Developer Q&A - Grand Arena Championship Changes - 12/02

Replies

  • Drathuk916
    624 posts Member
    edited December 2021
    mzftsttah2zl.jpeg

    So I didn’t do the math someone else did and they give themselves credit on the image. I understand the resistance to the idea. Your assumption that 8-4 wouldn’t keep you in a division may be correct @Ragnarok_COTF Personally, I’m not sold on the fact it would get you promoted, but can’t say for a certainty it wouldn’t either.

    The idea wouldn’t be to ping pong between the two divisions. It would be to do as well as you could without getting promoted. Now we know each week 12.5 percent of a division goes 3-0 and 12.5 percent of the division above goes 0-3. We also know that cg seems to have stated that only 10% of a league will be in division 1 and 25% in division 2. If we assume 100k per league than only 1250 people will go 0-3 in division 1 while 3125 will go 3-0 in division two. It’s hard for me to imagine that consistently going 2-1 will get you promoted due to the backlog of 3-0 accounts remaining in kyber 2. I actually think you can get the occasional 3-0 and avoid promotion.

    Would it take planning and studying the rankings? Absolutely. Would it involve some guessing on your part? Also absolutely. However multiple seasons of intentionally not doing your best to avoid rising to your 6-6 division should result in more crystals.

    This obviously goes out the window if they are allowing the divisions to equalize or if the division you are trying to avoid promoting into has more players than the one you are trying to stay in.

    Finally, the gain is minimal if you can’t occasionally go 9-3 or better but while only 500 or so crystals at 8-4 it is also 2 zeta scraps more for the extra wins. You’ll also might enjoy gac more because you’ll be winning more and when you do lose some will be intentional losses and less often because you ran into an account you can’t beat.

    I’m currently not pursuing this strategy btw but I do think it is a viable one.
  • Drathuk916 wrote: »
    However multiple seasons of intentionally not doing your best to avoid rising to your 6-6 division should result in more crystals.

    How? You have to stay 50-50 in the long run or you move up. So what is this long term strategy, specifically, that does this?
  • Drathuk916 wrote: »
    However multiple seasons of intentionally not doing your best to avoid rising to your 6-6 division should result in more crystals.

    How? You have to stay 50-50 in the long run or you move up. So what is this long term strategy, specifically, that does this?

    Please reread the portion where I state that it seems division 1 has fewer players than division 2 according to both the q and a and the info tab in game
  • Drathuk916 wrote: »
    Drathuk916 wrote: »
    However multiple seasons of intentionally not doing your best to avoid rising to your 6-6 division should result in more crystals.

    How? You have to stay 50-50 in the long run or you move up. So what is this long term strategy, specifically, that does this?

    Please reread the portion where I state that it seems division 1 has fewer players than division 2 according to both the q and a and the info tab in game

    I see what you're saying, that some 3-0 folks won't move up.

    Issue remains that your SR will gradually increase if you win more than you lose. I suppose if you tossed matches vs higher SR opponents, you could stay at an equivalent SR with a better than 50% win rate. How much better? We don't know. As such, we don't know how to manipulate our SR enough to increase crystal income.

    Bottom line, a claim has been made. At best, currently, said claim is a conjecture.
  • Drathuk916
    624 posts Member
    edited December 2021
    Of course it’s just conjecture. Yes your rating would be rising, but after an initial dropping because of the arbitrary gp rating gaps placed by cg the rating required to get into kyber will rise due to the difference in league sizes. If it is indeed a 10 percent division 1 and 25 percent division 2 then once that initial gap has been bridged by the consistent 3-0 2-1 players of kyber 2 the kyber 1 threshold rating will start to rise again. This just has to happen given that there will be 7500 more accounts gaining rating in division 2 than losing rating in division 1 each week if league size is around 100k as has been suggested on Reddit.

    Now eventually things will stabilize enough that going 1-2 in kyber one will get you demoted because the rating difference between kyber 1 and 2 is so small that all 3-0 accounts will promote plus some 2-1 accounts in kyber 2. I have no idea how long this will take obviously, but it isn’t until this point that the strategy would consistently fail and place you back into a ping pong situation imo.

