come to expect kludgy, gameable solutions to simple problems in this game, so maybe this is the best we can hope for.
So in order to fix un-equal matchups which lead to a specific type of player having a really hard time winning you'll change the system in such a way that the same specific type of player always plays for lower rewards?
I rather take my chances with the random opponent who has roughly the same amount of GP.
Yes exactly. I think that’s a feature, not a bug. If you’re locked into a matchup that you have a 5% chance of winning, the decisions you make have very little impact on your rewards. There’s a 95% chance you get the lower rewards anyway. If you have about a 50% chance of winning, then your strategic decisions can make a big impact on your rewards. Even if you’re playing for lower rewards, you’re having more fun because your performance in the competition is actually going to have an impact on your winnings.
That’s an odd preference, I think, but I suppose I believe you. To me, that’s like saying that an advanced club player should prefer to show up at a chess match and be paired with Magnus Carlsen because they’ve both been playing for roughly the same amount of time, and the rewards for the winner of that match are better than the rewards for the winner of the 1800-2000 Elo grouping. The entire assumption in most strategy games is that the players will prefer to be matched with competitors who are roughly as good at the game as they are. The best players get the best rewards, and play against the best competition. It’s hard for me to imagine that anybody would prefer for it to be otherwise.
In the games you’re talking about, there’s an objective measure of who is better. The problem here is that the people who are getting outmatched in GA *don’t believe* that their opponents are better at the game than they are. They think the opponent has played longer or spent money or focused differently or whatever reason there might be for why their roster is better suited to this game mode... but they fully believe that they should have an equal opportunity to win the same rewards as that person, because that person isn’t truly “better.”
Also, how is this different from the existing arena reward structure? Players with PVP-focused rosters rise to the top, fight better squads, and get better rewards. Players with collector-focused rosters fall to the bottom, fight weaker squads, and earn lower rewards. Is there something wrong with that?
Great. Now you just have to tell us all how initial matchups are determined...
Also need to define "performance."
Everybody is making it as hard as it needs to be, because it's not easy.
Maybe some do. That isn’t what I think, but it may be what some people think. I honestly haven’t seen many people complaining that PVP-focused players will get better rewards in this mode—the complaints seem to be about the competitiveness of the matches for the sake of competition itself. They’d prefer matchups where both sides have a fighting chance. I don’t think the reward structure affects that preference very much.
And for those of us who do badly in GA, but crave the better rewards, we can improve our rosters for GA, just like we can improve our rosters for arena. The decision about whether to invest resources in GA squads is affected by the reward structure (the better the rewards at higher levels, the more I’ll be incentivized to invest), but also by the quality of the matchmaking. The matchmaking needs to provide pairings that are at least competetive enough that marginal investments in my GA roster lead to noticeable improvements in my GA performance over time.
I don’t mind at all if the initial matchups are determined by total GP. Or you could just start everybody off at a rating of 1000, and see what happens. The Elo algorithm will eventually settle everybody in to a region where they’re competetive no matter what the initial ratings are. What I care about is that the matchmaking gets appreciably better over time. It doesn’t have to be perfect (or even very good at all) right now, as long as it is making progress towards reflecting our true performance.
If you want to get better rewards eventhough you suck, wouldn't you say you just want a "more fair matchup"? Ever noticed you see very few complaints about unfair matchups from players who won?
Here is the problem with that...
If you have 60 well geared toons, compare the GP of your top 60 toons to that of a kraken who has every toon in the game maxed. It will be roughly equivalent.
So someone with 60 strong toons will be matched against someone with 160+ strong toons. Your opponent will have way more options to use to take out whatever defense you set. They will have infinitely more options than you and will beat you nearly every time.
I’m not asking for a guarantee of a 50/50 chance of winning. That’s probably impossible. What I’m asking for are pairings that generally are competetive. True that the 95/5 example is out of thin air, but it’s by no means the “absolute worst in the current system.” Some of the mismatches people have reported would have been literally impossible for the weaker player to win, unless the better player just forgot to play. I still find it odd that you’re pegging your expectations to your “GP region.” GP is a really bad and arbitrary measure of competetiveness in this game mode. I thought we all agreed on that. So suppose I tell you that over time you will get equal rewards for either (1) winning 75% of your matches against players in your GP region, or (2) winning 50% of your matches against players in your true “competetiveness” region. Same distribution of rewards over time to everybody, but more competetive matches in which the players’ on-the-board choices, not the matchmaking, are decisive. Do you still truly prefer the less competetive system?
If somebody who sucks wants better rewards in GA, they should work on a GA roster that sucks less. Just like in arena.
Maybe not complaints, but I’ve definitely seen people on these boards and in my guild remark that they felt sorry for the other guy having to go up against them.
Could match based on the number of characters within the 3 highest gear tiers you have. (5-7, 10-12 for example) would be a more accurate count of your viable units against others at your level, and should produce closer matchups.
What do you consider my "true competetiveness region" though? equal roster, equal skill or both equal roster and skill?
There are plenty noobs with amazing rosters and vice versa.