Red Post Collection: Ekko PBE Feedback, Latest Tentative Karma Changes on PBE, and more

Posted on at 7:04 AM by Moobeat
This morning's red post collection includes Gypsylord with a few comments on Ekko now that he's on the PBE for testing, ricklessabandon with another iteration of his tentative Karma changes,  Meddler on the Blade of the Ruined King PBE buff, a reminder that ARAM Skin boosts are enabled, and more!
Continue reading for more information!

Table of Contents

Ekko PBE Feedback

With out next champion Ekko up on the PBE for testing, Gypsylord has swung back the PBE feedback thread to respond to some of the initial feedback and share tentative changes coming in the next update:
"Hey everyone,

Thanks for the feedback so far. Taking it all in and delivering things that come up often to the relevant contributors.

Bug Fix 
Thought I'd first let you in on a major bug we're trying to clear up. Currently the after-image on his ult is not working as intended. Whether or not you can see it on the enemy team is kind of up to random chance at the moment, sometimes it's visible, sometimes it isn't, sometimes it looks correct, and sometimes it looks like an exact copy of Ekko!

Intended functionality for enemy team vision of the clone is as follows: If you can see Ekko, you can see the clone. The clone disappears if you can't see Ekko. The clone looks like a team-colored version of the one Ekko sees, not a fully textured copy of him. We're working to get it there so bear with us. 
W delay 
Second, a note on the W's 3-second delay. Frankly, it's very much intended. It's a predictive skill that you should be thinking about using BEFORE you fight, not something you should be using to react to the Vi who just jumped on your face out of nowhere. This feeds in to the goal of giving Ekko numerous things to feel "smart" about. Is it skillful to split second spell-shield a Vi ult right as she hits you? Sure. It's SMART to predict that Vi's going to ult you in 3 seconds and detonate a giant shield and stun on her face right when she's about to get her punch on. It may feel unreliable at times but my hope is that these failures make the success case feel all that much sweeter.

Why AP? 
So Ekko to me thematically does not feel all that out of place in the AP space as he's a kid who isn't really all the physically strong but instead is leveraging the hextech powers of his Zero-Drive to move through and influence the timeline as well as damage opponents.

Mechanically AP scaling is important on Ekko because it helps us regulate his maximum damage potential (no crazy AD basic attacks to worry about) which in turn allows us to give him crazier stuff in terms of mobility and CC. The power of Ekko's slows, hastes, stuns, and movement is only okay because the amount of benefit he gets out of standing there wacking an opponent affected by them is capped by his spells and AP ratios. Also helps to regulate things like tower pushing and vamp, which tend to get pretty crazy on chars like Zed and Talon. They kill you, heal up off the wave, take your tower, and then gank bot. For as powerful as Leblanc is at least she doesn't wreck your towers or remove all the damage you did to her before she died (and yes Ekko's R does that but it's on a CD). Are AD assassins bad to make? No, not necessarily, they just gain access to a whole new set of powerful things that I felt should not be as strong on Ekko..

Putting in a few nerfs for Ekko that should be out sometime tomorrow. After watching a number of PBE games we feel that Ekko tends to get pretty trainy (runs you down and beats you to do death with basics) especially when ahead. Additionally, his tower taking is pretty crazy. To combat this we're reducing the bases and scaling on his basic attack and hitting the AP ratio on the W passive. Hopefully this should make it a bit harder for Ekko to snowball and make him a bit less oppressive when doing so.

Reduced base attack speed to .644 from .658
Reduced AS/lvl to 3% from 3.5%
Reduced AD/lvl to 3 from 3.5



On-hit AP ratio reduced to 1% missing health per 45 AP from 1% per 35

Thanks again for all the feedback, will keep reading it and responding when I can."

Check out these links for more on Ekko:

Latest iteration of tentative PBE Karma changes

Speaking of the PBE and tentative changes, ricklessabandon has also tweeted out his latest iteration of the PBE Karma changes. These are set to hit the PBE on 5/15.
[1] "new iteration of karma should be going out in the -next- pbe update (not tonight)."
[2] "changes should be "as compared to live" so no more 'gang up' on renewal, and no more monster targeting on focused resolve."

