Forum Discussion

NicWester's avatar
6 years ago

Grand Arena matchmaking.

I'll start by saying that I love Grand Arena as a feature and hope you guys keep running it on a frequent basis. Two a month would probably be ideal, but I'm sure that's going to be tough to balance alongside Territory Wars and Territory Battles. This is a long way of saying that even though I'm going to criticize matchmaking below, I'm not approaching this from an "I hate everything, stop doing it all" point of view.

The problem with Grand Arena is competitiveness. Your current matchmaking algorithm is working fantastically in creating matchups with similar GPs, I've never been more than 1% above or below anyone that I've fought against. So congratulations on that! Making an algorithm that works that well on the first go isn't easy. But I think that basing matchups on raw, aggregate GP is a mistake. The problem is that not all GPs are created equal, which will lead to non-competitive pods (I don't want to keep using the same words, so just to define my terms ahead of time "pool" is for the groupings of every player within a reward tier, while pods are a grouping of 8 or 5 players that are competing against one another. In short, pods are drawn from a pool of players.) and that will have the negative result of player dissatisfaction in the long run. I'll give hard data below to explain what I mean by "broad" and "narrow" rosters, but for now just assume that both have roughly similar GPs, but that the broad player can't possibly beat a narrow player.

If a player with a broad roster consistently gets paired up against players with narrow rosters, they're going to eventually stop actively participating in Grand Arena events. If I have no hope of beating my opponent then I'm not even going to try, I'll just sign up and collect my last place rewards in a couple days with no effort put into the event--hey, maybe I'll match up against someone who also doesn't put defenses and I'll win the tiebreaker so I'll get better rewards! On the other hand, if you have a narrow roster and consistently get paired up against other players with narrow rosters you're going to have a competitive pod, which will be fun a couple times--but once you find out that there are other people out there who keep getting similar or better rewards than you because they're in non-competitive pods, you're going to feel slighted. Why are you busting your butt and working harder for less than someone who just lucked into free stuff because they had opponents who either didn't care or were pushovers.

So you need to create a matchmaking system that will create pods that are more competitive if this thing is going to survive. I think there's an easy way of doing that:
1) Have the algorithm check if fleets are going to be used at all in the event. If they are, use fleet power as part of matchmaking. If they aren't, remove fleet power from the matchmaking.

2) Figure out the maximum number of characters needed to field a full defense at each tier and multiply that by some number (let's say 2.5, I'll explain why 2.5 below, but for now just bear with me) and then have the algorithm only look at the GP of that number of characters on each player's roster.

3) If fleets are used, then do the same for fleets and add their GP to the GP in part 2.

This GP, drawn from parts 2 and 3, is your Refined Galactic Power (a term I'm making up to show that it's different from the current matchmaking, which uses raw aggregate galactic power). Create pods using this Refined GP. This should ensure that every pod is competitive--it won't mean no one ever tanks just to get the lowest reward with zero effort, there will always be people looking for the free ride, but for people who actually play the event there will always be hope that they stand a chance to win at least a couple matches.

Alright, so:
Why 2.5?
Above I said use 2.5 as the multiplier for determining the number of characters to use when calculating Refined GP. I used that number as a starting place because you need to use SOMETHING, right? I picked it because you need at least as many characters on offense as you do on defense, which would mean 2x, right? So if there are 6 defensive teams, you need 6 of your own to beat them. That means 12 teams total, see? But it's not realistic to expect all attacks to work on the first shot, and a lower multiplier would overly-reward broad rosters because it would mean excluding more of their characters from the calculation. So 2.5 gives a little more leeway. The actual multiplier should be rigorously researched and adjusted upward or downward regularly to keep things as competitive as possible. But, for the purposes of this post, let's just assume 2.5.

What's a Narrow or Broad roster?
Simply put, a narrow roster is a player who only puts gear on their best characters and leaves characters they're not actively using at as low power as possible. They probably still farm characters so that they can have a 7* g1, level 1 character to deploy in Territory Battle platoons, but otherwise they don't actually invest in a character unless they're actively using them for something--raids, legendary events, etc. This leads to a roster with a lot of very, very high GP characters and a lot of very, very low GP characters.

A broad roster is a player who puts gear on anyone who needs it when they need it. Maybe they're doing it because they're a collector who wants every character to be as good as they can be, or maybe they're doing it because they just don't like seeing unsightly green crosses and red numbers yelling at them saying there's something they can upgrade. But, for whatever reason, they have a roster with a much more even distribution of high and medium invested characters.

