Recent Topics

Ads

Ranked Season 2 - Results & Insights

The latest news from the battlefront.
Find out what the developers are working on.
Forum rules
Before posting on this forum, be sure to read the Terms of Use
Caduceus
Posts: 653

Re: Ranked Season 2 - Results & Insights

Post#21 » Wed Sep 15, 2021 6:24 am

Even though I've seen people say it wasn't a problem, I think stream sniping, queue stacking, queue dodging etc. was a serious issue.

Matchmaking has to be as random as possible to function properly, but in the last season it was very easy to gather outside information on one's next matches.

This basically created a system where weaker players were preyed on and relentlessly farmed. Names of these players were readily shared as soon as they logged on.

The Ranked discord server is a large part of what made this possible, and even though that server did good things for Ranked as well, I am not convinced it amounted to a net positive.

It is also the lack of seperation ("bracketing") between players of different skill levels which made it possible to fish for matches against easy opponents. This is why I've argued from the beginning that some sort of beginner bracket is needed to resolve at least the worst of the skill imbalances. 0-999 was simply too large of a bracket given the size of the population, and mixed players that could easily have reached 1k MMR (given more games played) with players who would never have a hope of even getting positive MMR.
"I watched a snail crawl along the edge of a straight razor. That's my dream; that's my nightmare. Crawling, slithering, along the edge of a straight razor... and surviving." - Colonel Walter E. Kurtz

Ads
Husti
Posts: 111

Re: Ranked Season 2 - Results & Insights

Post#22 » Wed Sep 15, 2021 11:49 am

yeah, with the low amount of people queuing it is very likely for a bunch of players queuing at the same time to get into the same scenario. That is a problem. I've seen a lot of scenarios with the same people and it's usually the very good and well geared players doing that making it even worse for the other side.

I have no idea how to fix that because if the scenario manager tries to take people more randomly from the queue it might result in having certain people to wait much longer than others. The best solution for that is probably to get more people into ranked and HIDE THE NUMBER of queuers. Only show if there is someone queued for an archtype, not how many. Maybe don't show even that information. You don't see that for normal scenarios as well.

User avatar
Acidic
Posts: 2045
Contact:

Re: Ranked Season 2 - Results & Insights

Post#23 » Wed Sep 15, 2021 12:09 pm

Not very convinced by the stats produced here, kind of only showing winning avoids the whole point of mmr which is the basis of ranked as I understood it.
When I ran the numbers a few weeks again in a different way I got obviously different results.
I ran median mmr of played more than 10 games

Rational was mmr is what ranked was about
Median as it clips out extreme good/bad numbers out of the result

I won’t post my numbers as it would be more interesting to see what ppl consider a reasonable measure before the numbers appear.

User avatar
Fenris78
Posts: 787

Re: Ranked Season 2 - Results & Insights

Post#24 » Wed Sep 15, 2021 12:34 pm

Husti wrote: Tue Sep 14, 2021 6:21 pm I think the key to a fun and lasting ranked scenario experience is to give the losing side also rewards, like it is for scenarios. Why punish the losing side with -MMR when they did a great job and fought like lions but the other side was just better?

If you relax the system and give the losing side also rewards then the environment would be less toxic and people would queue more because even if they lose they would get reward.

Id remove the MMR system to speed up queuing and do something like this:

For the winning side: everyone gets the same amount of badges/currencies
For the losing side: do a roll at the end like for public quests: best two get 3/4 of the reward of the winning side, 3rd 4th 1/2 of the reward, 5th 6th 1/4 of the reward the winners got. The more damage/heal/protected you did the more contribution gets added to your roll.

Everybody gets someting, but on the losing side you get less and it depends on your contribution.

That's how Id make it.
I and some other asked for a system like that multiple times (see all previous topics about ranked), but I know it could be hard to get right : you have many factors influencing the raw numbers on the match result, like aoe damage, some tanks doing naturally more protection than others, or less damage, etc.)

The "match quality" is hard to grasp from an automated system point of view, and I guess it's far from easy to code it right (not to mention it's most likely not the top priority atm).

