How to fix GAC

AhnaldisGOAT
538 posts Member
edited June 24
Let me preface this by saying that this isn’t a thread to whine about how you dislike GAC matchmaking. We get it. This is also not a thread to debate how good or bad GAC matchmaking is. Spoiler alert, I think it is too simplistic for this complex of a game, which is why I’m posting this. This thread is about how to change GAC matchmaking for the better.
Now, here’s my system. All numbers are obviously not final, I’m just spitballing. We keep the “skill rating” archetype of win and move up the leaderboards for better rewards, but we don’t make it solely based on GAC W/L record. How? By factoring more things into the skill rating equation, and by adding a small change to the matchmaking system.
#1, GP. For example, for every 5,000 GP, you’d get…I don’t know, 1 skill rating point. If you lose GP by switching mods, you wouldn’t lose the skill rating you’ve earned, but it would simply increase how much GP you would need for the next point. This would help smaller rosters avoid getting matched up with rosters 3-4x larger.
#2, GL count. This one’s probably the least popular to implement, but I really would like this. Give something like 200-300 skill rating points for each GL in an opponent’s roster. That way, it would become more difficult (but not impossible) for people with 1-2 GLs to be matched up with people with both 0 GLs and 0 non-GL counters.
Now, for the system change. This one might anger some people. Sort people by time played in GAC along with skill rating. By “time played”, I mean the number of rounds where people earned at least 10 banners. So, if you attack 12 times, you get matched up with other people who attacked 12 times, if you didn’t attack once, you get matched with people who didn’t attack once, etc etc. This would help filter out the massive accounts who barely play GAC from dropping all the way to Bronzium/Carbonite, and beating up on weak rosters when the large ones randomly decide to play.
Thoughts? Feedback on numbers? Additions? Things to cut?

Replies

  • Antario
    661 posts Member
    Just ditch the datacrons and GAC is all good.
  • Lumiya
    342 posts Member
    Antario wrote: »
    Just ditch the datacrons and GAC is all good.

    Disagree here, it's not all good.
    We are all made of star-stuff
  • Let me preface this by saying that this isn’t a thread to whine about how you dislike GAC matchmaking. We get it. This is also not a thread to debate how good or bad GAC matchmaking is. Spoiler alert, I think it is too simplistic for this complex of a game, which is why I’m posting this. This thread is about how to change GAC matchmaking for the better.
    Now, here’s my system. All numbers are obviously not final, I’m just spitballing. We keep the “skill rating” archetype of win and move up the leaderboards for better rewards, but we don’t make it solely based on GAC W/L record. How? By factoring more things into the skill rating equation, and by adding a small change to the matchmaking system.
    #1, GP. For example, for every 5,000 GP, you’d get…I don’t know, 1 skill rating point. If you lose GP by switching mods, you wouldn’t lose the skill rating you’ve earned, but it would simply increase how much GP you would need for the next point. This would help smaller rosters avoid getting matched up with rosters 3-4x larger.
    #2, GL count. This one’s probably the least popular to implement, but I really would like this. Give something like 200-300 skill rating points for each GL in an opponent’s roster. That way, it would become more difficult (but not impossible) for people with 1-2 GLs to be matched up with people with both 0 GLs and 0 non-GL counters.
    Now, for the system change. This one might anger some people. Sort people by time played in GAC along with skill rating. By “time played”, I mean the number of rounds where people earned at least 10 banners. So, if you attack 12 times, you get matched up with other people who attacked 12 times, if you didn’t attack once, you get matched with people who didn’t attack once, etc etc. This would help filter out the massive accounts who barely play GAC from dropping all the way to Bronzium/Carbonite, and beating up on weak rosters when the large ones randomly decide to play.
    Thoughts? Feedback on numbers? Additions? Things to cut?

    It's a start...

    What data would it take to actually come up with an example league? Is that data available to us as players or those of us classed as mammals that live in the sea, or those that moderate?

    Without a doubt not everyone will agree, but surely it's worth knocking something together to demonstrate the theory???

    my Skill Rating would be c.1700 which seems low, 1SP/2500 GP maybe, which would be 3115, my actual skill rating is 2911

    Factor in Sort people by time played in GAC

    Would be good to see other factors that have potential.
    This is the Way
  • Jkane
    154 posts Member
    I like factoring GP back into the equation while not making it too significant. I also like the system change that takes into account "time played. I don't think it has to be an exact match between attacks. I feel a player who makes 10 attacks, for example could be matched up with opponents who made 8-12 without destroying the intent.

    Regarding GL's, I'm ambivalent. I have 3 and the Executor, so I've regularly had matches with people who had 1-5 GL's. Haven't won every match against those with less and I haven't lost every match against those with more, so it doesn't bother me personally. Also wouldn't bother me is such a change was implemented.
  • Nothingface
    620 posts Member
    edited June 24
    GLs have pretty established counters so I wouldn't take them into consideration but establishing parameters like +/-50% gp would probably help sub kyber matchmaking.
    Honestly, matchmaking in kyber isn't bad it just falls apart in lower divisions.
  • GLs have pretty established counters so I wouldn't take them into consideration but establishing parameters like +/-50% gp would probably help sub kyber matchmaking.
    Honestly, matchmaking in kyber isn't bad it just falls apart in lower divisions.