To be clear, neither of these is inherently "bad" or "wrong." Play the way you want to play. But, currently, a Narrow roster clearly beats a Broad roster, which makes having a broad roster "bad" and "wrong" for the time-being. Before now, you could argue that either was viable because a narrow roster would be good at what it does, but inflexible and bad at reacting to new stuff--got to farm Pao up from nothing if you want to make Chex Mix, for example--while a broad roster wouldn't be as good at specific things (Arena, for example, because while you've got a couple teams on par with a narrow roster, you're not constantly investing in them the same way) you'd have more to contribute in Territory Battles and Wars and be a little more flexible when the game changes to make certain characters necessary (BB-8 gets released and all the narrow folks are mad because now they have to farm First Order from hard nodes, but broad rosters are laughing all the way to the bank because they already did that, for example--a real thing that happened to me). But now there *is* a definite answer because in every other mode of the game it's a matter of preference, but here in Grand Arena you have that answer: Go narrow or go home.

Some Hard Data:
"Go narrow or go home" seems like a hyperbolic claim and you might think I'm just a sore loser (Nah, man, I'm a great loser--plenty of experience at it ;) ) or hopelessly defeatist because I said a broad roster will never beat a narrow roster. So let me give you the hard data from my current Grand Arena opponent. (My ally code is 843-944-478, in case a dev wants to take a look and verify this, or take a deeper look at things I didn't even think to look at)
My GP: 3,925,484
Opponent's GP: 3,957,501
Raw GP differential: 32,017 (So raw GPs are within 99.19% of one another)

My G12s: 22
My G11s: 24
His G12s: 57
His G11s: 37
Total "high gear" characters: 46 to 94

(6 defensive teams at my tier, multiplied by 2.5 as per above)
Power of my top 75 characters: 1,308,096
Power of his top 75 characters: 1,449,831
Refined GP differential: 141,735 (So refined GPs are within 90.22% of one another)

Not only does my opponent have double the number of high gear characters, he also has nearly 10% more GP than me where it counts. I say "where it counts" because I have a lot of g8 and g9 characters, but they're realistically not going to do a whole lot against his teams. So while on paper we have roughly equal GPs within 1% of one another, in reality he grossly outclasses me.

How does such a mismatch happen? Here's an interesting number set:
Power of my BOTTOM 75 characters: 834,411 (Average GP of the bottom end of my roster: 11,125)
Power of his BOTTOM 75 characters: 665,475 (Average GP of the bottom end of his roster: 8,873)

Interestingly, he has 13 characters with under 6k power. The point isn't to say that the bottom of my roster would beat up the bottom of his roster (although that's kind of an interesting idea for a future Grand Arena, only allowing you to use the weakest characters on your roster--oh MAN would narrow folks be mad at that!) but rather to point out that the bottom of his roster, the characters that will never see the light of day in a Grand Arena, are keeping his Raw GP artificially low, while the bottom of my roster is inflating my Raw GP. This is the source of mismatches--a disparity between total GP and useful GP. If we only counted the useful characters, I would be facing someone with nearly a full 10% more GP than myself.

So, please, when you inevitably look at the data for the first couple Grand Arenas and look at how to improve matchmaking (And I know you guys will... As I said, I really like this new mode and I'm not coming her to dump on it. You guys take your job seriously and put out a quality product. I know that the initial matchmaking was put together with an eye towards refining it in future iterations.) please consider a Refined Galactic Power over Raw Galactic Power. I think that'll create much more competitive matchups going forward which will lead to a very popular mode of play, and encourage people to invest resources into characters--and any time people are spending resources they're going to be more likely to eventually spend money.

Thanks for reading this far.
  • "jayjonbeach;c-1787917" wrote:
    What you seem to want, is exactly that, you want uneven match-ups so stronger rosters, can walk over weaker ones, so they DON"T have a chance, you don't want the weaker rosters to have more of a chance. And you think that is more fair than what the OP and I are suggesting?

    That is exactly what he and others who share his point of view want: easy wins.
  • "Rath_Tarr;c-1788057" wrote:
    "jayjonbeach;c-1787917" wrote:
    What you seem to want, is exactly that, you want uneven match-ups so stronger rosters, can walk over weaker ones, so they DON"T have a chance, you don't want the weaker rosters to have more of a chance. And you think that is more fair than what the OP and I are suggesting?