There are multiple issues about ranked going on a regular basis, wich fails because of :

- 6v6 class balance discrepancies ; many classes/specs are simply not fitted for those kind of fights, and many mirrors are far from equal when it comes to ST burst (Sorc/BW, SW/SH, WE/WH) or survivability (shaman/AM, BO/SM), and some are even globally not fitted at all for dynamic fights (Magus/Engineer).

- Player repartition : 70% rdps on Order (3 ranged classes which are globaly less fitted for 6v6), and some good mdps ones (Slayers, WL, even WH with a good build), and lack of damage from tank (KotBS 2H is half of damage AND survivability of a Chosen for instance) saving SM, which also lack a ST punt to help securing kill.
While Destru got very good ST classes, most of them being mdps (mara, choppas, WE, DoK), supported by heavy ST bursty ones (sorc and SH) wich are totally fitting the 6v6 meta needed for ranked.

Basically player choice of characters are heavily shifting the balance to one side, and it looks like not many Order players want to bear the destru meatgrinder anymore ; the odds of a bad setup vs a good one being too high.

- Player base : 90% of players are either good players, but undergeared and/or not wanting to commit into stressful ranked scenarios (even for the sake of BiS gear), or simply casuals that dont care at all and only want to do RvR or regular scs.

- Rewards : as we already mentioned multiple times, rewards are good, BUT the 1k entry barrier + MMR loss is making all but incentivize participation, when most players never see their score going above zero ; they simply quit without a solid (even on the long term) hope to a foreseeable BiS ring.
The recent addition to weekly rewards is a first step, but need to accomodate a change over the MMR system too, otherwise for most of us I guess six months of ranked (at minima, with 10 to 20+ needed matches per week) is too far away (if you compare to Sentinel ring for example wich is about 7 weeks to obtain, without risk).

So a first step have been made to bring back interest into ranked again, now more need to be tweaked to make the MMR rewarding, and engaging the participation on a regular basic, accounting the small playerbase we only have.

To conclude, I can also suggest making two brackets : one "entry level" only allowing Conqueror/Vanquisher, another with Invader+ gear. The gear gap is simply too deep betweenVanquisher and Sovereign to allow any form of competitive matches between such geared teams.

Husti
Posts: 111

Re: Ranked Season 2 - Results & Insights

Post#25 » Wed Sep 15, 2021 3:49 pm

Fenris78 wrote: Wed Sep 15, 2021 12:34 pm
Husti wrote: Tue Sep 14, 2021 6:21 pm I think the key to a fun and lasting ranked scenario experience is to give the losing side also rewards, like it is for scenarios. Why punish the losing side with -MMR when they did a great job and fought like lions but the other side was just better?

If you relax the system and give the losing side also rewards then the environment would be less toxic and people would queue more because even if they lose they would get reward.

Id remove the MMR system to speed up queuing and do something like this:

For the winning side: everyone gets the same amount of badges/currencies
For the losing side: do a roll at the end like for public quests: best two get 3/4 of the reward of the winning side, 3rd 4th 1/2 of the reward, 5th 6th 1/4 of the reward the winners got. The more damage/heal/protected you did the more contribution gets added to your roll.

Everybody gets someting, but on the losing side you get less and it depends on your contribution.

That's how Id make it.
I and some other asked for a system like that multiple times (see all previous topics about ranked), but I know it could be hard to get right : you have many factors influencing the raw numbers on the match result, like aoe damage, some tanks doing naturally more protection than others, or less damage, etc.)

The "match quality" is hard to grasp from an automated system point of view, and I guess it's far from easy to code it right (not to mention it's most likely not the top priority atm).
Yeah you are probably right. There are situations where for example the healers don't have much to do because the dps are stomping. In this case they would get not contribution.

It would be easier to give the losing side just 1/2 or 1/3 of the winners reward. Done. With the weekly event rewards after 10 completed scenarios and 10 wins (Id reduce that one to 5 because you need roughly 20-30 games for 10 wins, which is a lot) the amount of rewards should be fine and there should be enough initative to participate in that game mode.

