|
they just dont want people fighting over vehicles as the omg i am gonna get so much rating from being in a terror. They really will have to do something about mercs though. Since you don't get damage stats from them even less people will bother doing them.
With all the stats gathering sites, it will not take long till somene sees a sub 50% winrate player in grandmaster and then more will start to copy this cancerous way of playing. And then we will have a Gazlow one trick invasion. And I am talking about robo gaz ! Oh the dark days ahead ! ;P
The best advice in this new system is, stop quick on losing streaks. Because you will fall deeeeep. Its faster to return, but it still hurts.
|
On November 28 2017 21:25 FeyFey wrote: they just dont want people fighting over vehicles as the omg i am gonna get so much rating from being in a terror. They really will have to do something about mercs though. Since you don't get damage stats from them even less people will bother doing them.
With all the stats gathering sites, it will not take long till somene sees a sub 50% winrate player in grandmaster and then more will start to copy this cancerous way of playing. And then we will have a Gazlow one trick invasion. And I am talking about robo gaz ! Oh the dark days ahead ! ;P
The best advice in this new system is, stop quick on losing streaks. Because you will fall deeeeep. Its faster to return, but it still hurts.
I am pretty sure there is a damage done to mercs statistics internally in blizzard that they are using.
|
Obviously nobody outside of Blizzard has any personal experience with the system yet, but I have read/watched all the interviews, Q&A's, and reddit questions I can, so this is my thoughts based on what I know so far:
- It's impossible to have a sub 50% win rate and advance based on the information shared so far. Winning is worth a minimum of double the performance adjustment, and I suspect most people will very rarely earn the maximum bonus or penalty.
- Remember that it's possible to get a negative modifier if the stats you are producing mirror those of a lower MMR player, so attempts to "abuse" the situation better earn you a win or your efforts might actually earn you a penalty rather than a bonus.
- They are measuring things that aren't on the scoreboard like capturing merc camps so that will be rewarded. There are over 20 stats being measured for every hero, though obviously some like CC time won't apply if your hero has no CC.
- Not knowing all the metrics and how much they are being weighed for each hero is part of what makes it hard to abuse. No single stat is going to guarantee you a good performance rating.
- My impression based on the Q&A's and interviews given so far is that there is probably a max contribution for each stat, so "abusing" an obvious stat like damage on an assassin has a point of diminishing returns once you have earned the maximum contribution for that stat.
- On top of that, Travis said that if the assassin hero you are playing has high damage numbers at all levels of play, that wouldn't be a weighted stat, but rather a stat that might give a negative adjustment if it wasn't high.
- Having caps on how much each stat can contribute to your overall performance is one way they could make the Garden Terror not matter as much. They could measure what a grand master MMR player on the hero you are playing does in siege and hero damage without the terror and set that or slightly above that as the level needed to earn max bonus for that stat on Garden of Terror. The terror would make it easier to hit that cap of course, but each stat only contributes a portion to your performance adjustment.
- Weighted stats are only worth ~10-20% more than unweighted stats.
- It's not going to be perfect and 100% accurately reward every single action you do in every game, which is why winning and losing matters much more to your advancement.
- It's likely that there will be flaws that need tweaks. If it works the way they claim it does, people cause their team to lose because they are trying to find a way to "abuse" the system is probably going to be a bigger problem than people getting unfair advancement because they found a way to abuse the system.
- Obviously if one or more things doesn't work the way they have described, then many of these points won't be accurate. Until we see it live we can only speculate based on the information given.
|
On November 28 2017 22:16 Hider wrote:Show nested quote +On November 28 2017 21:25 FeyFey wrote: they just dont want people fighting over vehicles as the omg i am gonna get so much rating from being in a terror. They really will have to do something about mercs though. Since you don't get damage stats from them even less people will bother doing them.