    Edit: I’m obviously ignoring the upper end of kyber 1 that got a rating bump and typically have 5/6 gls plus executor.
  • Drathuk916 wrote: »
    Of course it’s just conjecture. Yes your rating would be rising, but after an initial dropping because of the arbitrary gp rating gaps placed by cg the rating required to get into kyber will rise due to the difference in league sizes. If it is indeed a 10 percent division 1 and 25 percent division 2 then once that initial gap has been bridged by the consistent 3-0 2-1 players of kyber 2 the kyber 1 threshold rating will start to rise again. This just has to happen given that there will be 7500 more accounts gaining rating in division 2 than losing rating in division 1 each week if league size is around 100k as has been suggested on Reddit.

    Now eventually things will stabilize enough that going 1-2 in kyber one will get you demoted because the rating difference between kyber 1 and 2 is so small that all 3-0 accounts will promote plus some 2-1 accounts in kyber 2. I have no idea how long this will take obviously, but it isn’t until this point that the strategy would consistently fail and place you back into a ping pong situation imo.

    Edit: I’m obviously ignoring the upper end of kyber 1 that got a rating bump and typically have 5/6 gls plus executor.

    Your claim that the K1 barrier will inflate is interesting. I understand your logic. You may be right. I am not sure, though, because SR doesn't increase the same for all wins. A 4000 SR beating a 3500 SR player will result in a smaller swing than a 3700 beating a 3600. The upsets from new K1 players beating "inflated" K1 players will rapidly coarse correct the SR values as well.

    All that said, there is certainly more merit to your position than I originally gave credence to. We shall see.

    Even with my low opinion of CG, they did say they got the algorithm from EA (whom I also loathe, but for different reasons... and some of the same), so I'm inclined to believe it's been field-tested.
  • It doesn't look like they are strictly keeping to the 10% in div 1 etc. If you look at the GAC main screen it has a set score range for your current division. That doesn't seem like it will definitely keep the same percentage of players in each division all the time.

    If you go 0-3 and your score drops under the threshold, you drop. So if there is a larger population in div 2 than div 1, then surely more players move up than move down, changing the population spread. Likewise div 5 is only 10%, and div 4 is larger, so more players drop into div 5 than move up to div 4.
    Div 3 is larger again, so they would have more players moving both up and down than were coming in from div 2 and 4. I don't really get how this can work as a stable population in each division. Doesn't this mean that all players are migrating towards the outer divisions? I might need some maths help here.

    Need to keep an eye on that skill range number and see if it changes over time. Here is a screen grab for later reference.
    m7xsbedn0svn.jpg
  • Drathuk916 wrote: »
    Of course it’s just conjecture. Yes your rating would be rising, but after an initial dropping because of the arbitrary gp rating gaps placed by cg the rating required to get into kyber will rise due to the difference in league sizes. If it is indeed a 10 percent division 1 and 25 percent division 2 then once that initial gap has been bridged by the consistent 3-0 2-1 players of kyber 2 the kyber 1 threshold rating will start to rise again. This just has to happen given that there will be 7500 more accounts gaining rating in division 2 than losing rating in division 1 each week if league size is around 100k as has been suggested on Reddit.

    Now eventually things will stabilize enough that going 1-2 in kyber one will get you demoted because the rating difference between kyber 1 and 2 is so small that all 3-0 accounts will promote plus some 2-1 accounts in kyber 2. I have no idea how long this will take obviously, but it isn’t until this point that the strategy would consistently fail and place you back into a ping pong situation imo.

    Edit: I’m obviously ignoring the upper end of kyber 1 that got a rating bump and typically have 5/6 gls plus executor.

    Your claim that the K1 barrier will inflate is interesting. I understand your logic. You may be right. I am not sure, though, because SR doesn't increase the same for all wins. A 4000 SR beating a 3500 SR player will result in a smaller swing than a 3700 beating a 3600. The upsets from new K1 players beating "inflated" K1 players will rapidly coarse correct the SR values as well.