User avatar
theoddone
Posts: 127

Re: Ranked Season 2 - Results & Insights

Post#26 » Wed Sep 15, 2021 5:47 pm

What does ranked need the most atm? People.

What do people need to be motivated to play? Rewards.

What happens when people can not get the rewards? They don't play.

So what do the pugs need atm in ranked? The hope of getting rewards. Without any hope of getting rewards, pugs won't play.

Increase the weekly reward to 10 insignias. That way people will at least have hope of getting the ring. People tried at the start of the season but realized it was hopeless so they stopped.
If people can get the ring in 5 weeks then they might play. You still have to win 10 matches each week and have a ton of emblems to buy the gear.

Other than that I think the devs have done a really good job with the interface and the MMR system is quite an improvement compared to before. Although MMR matching should be removed in the group ranked because the population isn't big enough.
Hopefully, there will be some balance changeups because it gets boring to play against double slayer 2/3 of the matches in soloq.

I think pre-season is the perfect time to test things.
-Theo

ashton007
Posts: 380

Re: Ranked Season 2 - Results & Insights

Post#27 » Wed Sep 15, 2021 6:18 pm

Sulorie wrote: Tue Sep 14, 2021 6:52 pm
ashton007 wrote: Tue Sep 14, 2021 6:25 pm Buff IB, thanks. Grudge building is so **** and too dependent on others-
It depends on your ability to swap oath friend.
So you agree, it’s too dependent on others. Why does my class mechanic require someone else? Makes 0 **** sense and IB is the only class that depends on others to build mechanic, meanwhile chosen/knight has effortless auras now. (Let’s not forget armor pots did a number on IB damage already) it’s a pain in the ass trying to armor debuff and punt among all my other tools-

Sulorie
Posts: 7219

Re: Ranked Season 2 - Results & Insights

Post#28 » Wed Sep 15, 2021 8:48 pm

ashton007 wrote: Wed Sep 15, 2021 6:18 pm
Sulorie wrote: Tue Sep 14, 2021 6:52 pm
ashton007 wrote: Tue Sep 14, 2021 6:25 pm Buff IB, thanks. Grudge building is so **** and too dependent on others-
It depends on your ability to swap oath friend.
So you agree, it’s too dependent on others. Why does my class mechanic require someone else? Makes 0 **** sense and IB is the only class that depends on others to build mechanic, meanwhile chosen/knight has effortless auras now. (Let’s not forget armor pots did a number on IB damage already) it’s a pain in the ass trying to armor debuff and punt among all my other tools-
When your own team gets no dmg, why would you need armor debuff against an inferior enemy? IB hardly loses grudge for his buffs, to counter this you need more time to rise it. BG gets it faster but spends more. Bringing other tanks into the discussion doesn't help.
As knight/chosen play is too passive, think about an engaging aura mechanic.
Dying is no option.

Ads
doky
Posts: 112

Re: Ranked Season 2 - Results & Insights

Post#29 » Wed Sep 15, 2021 10:46 pm

theoddone wrote: Wed Sep 15, 2021 5:47 pm What does ranked need the most atm? People.

What do people need to be motivated to play? Rewards.

What happens when people can not get the rewards? They don't play.
this dude is a genius!

Bloodmasked
Posts: 200

Re: Ranked Season 2 - Results & Insights

Post#30 » Thu Sep 16, 2021 7:09 am

GamesBond wrote: Mon Sep 13, 2021 10:54 pm
Husti wrote: Mon Sep 13, 2021 9:26 pm Sorc highest win rate, BW lowest :roll:

I hope you don't take these percentages for balancing. Some WEs did very well in there, yet there overall win rate is low. SW couldn't be playing in range because it was that bad. They were all assault. DPS RPs were not allowed to queue because they were that bad. Order got lots of wins because of slayers. And so forth.
8 sorcs and 6 BWs - I doubt that's even relevant for balance. We also do not balance the whole game based on ranked, ranked is not and will never be the main game mode of Return of Reckoning.
there have been several shield dok nerfs specifically due to ranked "balance" changes.

Who is online

Users browsing this forum: No registered users and 21 guests