    That is exactly what he and others who share his point of view want: easy wins.


    How ironic....
  • "Waqui;c-1788209" wrote:
    "Rath_Tarr;c-1788057" wrote:
    "jayjonbeach;c-1787917" wrote:
    What you seem to want, is exactly that, you want uneven match-ups so stronger rosters, can walk over weaker ones, so they DON"T have a chance, you don't want the weaker rosters to have more of a chance. And you think that is more fair than what the OP and I are suggesting?

    That is exactly what he and others who share his point of view want: easy wins.


    How ironic....

    How so?
  • "dimi4a;c-1790348" wrote:
    @NicWester bravo sir, very well explained and argumented! It seems odd that matchmaking counts a player's Fleet GP as part of his total GP when there are no ship battles in GA.


    "Waqui;c-1787807" wrote:
    "Fürkész;c-1787769" wrote:
    If I remember well there were a dev. post if no fleet in GA, then only the the squad GP will be used in the matchmaking. Obviusly it is no live yet(avg. 200k less my squad GP than my opponents).
    Do we know when this will happen?

    (P.S.: My rooster is fleet heavy, but until now only lost two of my GA matches, so do not matters a lot)


    It was in the road ahead. We will just have to be patient.


  • "Waqui;c-1790408" wrote:
    "dimi4a;c-1790348" wrote:
    @NicWester bravo sir, very well explained and argumented! It seems odd that matchmaking counts a player's Fleet GP as part of his total GP when there are no ship battles in GA.


    "Waqui;c-1787807" wrote:
    "Fürkész;c-1787769" wrote:
    If I remember well there were a dev. post if no fleet in GA, then only the the squad GP will be used in the matchmaking. Obviusly it is no live yet(avg. 200k less my squad GP than my opponents).
    Do we know when this will happen?

    (P.S.: My rooster is fleet heavy, but until now only lost two of my GA matches, so do not matters a lot)


    It was in the road ahead. We will just have to be patient.




    I bet in a half carbanti in april (after 3 GA in march) you still need to say the same thing. (Patience)
  • "Fürkész;c-1790585" wrote:
    "Waqui;c-1790408" wrote:
    "dimi4a;c-1790348" wrote:
    @NicWester bravo sir, very well explained and argumented! It seems odd that matchmaking counts a player's Fleet GP as part of his total GP when there are no ship battles in GA.


    "Waqui;c-1787807" wrote:
    "Fürkész;c-1787769" wrote:
    If I remember well there were a dev. post if no fleet in GA, then only the the squad GP will be used in the matchmaking. Obviusly it is no live yet(avg. 200k less my squad GP than my opponents).
    Do we know when this will happen?

    (P.S.: My rooster is fleet heavy, but until now only lost two of my GA matches, so do not matters a lot)


    It was in the road ahead. We will just have to be patient.




    I bet in a half carbanti in april (after 3 GA in march) you still need to say the same thing. (Patience)


    I have no problem repeating it, if it's still true.
  • We have been debating GA matching on a thread in general forums...GA, matchmaking & amp & fluff. Where we were asked to provide screenshot of how CG encourage buffing your roster which we have found several DEV announcements w/ proof of that. And we also have another announcement where Dev's say no one should be disadvantage for leveling up past a point in any game mode. Which would be counter to "spirit of the game". If you could review the comments and provide some feedback on the current matchmaking would be greatly appreciated. It has a lot of feedback available on the current GA matchmaking.
  • After so many Grand Arenas... Counting TOTAL gp for a GA with no ships is just unfair and ridicolous. This gives advantage to players who ignore ships and makes the game frustrating for people who upgrade both characters and ships. How can I beat someone with 300.000 characters gp more than me? Please fix this as soon as possible :(
  • "TRanger;c-1791295" wrote:
    We have been debating GA matching on a thread in general forums...GA, matchmaking & amp & fluff. Where we were asked to provide screenshot of how CG encourage buffing your roster which we have found several DEV announcements w/ proof of that. And we also have another announcement where Dev's say no one should be disadvantage for leveling up past a point in any game mode. Which would be counter to "spirit of the game". If you could review the comments and provide some feedback on the current matchmaking would be greatly appreciated. It has a lot of feedback available on the current GA matchmaking.


    CG encouraged players to broaden their rosters and buff up their characters before the long gone tournaments, yes. They didn't advice to fluff up your roster by leveling characters, you wouldn't be using anyway.