    My thought process with the GL thing is that a person with a GL is probably likely to have at least one non-GL counter, even if it’s soft. For example, someone with SEE probably can try the DR vs JML counter, someone with JML probably has the resources to use JKR/JKL vs SLKR, etc. I want to try to make sure that having +1 GL isn’t an auto-win like it sometimes is in the lower leagues because very few people have the specific non-GL counters.
  • harvestmouse
    391 posts Member
    edited June 24
    1. I don't want to go back to GP match making at all. It makes for negative building and suits newer players that have a better overall vision of the meta. I've played under GP match making and I didn't think it was better.

    2. I'd be all for having a GL factor. However, GLs are the money maker. There's no way that would be productive for CG. Also it would really skew divisions. If let's say you're the best player in the world without a GL, what division do you end up in? Kyber 1? Then who do you face? The players in K1 with the least GLs. Eventually the winner of the low GL count brackets ends up near the top.

    How I'd fix GAC would be:

    1. No move ups and downs mid week. Only at the end of the week.

    2. There's no difference between moving up divisions and leagues. You can move up from A1 to K5 just as you would A2 - A1. League differences at this point make no sense bar the title.

    3. No more squish. Divisions are transparent with a regulated amount moving up and down. And that amount is very clear. If you're in the relegation zone it's marked in red, if in the promotion zone it's marked in green.

    4. No more 3v3 and 5v5 under the same system. Either separate them or remove 3v3 altogether.

    5. A system that acknowledges 'sandbagging'. So if you're in a division lower than the recommended GP, you face players of a similar GP that are also in your division.

    6. No more 1 toon dominating a match. A player has the choice of removing 1 character from the opponent's defence. However, that player (the opponent) gets a max banner win (the player having the character removed).
  • tbh, the current system would be MILES BETTER if it implemented a few changes. My most favorite one is to weight the matchups. If I beat someone with 6 GLs, having only 3, I should get more points than I get for beating someone who has 2 or 3 GLs. It's not unlike alot of PVP systems in alot of games where beating a squad that's more powerful than yours and beating a squad that's alot weaker than yours do not earn you the same points. You get more points beating that stronger squad. You get less points beating a weaker opponent.

    Just changing THAT would really impact the lower divisions and the bullying going on down in them. It might not completely eliminate it but it would certainly do something CG likes- incentivize building your roster. It also wouldn't be a change that would enrage the whales up in K1 as it mostly wouldn't impact them.
  • I believe at present it simply looks at those with your skill points, in whichever league/division you are, and in blocks of 8 allocates the GAC groups… so looking at the standings (pre matchmaking) you could guess who would be in your group, as it’s up to +/- 7 positions either side…
    Incredibly simple, and that’s what we need to keep in mind, it’s gotta be simple, coz let’s be honest, CG isn’t gonna invest (and probably wouldn’t do a good job even if they tried) into overly complicate the matchmaking algorithm.
    So my suggestion is this, simply add a second ‘sorting’ criteria, other than skill rating, GP being the most obvious… so when groups are allocated everyone would have a very similar GP.
    It may perhaps alleviate those incredibly unfair circumstances where someone at 4M GP fights an opponent twice his GP.
    It’s still extremely basic, I know, but it’s gotta be if we want CG to consider a potential change… 1 more variable, nothing more, otherwise CG will never consider taking actions imo.
    But personally all of the ideas mentioned above have merit and would no doubt improve the status quo.
  • Why stop at counting the GLs? What about the number of omicrons? Legendaries? the number of lvl9 datacrons? the number of lvl6 datacrons?
  • @AhnaldisGOAT What is your objective with these changes? What result of the current system are you trying to remove? What new results are you hoping your changes bring about?

    Putting aside bruised egos and the varied experiences we've all had with the new and old systems, the issue with these discussions is that different folks have different ideas of what an ideal system would bring about. So I think you're putting the cart ahead of the horse here by discussing changes before agreeing on the desired outcome.
  • Why stop at counting the GLs? What about the number of omicrons? Legendaries? the number of lvl9 datacrons? the number of lvl6 datacrons?

    Cool, I won't add any omicrons or datacrons at all then. Most likely the less dedicated players won't be using them.
  • and that’s what we need to keep in mind, it’s gotta be simple, coz let’s be honest, CG isn’t gonna invest (and probably wouldn’t do a good job even if they tried) into overly complicate the matchmaking algorithm.

    The current demotion/promotion system is very complicated. I think it'd work better with a much simpler method.
  • and that’s what we need to keep in mind, it’s gotta be simple, coz let’s be honest, CG isn’t gonna invest (and probably wouldn’t do a good job even if they tried) into overly complicate the matchmaking algorithm.

    The current demotion/promotion system is very complicated. I think it'd work better with a much simpler method.

    It was fairly straightforward at the beginning, then CG introduced mid-rounds promotions/demotions, we learned about the squish… I mean, that side could and should have stayed simple and straightforward, but CG goes messing with the things that work rather than focus on what’s broken 😖

    I’d get rid of this silliness mid-rounds, it’s meaningless, adds no value whatsoever imo.
  • They need to ban starkiller on defence until the bugs are fixed
  • Rius
    161 posts Member
    edited June 25
    I don’t think matchmaking should be GP or GL based. When you unlock a new GL or team you want to experience a burst of improvement (a good season or two) and re-plateau. We should all experience this as we level up from time to time. The win loss should stabilise similar rosters and skill if players participate.

    I just think sandbagging should be penalised if you are significantly higher GP than the average in the division, your rewards should be handicapped to promote you to win and rise to a division more suited to your roster and play against your equals competitively rather than punch down on rosters with no effort.
Sign In or Register to comment.