    All that said, there is certainly more merit to your position than I originally gave credence to. We shall see.

    Even with my low opinion of CG, they did say they got the algorithm from EA (whom I also loathe, but for different reasons... and some of the same), so I'm inclined to believe it's been field-tested.

    Well that obviously depends on mm. My understanding of what cg has been selling is that similar ratings will be paired. Therefore large mismatches by skill rating should not happen. This is obviously not the case for roster size and construction.

    I’m also not sold on the idea yet that there is a modifier for playing an opponent with a different skill rating. I point to the fact that I lost 42 points in round one, gained 41 in round 2 (makes sense fought an opponent with lower skill rating and had a lower skill rating), then gained only 39 points for the third match. In that third match I only had one fewer skill rating than match one and faced an opponent with one more skill rating than match one but the match caused the least amount of change for both myself and my opponent.
  • It doesn't look like they are strictly keeping to the 10% in div 1 etc. If you look at the GAC main screen it has a set score range for your current division. That doesn't seem like it will definitely keep the same percentage of players in each division all the time.

    If you go 0-3 and your score drops under the threshold, you drop. So if there is a larger population in div 2 than div 1, then surely more players move up than move down, changing the population spread. Likewise div 5 is only 10%, and div 4 is larger, so more players drop into div 5 than move up to div 4.
    Div 3 is larger again, so they would have more players moving both up and down than were coming in from div 2 and 4. I don't really get how this can work as a stable population in each division. Doesn't this mean that all players are migrating towards the outer divisions? I might need some maths help here.

    Need to keep an eye on that skill range number and see if it changes over time. Here is a screen grab for later reference.
    m7xsbedn0svn.jpg

    Those skill ranges are not accurate. There are already people with less than 3540 in kyber 1 for this week.

    This all brings me back to my original post. It really sucks that cg hasn’t made clear how the skill rating system works. How promotion/demotion works. The actual size of each division not a percentage. I don’t expect them to ever, but it hurts the players ability to strategize regardless of their goal. It’s harder to decide if a win will likely promote them. Will a loss demote me. Is it really only wins and losses that matter or do banners matter? Does margin of victory matter? Etc etc.
  • Drathuk916 wrote: »
    Is it really only wins and losses that matter or do banners matter? Does margin of victory matter?

    Yes, no, no.

  • When there are enough people throwing matches and dropping to a lower div, it would become harder to maintain a profitable win rate even in the lower div.
  • p8raffvbkxye.jpg

    Top of gac ranking foy kyber1 displayed by my phone.

    Seems completly bugged.
  • o8h4ow907dtc.jpg

    Allies ranking, no consistency between ranking with # and total of points.
  • Drathuk916
    624 posts Member
    edited December 2021
    When there are enough people throwing matches and dropping to a lower div, it would become harder to maintain a profitable win rate even in the lower div.

    Absolutely and as more and more people relic higher and higher percentages of the 110 characters required to make the 22 meta squads for gac the harder it will be to both get into kyber 1 and maintain kyber 1.

    From the beginning, I think both cg and the player base as a whole have not fully considered the ramifications. Cg has assumed more crystals for most and close to all is the only thing that matters. The player base has assumed basically the same.

    I don’t think anyone at cg has truly considered whether the new gac is actually an improvement on the old system from the actual playing experience. I’m not sold that it is in its entirety. I appreciate that a strong defense is no longer harmful because banners don’t appear to matter beyond needing to get more than your opponent. This doesn’t mean I’m positive I’ll enjoy the new gac more.
  • Has anyone gotten omicrons for the catina battles....never seen it happen.

    Heck even 1 out of 100 would nice....
  • Has anyone gotten omicrons for the catina battles....never seen it happen.

    Heck even 1 out of 100 would nice....

    I have gotten 3-4
    Why wasn't Cobb Vanth shards a reward for the Krayt Dragon raid? Why wasn't Endor Gear Luke shards a reward for the Speeder Bike raid?
  • Starslayer
    2413 posts Member
    edited December 2021
    When there are enough people throwing matches and dropping to a lower div, it would become harder to maintain a profitable win rate even in the lower div.

