Red Post Collection: Quick Gameplay Thoughts: 2/28, /Dev: Matchmaking Real Talk & More

Today's red post collection includes Meddler's quick gameplay thoughts for February 28th, including thoughts on Sion, the Conqueror rune testing on the PBE at the moment, as well as the first of three /Dev blogs on Matchmaking, red comments from around the wed, and more!
Continue reading for more information!

• Red Post Collection: 2018 Global Esports Event Locations, 50% off Select Emotes, Kai'Sa discussion, & more
• 2/27 PBE Update: New Summoner Icon & Tentative Balance Changes
• 2/28 PBE Update: "KAISA_EXOSUIT_LOADER.log" & More Tentative Balance Changes

Table of Contents

Quick Gameplay Thoughts: February 28

Here's Meddler's quick gameplay thoughts for February 28th, covering Sion changes, the Conqueror rune in testing on the PBE, and more:

"Hey all, 

Usual Disclaimers 

These posts will often contain talk about future work we're doing, or planning to do, that isn't yet guaranteed to ship. The nature of the work could change or, depending on what we discover, projects mentioned may get put delayed or even stopped. If you'd like to see a Tweet whenever a new one of these posts goes up: https://twitter.com/RiotMeddler 

Sion 

We fixed a couple of bugs on Sion in 8.3, which likely made him a bit stronger. At the same time it turns out we also unknowingly introduced a new bug as well, which is likely adding quite a bit more strength to him than those bug fixes did. In order to address an issue where minions he E'd would sometimes die before the end of the knockback we ended up rebuilding the spell. The rebuilt version was unintentionally set to use a different check to determine what was hit (distance between the E and the edge of the target's model, instead of the E and the center of the target's model). That basically means Sion E is wider than it's meant to be at present, which adds noticeable extra power. That will be fixed in the next patch. 

Conqueror 

We're continuing to test the new Precision keystone, Conqueror, with the goal of shipping it in 8.6. Been seeing quite a bit of discussion about it and wanted to talk about out a few things with its design:  

• While you do have to be in combat for 4s before it triggers that's not the same as you have to be hitting your target for 4s before it triggers. Like Grasp of the Undying you can prep it by taking or dealing damage (e.g. poke spells or other targets) before getting onto the enemy you particularly want to fight.
• The true damage conversion it offers is somewhat anti synergistic with champions who build pen or have it on their kits, or have true damage already. Converting a % of your damage to true is much more valuable when that damage was previously getting mitigated at a high rate. If you're got significant % pen/shred already (e.g. Darius with BC and his E armor pen) and/or deal a lot of true damage regardless (e.g. Fiora) you'll be getting less of a damage amp than champs without either. Might still be a good choice, especially if you're facing really tanky enemies, but less likely to be the strongest use case.
• The delay and the true damage are both there in part to create more distinction from other keystones. To get meaningful choice between runes they need to be offering significantly different things, which can be tricky when their output is just damage. That means creating distinction between damage runes through things like when in the fight they offer damage (e.g. up front versus after a delay), who they're good against (e.g. squishies versus tanky targets versus low health enemies etc), whether they're pure damage or also include some other effect (e.g. healing) etc. Lack of that sort of distinction was one of the big issues with Fervor, given it switched on almost immediately (most of its users could rush its ramp up significantly) and then stayed on while being equally effective against whoever (given it was raw AD). 

Longer thoughts on Marksmen delayed 

I mentioned a while back I was looking to talk about our thinking on the state of marksmen in detail soon. Going to have to delay that until quite a bit closer to mid-season. We're still working through a lot of interconnected issues there around marksmen and bot lane overall that will need a fair bit more time before we've got some clearer conclusions. Apologies this is taking quite a bit longer than originally hoped. 

Some of the things we're looking at: 

• The arms race between champs who build multiplicative damage and multiplicative defense and how that affects champs building a mixture of both
• The need to be able to kill full tank builds late game
• Whether the balance of power between marksmen general damage and marksmen tank busting effects is correct
• How front loaded marksmen damage can get, later in the game especially (e.g. crit + Shiv crit)
• What stats are needed to make playing a marksmen feel good earlier in the game versus what power's actually needed
• How much reliable burst mitigation there is available for marksmen, both from supports and from other sources
• How marksmen power curves compare to other classes and whether marksmen have sufficiently distinct power curves from each other
• Marksmen wave clear late game (and wave clear in general)
• Amount of power having decent resourceless ranged damage early game offers (while many champs are still quite mana gated) 

Plus likely a few other things that will emerge as we keep working. One of our next steps is going to be narrowing down to a subset of issues, with the goal of making significant progress on some rather than a little bit on everything. We'll also going to need to do some work on non marksmen too as part of our work to open up bot lane more."

When asked his thoughts on the jungle after Tracker's Knife was removed in Patch 8.4, Meddler commented:

"Quote:

Thoughts on the jungle landscape in Pro Play/Normal Play now that Tracker's Knife has been gone for about a week? 

How has it effected Lee Sin's performance so far? 

Will we ever see a jungle ward item/a new smite upgrade to give us more options than Red or Blue Smite?

Doesn't seem to have affected normal play too much at most MMRs at least, which isn't surprising given it wasn't bought much either. 

Most pro play has still been on 8.3, so need a bit more time to assess that. 

Lee Sin looks to have taken at least a small hit. He was on the weaker side anyway, so some argument for giving him a bit of power. At the same time we're going to be hitting Cinderhulk, Warmogs and some of the strongest junglers, so he'll get indirectly stronger. He's a possible 8.6 candidate, more thoughts on 8.6 balance on Friday. 