With all the stats gathering sites, it will not take long till somene sees a sub 50% winrate player in grandmaster and then more will start to copy this cancerous way of playing. And then we will have a Gazlow one trick invasion. And I am talking about robo gaz ! Oh the dark days ahead ! ;P
The best advice in this new system is, stop quick on losing streaks. Because you will fall deeeeep. Its faster to return, but it still hurts. I am pretty sure there is a damage done to mercs statistics internally in blizzard that they are using.
And I am pretty sure that people aiming only for shown stats will not care . As they are the ones that stopped doing mercs after they changed that. Before they were the ones rushing to mercs.
To be a bit more serious about the weighted stats. I am sure they learned alot from Overwatch in that regard. But I dont think they found a solution to the mercy situation. Though it will probably never be that bad. Why take Gust if Hinterlands will most likely help your rating more. Winrate might be a bit lower, but the individual rating will offset that. If talent choices change the weighting, people will find the best weighted talents and so on.
In the end that system will either bring normal players faster to where they should be. But it will introduce anomalies (behaviour not aimed at winning) and tilts will have more severe demoralizing effect. Or it won't matter, because its effect is to low compared to winning or losing and they just wasted their time implenting it.
The only positiv thing is, that they can introduce rewards if someone performs exceptional well :D.
|
Not knowing all the metrics and how much they are being weighed for each hero is part of what makes it hard to abuse. No single stat is going to guarantee you a good performance rating.
Not really when they openly admit they don't take into account Guardian of Terror. That's not even that complex so more complex stuff like getting vision of an area, winning time for your teammates by wasting multiple opponents time is 99.99% not something they are looking at.
There are decisions that marginally increases the teams chance of winning but puts makes your stats worse. There are tons of those decisions that you very easily can identify when playing the game and where it's very obvious blizzard isn't quantifying that.
Only looking at 20 numbers is very very little in such a complex game.
They could measure what a grand master MMR player on the hero you are playing does in siege and hero damage without the terror and set that or slightly above that as the level needed to earn max bonus for that stat on Garden of Terror. The terror would make it easier to hit that cap of course, but each stat only contributes a portion to your performance adjustment.
They could definitely do that. But judging by the interview, they aren't.
|
My 2 cents regarding support changes which seems to gather a lot of criticism: I think its a good idea on paper. One of the reasons I only play QM and not hero league is because I don't want to be forced into a role I don't enjoy. And I just don't find satisfaction in playing supports. I want to be making plays, and I think nerfing supports is a bit of a step in the right direction. Ideally you could play comps with only like a Tassadar or Tyrande as the only support player.
|
On November 28 2017 23:57 Hider wrote:Show nested quote +They could measure what a grand master MMR player on the hero you are playing does in siege and hero damage without the terror and set that or slightly above that as the level needed to earn max bonus for that stat on Garden of Terror. The terror would make it easier to hit that cap of course, but each stat only contributes a portion to your performance adjustment. They could definitely do that. But judging by the interview, they aren't.
I think they are doing that or something similar based on previous comments they have made. Travis has said that you can't focus on a single stat and hope to get a good performance adjustment and that weighted stats are only worth ~10-20% more than other stats.
My guess is that there is a maximum on how much each stat contributes to your over all performance adjustment. Once you have earned the max bonus from that stat it won't matter if you keep getting more with the terror or by focusing on that stat above anything else you could be doing. If you have already hit the max in that category then you would progress faster by doing something in an area you haven't maxed out. Of course even if I am right, we won't know what the max is for each stat, and it would be different on each hero, map and game length.
As for double supports and the upcoming support nerfs, there is a new article from Blizzard about the changes here.
|
nerfing supports lol. How about nerfing the overbuffed dps and tanks so that you can't run 2 supports. Or release more dps with support skills, like Tyrande. It would help if they stop streamlining supports to do the exact same thing with just other graphical effects. So that more people enjoy playing support.