Here's the list typed out:
Gathering Fire (passive)
  • CD refund on spell hit increased to 2/2.5/3 at level 1/7/13 from 2 at all levels.
  • CD refund on basic attack increased to 1/1.25/1.5 at level 1/7/13 from 1 at all levels
Renewal (R+W)
  • No longer deals damage
  • New Effect: Increases root duration (if triggered) by .75/1/1.25/1.5 seconds
Inspire (E)
  • Cooldown lowered to 10/9/8/7/6 seconds from 10 seconds at all ranks
  • Mana cost lowered to 60/65/70/75/80 from 60/70/80/90/100.
  • Shield amount decreased to 80/110/140/170/200 from 80/120/160/200/240
Defiance (R+E)
  • No longer deals damage
  • Now increases primary shield by 30/90/150/210 (+30% AP)
  • AoE shield amount increased to 50% of the primary target's shield from 30/70/100/150

Expect these changes in the 5/15 PBE update.


Meddler also popped on the boards to mention they are experimenting increasing Blade of the Ruined King's unique active range up to 550 from 450.
"We've actually got a buff to just that being tested at the moment (trying BotRK at 550 again instead of 450). We originally nerfed the range when the active was more powerful, now that it's weaker however we want to see if we can get away with putting the range back up, make it a bit smoother to use."
As of the 5/14 PBE update, this change and a similar range increase for Bilgewater Cutlass are up for testing.

Zed's 5.9 R change

In a boards thread discussion the 5.9 changes to Zed's R and if the team had considered perhaps lowing his health regeneration instead of nerfing the R, Meddler replied:
I'm all for Zed nerfs, though this one seems a little extreme and kills some of his core power: evasiveness. On the other hand, Zed has above-average sustain for an assassin that has manaless poke and good last-hit capability, so I'd have thought that nerfing his health regeneration to a level similar to Katarina would be a good type of nerf. 
Unfortunately, it would seem you don't think that way. Why is that?
The nerf to Zed's ult's there because we believe other players need to be given at least a brief period in which to respond. That's a principle driven change, rather than a power driven one, so changing base stats doesn't address the same issue. 
If, as a separate problem, Zed's shrugging off damage too quickly in lane we certainly could nerf his health regen. There is the risk with that approach on a character that builds lifesteal early of creating a more snowbally pattern of success/failure however, something assassins are already prone to. Would want to dig into what the laning problem was, when in the lane it occurred, against who etc before picking a specific solution."

Shop Freeze Bug fixed in 5.9

Over on reddit, Reinboom commented on a thread about a long standing  and elusive "shop freeze bug" (which as the name suggests would freeze the in-game shop when purchasing items) being fixed in 5.9:
";_; Yay. 
There was a few fixes for 5.9 that "guessed" at trying to fix this. I've actually been looking at this problem on and off in bursts since season 3 and we didn't have a strong lead until a few weeks ago. The most major issue being... we never had a machine that would consistently get the issue. We had data that showed it occurred but in house repros was... non-common. 
If you get freezing again pleeeease be outspoken on it. 
There's been quite a few teams or individuals working hard on getting performance in place in other areas of the game as well. I noticed some comments about "5.9 seems to be more smooth overall". If that's true, it's totally on some of awesome engineers (that I don't know the summoner name of unfortunately) doing awesome performance work."
When asked if this had anything to do with Ekko's release, she noted:
"Nope, pure coincidence. We would've liked to fixed it at any time before.

Or a more elaborate story of "why now and what was it?":

We were adding tech to lock and unlock items based on any condition we generally desire. We did "targeted" features such as this before (see required level), but never something completely generalized. Since tech like that has performance concerns since it would require "refreshing" the requirement cache of each item (the same as any inventory change), I tried to put in a couple performance pay downs in to those changes.

One of those pay downs made my own machine use it SLIGHTLY faster... but the machine of a different Rioter suddenly get the freezing and get it REALLY badly. More so, it had the same pattern that players reported (gets worse with game time, happens on opening the shop and purchasing, selectively really bad with undo).

Since this was in a paydown with a bunch of performance related changes (as well as new tech), there wasn't a definitive origin point. Worse, there wasn't a repro on my machine (as the dev doing the work) so I could only verify changes by kind of guess work and passing it over. More exactly, I would "binary search" additions (removing a large portion of the changes and then passing a new build over to his machine) in order to try to hone down the scope.