    I still don’t get this. To drop you need to lose more than you win. Awesome. Then, if you want to maintain your spot in a place where you get easier victories, you need to win as much as you lose, because if you 8-4 or something you’ll go back up where you didn’t want to be in the first place. So instead of 6-6ish at your designated spot, you will 6-6ish in a lower spot with less rewards. Yeepee.

    If someone can explain me the maths behind this strategy, i’m genuinely curious.
  • Starslayer wrote: »
    When there are enough people throwing matches and dropping to a lower div, it would become harder to maintain a profitable win rate even in the lower div.

    I still don’t get this. To drop you need to lose more than you win. Awesome. Then, if you want to maintain your spot in a place where you get easier victories, you need to win as much as you lose, because if you 8-4 or something you’ll go back up where you didn’t want to be in the first place. So instead of 6-6ish at your designated spot, you will 6-6ish in a lower spot with less rewards. Yeepee.

    If someone can explain me the maths behind this strategy, i’m genuinely curious.

    Yes, generally speaking this strategy is flawed.

    I still think that 6/6 will make you gravitate toward division 3 in all leagues - this claim is disputed, and we shall see it in a couple of months.

    I made the claim however, that a leaugejump is something that you might want to avoid in certain scenarios, because the number of required teams and many other factors may cause a signifacant perfomanceloss, resulting in demotion one month later. Based on my disputed presumption I proved that back-and-forthing between K5 and A1 rewards you less crystals (roughly 50/day loss) than manipulating your position, floating around A2-A1.

    Jumping to another league is only beneficial for you if you can stabilize in it - and based on the system it is safe to assume that there will be some players who cannot do that.
  • Starslayer
    2413 posts Member
    edited December 2021
    .
  • If you’re skilled enough to find a perfect pattern to know when to throw a game and consistently win the necessary ones to earn a few extra crystals by a potential corner case in this system then I think you’ve deserved those extra crystals. Most players won’t bother trying to do that and if they would occasionally win a few extra crystals by chance then so be it.
  • If you're not going to give any rewards for players who join and don't attack, can you set everyone who hasn't already joined to join automatically at the end of the sign up period?
  • Jakdnels wrote: »
    If you're not going to give any rewards for players who join and don't attack, can you set everyone who hasn't already joined to join automatically at the end of the sign up period?

    If you draw group of 8 from the entire playerbase and not just from the players willing to play, you have far more chances to face opponents who won’t engage at all. Easy crystals, but not really exciting gameplay. Plus, you’ll up your SR thanks to the luck of the draw (well, more than it is now) and not because you earned it, which may put you later on against players who are beyond your reach.
  • Starslayer wrote: »
    Jakdnels wrote: »
    If you're not going to give any rewards for players who join and don't attack, can you set everyone who hasn't already joined to join automatically at the end of the sign up period?

    If you draw group of 8 from the entire playerbase and not just from the players willing to play, you have far more chances to face opponents who won’t engage at all. Easy crystals, but not really exciting gameplay. Plus, you’ll up your SR thanks to the luck of the draw (well, more than it is now) and not because you earned it, which may put you later on against players who are beyond your reach.

    I'd rather any or all of those things happen than penalize people who intended to play but missed the sign up and are locked out for weeks and have their SR negatively impacted. And also no one could complain that they signed up but some mysterious bug didn't cause it to register.
  • What about having an "always sign me up" option? You could turn it off if you want to, but won't accidentally miss out otherwise. You could also not have it turned on if the control of your roster at lock is important to you.
  • Jakdnels wrote: »
    Starslayer wrote: »
    Jakdnels wrote: »
    If you're not going to give any rewards for players who join and don't attack, can you set everyone who hasn't already joined to join automatically at the end of the sign up period?

    If you draw group of 8 from the entire playerbase and not just from the players willing to play, you have far more chances to face opponents who won’t engage at all. Easy crystals, but not really exciting gameplay. Plus, you’ll up your SR thanks to the luck of the draw (well, more than it is now) and not because you earned it, which may put you later on against players who are beyond your reach.