But lessening the double punish of lower damage and healing when lower level would be something i would love to see.
|
On November 29 2017 08:32 FeyFey wrote: nerfing supports lol. How about nerfing the overbuffed dps and tanks so that you can't run 2 supports. Or release more dps with support skills, like Tyrande. It would help if they stop streamlining supports to do the exact same thing with just other graphical effects. So that more people enjoy playing support.
But lessening the double punish of lower damage and healing when lower level would be something i would love to see.
Waveclear + healing supports are pretty boring. Definitely wish for more damage/cc+ a bit of heal supports like Tyrande and make them viable as solo supports.
|
On November 29 2017 09:24 Hider wrote:Show nested quote +On November 29 2017 08:32 FeyFey wrote: nerfing supports lol. How about nerfing the overbuffed dps and tanks so that you can't run 2 supports. Or release more dps with support skills, like Tyrande. It would help if they stop streamlining supports to do the exact same thing with just other graphical effects. So that more people enjoy playing support.
But lessening the double punish of lower damage and healing when lower level would be something i would love to see. Waveclear + healing supports are pretty boring. Definitely wish for more damage/cc+ a bit of heal supports like Tyrande and make them viable as solo supports.
We need more stukovs tbh. He has an interesting heal, ccs (silence with E, strong slow with W), area denial (with E), strong melee attack (good vs squishy divers like tracer/genji, or vs mercs), an "emerald wind" (flailing swipe). I really hope they keep giving these kind of tools to supports or else no one will want to play support in hero league. Being a healbot is really not fun to most people (myself included).
|
As someone who been spamming a ton of Samuro illusion master recently, I am looking somewhat confused/slightly dissapointed with the new changes.
Reduce the cooldown of Advancing Strikes to 8 seconds Now increases the attack damage of Mirror Images by 100% Now also increases Samuro’s Basic Attack damage by 10%
So assuming they mean Image transmission cooldown instead of Advancing strikes, that's a nerf from 6 seconds to 8 seconds. I assume this means you no longer can recall with image back to nexus and then retransfer back to the image with full health. That's unfortunate because the low CD imo made the skillcap very high on the hero.
This feels like a small nerf to how impactful/annoying a high skilled Samuro player can be with the hero.
Further, since mirror image damage is completely irrelevant, why do they bother increasing damage from 9 to 18?
What I right now enjoy about Illusion master is that it made it feel more like a well adapted dota-hero as you can be everywhere on the map. You can take camps. You can push lanes you can attack heroes - as long as you are mechanically good enough. There is nothing else in HOTS that feels like Illusion Samuro and it is quite rewarding to play him well.
And since not a lot of people have complained about this in terms of balance, I don't see why they had to change this?
New Samuro can still too some extent do this, but I feel he now is more of 65% old Samuro + 35% Zeratul.
|
Samuro was terrible in 5man compositions/hero league. So i guess they try and make him viable there.
|
Samuro has the second highest win rate in hero league.
|
On November 29 2017 14:36 Essbee wrote:Show nested quote +On November 29 2017 09:24 Hider wrote:On November 29 2017 08:32 FeyFey wrote: nerfing supports lol. How about nerfing the overbuffed dps and tanks so that you can't run 2 supports. Or release more dps with support skills, like Tyrande. It would help if they stop streamlining supports to do the exact same thing with just other graphical effects. So that more people enjoy playing support.
But lessening the double punish of lower damage and healing when lower level would be something i would love to see. Waveclear + healing supports are pretty boring. Definitely wish for more damage/cc+ a bit of heal supports like Tyrande and make them viable as solo supports. We need more stukovs tbh. He has an interesting heal, ccs (silence with E, strong slow with W), area denial (with E), strong melee attack (good vs squishy divers like tracer/genji, or vs mercs), an "emerald wind" (flailing swipe). I really hope they keep giving these kind of tools to supports or else no one will want to play support in hero league. Being a healbot is really not fun to most people (myself included).