Eventually we isolated it to one small area and some code interacting with that area. There are two different possible ways of readding and redrawing items to the display with the library we were using. We were doing the recommended method (which had trivial performance gains on most machines). The other method... worked fine on most machines, but was horrendous for when the machine caused the freezing case. Worse, one of the paydowns was misswritten and would cause that specific type of add/remove to occur a variable "large" number of times more (from 2x to 10x more often - basically, I forgot to clear the "cache is dirty" flag). This wouldn't have a massive effect on most machines, but it made it superbly obvious when the problem occurred.

This tends to happen with software engineering work. The common meme here is "spaghetti code" but that term speaks to unpredictable control flow (e.g. using try-catch for intentionally branching logic). The League game code's control flow is actually rather nice to navigate (basically... not spaghetti), especially as of late (the engineering team has done a TON to make it really nice to work in), but issues like this can still occur. In this case, a bad misunderstanding with an external library. Or in a lot of our other cases older scripting practices badly interacting with newer technology.

ARAM Skin Boosts are back

After being added in 5.8 then quickly disabled due to a bug, the new ARAM skin boosts are being re-enabled on live.

Here's Chager with more information:
"We waited a little while after the release yesterday before turning ARAM Skin boosts back on. They should be on and active for all regions except for EUW (there's an ongoing issue with the store there, but we hope to have it resolved and have skin boosts turned on tomorrow)."
For a refresher, ARAM Skin Boots are :
Skin boosts have been permanently enabled on ARAM. They cost 95 RP and grant you and your teammates a one-time random skin for your randomly-selected champion, as well as 100 IP (200 for you).

Increase of Drophack - May 2015

Last up we have Riot tmx with a message on the recent increase in drophacks:
"We have seen a resurgence of drophacks recently across all League of Legends servers worldwide. A drophack essentially results in a small subset of players facing the “Attempting to reconnect” message while playing their games, with a possibility of the game never finishing or disappearing from one’s match history entirely. 
Detecting and preventing cheating is an ongoing challenge, and we're constantly improving our technologies to this effect. We have multiple projects in the works that should improve the experience, but we still need some time before deploying them to Live. We can’t guarantee an end to all drophacks, as we all know this will forever remain an arms race, however we strongly believe that this will impede those with malicious intent. 
Simultaneously, we kicked off suspension proceedings against all accounts involved in drophacking, past or present. Similar to last year's ban waves, we won’t be showing any mercy and all corroborated accounts will be permanently banned. No warnings or two-week suspensions, just pure and simple account closure. The ban waves have already started, and we expect to get a vast majority of the accounts processed by the end of this week. It will take some time to complete, but the delay remains in line with our general approach: we have to be 100% certain that a given account is guilty. We are confident in our ability to detect offenders, and aim to decisively punish offenders. 
We don't expect the ban waves to immediately reduce this trend however. The decision to perpetrate drophacks has more to do with personal attitude, one's ethical standards and personal respect for rules and other players. We want League of Legends to be competitive, but we want a fair rivalry and a positive challenge. We believe that players attempting to drophack have forfeited their right to play our game and be a part of this community. 
Examining the data, we don't even see any great rewards coming from such behaviour. Guilty players only see temporary personal benefits: a few additional LP points saved that they would have lost otherwise; maybe a promotion, but is it really worth it when you’re jeopardising your entire future with League? The temporary and minor reward doesn’t make up for the extent to which drophacking hurts the experience of fellow players. 
In Europe we’ve already rolled out a number of “reinforced” permanent bans (ie: legal prohibition of accessing any of our services under penalty of prosecution) against extremely negative players, and if we notice the same people repeatedly displaying extreme toxicity or antisocial behaviour, we will expand this group further. We hope to deploy the fix soon, and thank you for your patience while we’ve taken measures to address this issue. We would’ve liked to have posted this message earlier, but wanted to make sure that you would know how seriously we take this problem, and hope the actions mentioned above ensure the best possible League experience for the community going forward. 
Thanks for taking the time to read this update, and we hope you’ll feel an improvement to the experience very soon."

No comments

Post a Comment