    I'd rather any or all of those things happen than penalize people who intended to play but missed the sign up and are locked out for weeks and have their SR negatively impacted. And also no one could complain that they signed up but some mysterious bug didn't cause it to register.
    Locked out for weeks? Plural?


  • Jakdnels wrote: »
    Starslayer wrote: »
    Jakdnels wrote: »
    If you're not going to give any rewards for players who join and don't attack, can you set everyone who hasn't already joined to join automatically at the end of the sign up period?

    If you draw group of 8 from the entire playerbase and not just from the players willing to play, you have far more chances to face opponents who won’t engage at all. Easy crystals, but not really exciting gameplay. Plus, you’ll up your SR thanks to the luck of the draw (well, more than it is now) and not because you earned it, which may put you later on against players who are beyond your reach.

    I'd rather any or all of those things happen than penalize people who intended to play but missed the sign up and are locked out for weeks and have their SR negatively impacted. And also no one could complain that they signed up but some mysterious bug didn't cause it to register.
    Locked out for weeks? Plural?


    No need to be pedantic. You understand the point
  • TVF
    36519 posts Member
    Jakdnels wrote: »
    Jakdnels wrote: »
    Starslayer wrote: »
    Jakdnels wrote: »
    If you're not going to give any rewards for players who join and don't attack, can you set everyone who hasn't already joined to join automatically at the end of the sign up period?

    If you draw group of 8 from the entire playerbase and not just from the players willing to play, you have far more chances to face opponents who won’t engage at all. Easy crystals, but not really exciting gameplay. Plus, you’ll up your SR thanks to the luck of the draw (well, more than it is now) and not because you earned it, which may put you later on against players who are beyond your reach.

    I'd rather any or all of those things happen than penalize people who intended to play but missed the sign up and are locked out for weeks and have their SR negatively impacted. And also no one could complain that they signed up but some mysterious bug didn't cause it to register.
    Locked out for weeks? Plural?


    No need to be pedantic. You understand the point

    Words mean things, and the longer the internet has continued the worse it has gotten. A line must be drawn HERE.
    I need a new message here. https://discord.gg/AmStGTH
  • What about having an "always sign me up" option? You could turn it off if you want to, but won't accidentally miss out otherwise. You could also not have it turned on if the control of your roster at lock is important to you.

    Problem with this is I doubt many people who quit the game will bother to turn it off, and then bottom divisions get filled with massive GP auto defenses.
  • Jakdnels wrote: »
    Jakdnels wrote: »
    Starslayer wrote: »
    Jakdnels wrote: »
    If you're not going to give any rewards for players who join and don't attack, can you set everyone who hasn't already joined to join automatically at the end of the sign up period?

    If you draw group of 8 from the entire playerbase and not just from the players willing to play, you have far more chances to face opponents who won’t engage at all. Easy crystals, but not really exciting gameplay. Plus, you’ll up your SR thanks to the luck of the draw (well, more than it is now) and not because you earned it, which may put you later on against players who are beyond your reach.

    I'd rather any or all of those things happen than penalize people who intended to play but missed the sign up and are locked out for weeks and have their SR negatively impacted. And also no one could complain that they signed up but some mysterious bug didn't cause it to register.
    Locked out for weeks? Plural?


    No need to be pedantic. You understand the point

    Pointing the difference between locked out for 4 weeks and locked out for 1 week isn’t really nitpicking.
    And the impact of not joining on SR (lose 2 rounds) is less than the impact of not playing at all because you were not available (and so chose not to join) and lose 3 rounds.
  • I know the devs say it isn't an exploit, but it seems rather unfair to face someone with a GL, who knows you can't beat it and sticks it in front of their fleets alone, saving them two whole squads for offence and there's no punishment for it even though they are clearly exploiting a known screw up. Surely it messes up your whole skill rating algorithm? Do you guys not test anything these days? Will it even be fixed before the next event?
Sign In or Register to comment.