I was in a game the other day and for whatever reason drafted stukov instead of malf to combo with stitches. I don't think I had appreciated how responsive e+d is for a disabling slow--- it was almost as good as malf roots for following up on stitches hooks. I also like that KT-like kit design of having to choose between empowering a disable and empowering a heal sometimes.
Really bliz deserves a huge round of applause for stukov's design.
|
|
Regarding Garden Terror & other vehicles - It's pretty likely that for many heroes Siege Damage is a stat that's barely weighted. Furthermore, because it's a per-map basis, the valuation for siege damage may be reduced even further, because the system will detect there's a low correlation between "high siege damage" and "this character is winning", since there will be plenty of games where a hero wins without ever piloting. Remember, the system is looking for patterns of stats that lead to wins - when there's a lot of noise (siege damage on GoT), there won't be much importance placed.
|
I've found myself wondering why they would do the support changes two weeks before doing the rest of the changes. It feels like you should just do it all in one go rather than leaving us figuring out a meta for two weeks before it all turns on end again.
|
On November 30 2017 01:11 ChaosOS wrote: Regarding Garden Terror & other vehicles - It's pretty likely that for many heroes Siege Damage is a stat that's barely weighted. Furthermore, because it's a per-map basis, the valuation for siege damage may be reduced even further, because the system will detect there's a low correlation between "high siege damage" and "this character is winning", since there will be plenty of games where a hero wins without ever piloting. Remember, the system is looking for patterns of stats that lead to wins - when there's a lot of noise (siege damage on GoT), there won't be much importance placed.
That's my guess, or they set a maximum bonus you can possibly earn from siege that is achievable even if you don't use the terror. Any extra siege above that cap doesn't benefit your performance rating.
|
On November 30 2017 08:10 karazax wrote:Show nested quote +On November 30 2017 01:11 ChaosOS wrote: Regarding Garden Terror & other vehicles - It's pretty likely that for many heroes Siege Damage is a stat that's barely weighted. Furthermore, because it's a per-map basis, the valuation for siege damage may be reduced even further, because the system will detect there's a low correlation between "high siege damage" and "this character is winning", since there will be plenty of games where a hero wins without ever piloting. Remember, the system is looking for patterns of stats that lead to wins - when there's a lot of noise (siege damage on GoT), there won't be much importance placed. That's my guess, or they set a maximum bonus you can possibly earn from siege that is achievable even if you don't use the terror. Any extra siege above that cap doesn't benefit your performance rating.
They won't do that since the entire point of the system is that they don't pick the winners and losers- the algorithm determines what kind of stats are meaningful from data.
Unfortunately, the data won't be evenly distributed. The only way the system will account for this is in heroes which commonly or always get in the terror. (vikings come to mind)
|
On November 30 2017 01:11 ChaosOS wrote: Regarding Garden Terror & other vehicles - It's pretty likely that for many heroes Siege Damage is a stat that's barely weighted. Furthermore, because it's a per-map basis, the valuation for siege damage may be reduced even further, because the system will detect there's a low correlation between "high siege damage" and "this character is winning", since there will be plenty of games where a hero wins without ever piloting. Remember, the system is looking for patterns of stats that lead to wins - when there's a lot of noise (siege damage on GoT), there won't be much importance placed.
This doesn't solve the issue. If you play a hero that is expected to do 60k siege damage in a game, but get an additonal +60k siege damage, why would that siege damage be any less valuable than a hero that is expected to get 120k siege damage?
Obviously a proper model is going to reward you for that as long as you are keeping up with the other metrics.
So if it doesn't look at which heroes have been in the terror, then it is going to reward players who gets an usual amount of siege damage no matter what. And siege damage is extremely valueable - for every hero.
Further, because it doesn't look at talents, it cannot effectively predict which heroes are supposed to get siege damage and which aren't as that depends on the talent choices.
|
|
|
|