Possible we add another option instead of Red/Blue Smites, nothing immediately planned though, and would want to avoid something like Trackers that's just flatly the best choice, even if many players aren't willing to use it (vision's not cool and all)."

When asked about the specifics on Marksmen and Crit builds, Meddler replied:

"Quote:

A very technical question regarding Marksmen and their Crit-builds:
Did you guys modifiy the C value of the PRD algorythem?
Especially, if you introduced a mechanic to lower the inital crit value after crits in a row? 

I would love to know that, if possible. It is sadly basically impossible to test for those things by myself, I tried it :/
This serves the purpose to create a fairer crit system/rng, while it still beeing random! 

Thanks in advice :D

Yeah, Crit does run on a pseudo random system. Your chance to crit initially is equal to the displayed value. The more the amount you crit diverges from the expected value though the more your future crit chance will get adjusted until it gets back close to expected values. If you've got a Zeal, and have got a string of crits in a row for example your chance of critting will be somewhat below 20% as a result. Or if you've got an IE and a PD and had a whole run of non crits your crit chance will be higher than 50% and continue to rise if you keep not critting. 

Having said that, it's not a dramatic change over a small number of attacks. Two non crits when at 50% chance to crit won't give you a massively increased chance on your third shot. It's more a system intended to counteract more extreme streaks over a larger number of hits."

On Xayah and Rakan, Meddler noted:

"Quote:

Any planned changes for Xayah and/or Rakan, I don’t have a problem with them I was just wondering if riot sees a problem with them? They are my two favourite champs in the game.

Nothing significant planned. Maybe some balance tweaks at some point, but even then nothing major certainly, they seem in a good spot overall."

On the Restoration passive, Meddler commented:

"Quote:

What ever happened to the Restoration passive on Hunter's Talisman? Was it deemed too weak/too strong? It still makes me sad that Hunter's Talisman's mana regeneration does not function in the river, the dragon pit, or the baron pit at all, while Hunter's Machete gets full benefit from all of its passives in all of these scenarios.

Yeah, there are some issues with both the Machete/Talisman relationship and with how mana and health sustain for junglers work in general. Some champs stay topped off too easily, which gives them a lot of extra power in clearing and ganking early. Others struggle to clear well enough in the first place. We're working on some tentative changes around that, possible they're in 8.6, depends how they test though."

Meddler noted when the midseason patch was planned:

"Quote:

Hey Meddler can I ask what patch the midseason patch will likely fall under.

Probably 8.10"

Meddler noted he would go into more detail Lissandra next week:

"I'll get some notes on Liss into a post for next week. TLDR is some things working, some not, should have more clarity by then."

/Dev: Matchmaking Real Talk 

Check out the first of three /Dev blogs on Matchmaking on the Nexus - /Dev: Matchmaking Real Talk:

"Ask Riot gets a lot of questions about how we decide on your teammates and opponents in League, and we wanted to take some of the frequently asked ones and give them to the designers that work on matchmaking, champion select, position select, and all that good stuff. Let’s dive in: 

WHY DO SOME GAMES OF LEAGUE FEEL IMBALANCED? 

When you encounter imbalanced games of League, we generally chalk it down to three major culprits: snowballing, matchmaking uncertainty, and/or perception. 

In terms of design, we want League to be a game where getting a small advantage and running with it is one path to victory. Any small successes that are snowballed can determine early leads, and then sometimes the result of the game. Early Pantheon double kills can feel unstoppable without a miracle teamfight, but there’s always itemization and strategic calls that can turn games around from the brink of defeat.

Within the actual systems, most teams have an expected win rate of 50 +/-1%. This means from all the data we have, we think we’ve made a fair match. But we can’t rely purely on the raw data. What if players are playing positions or champions they aren’t familiar with? What if there’s a large discrepancy in MMRs in a normal premade, or a player hasn’t played enough games to get an accurate rating (looking at you, smurfs)? On top of this, there could be personal factors that matchmaking can’t see—lack of sleep, one too many Graggy Ices, that sorta thing.

Depending on how you’re doing, it could just be your perception. Being wrecked is much more visceral than the joy of dominating your opponent, and the really bad moments can unfortunately be more memorable than the great ones. 

And sometimes you’ll just have a bad game. 

While we know our matchmaking systems are never going to be perfect, and sometimes you (or your opponent) will end up with a 20 minute win, we try everything we can to prevent imbalance before the game has even started, and consider it a huge priority to ensuring League remains competitive. 

— Riot Gortok, Designer, Get in Game Team 

WHY DO I GET WORSE PEOPLE IN MY GAMES WHEN I’M ON A HOT STREAK AND TRYING TO CLIMB? 

This is pretty much an urban myth. Mostly. 

There’s nothing in the MMR system that forces you to have lower-skill teammates or disproportionately higher-skill opponents. We expect every game you play to have a 50 +/-1% chance of your team destroying the Nexus. 

As your MMR goes up, you’ll stop being the “standout” player and your teammates’ skills will be higher. This can also happen to anyone else in the game, which can give the impression that you have “worse” teammates, when really they might just be newcomers to your skill bracket. 

Even if it all goes wrong and you feel like your teammates have been letting you down, you can make use of systems like demotion protection and promo helper to prevent you from being punished for a few “unlucky” games. We think this is the right tradeoff compared to having a volatile rank—being demoted feels rough, and reaching a tier should be a decent indication that you can play at that level. 

— Riot Gortok, Designer, Get in Game Team 

WHY DO I COME UP AGAINST HIGH-RANKED PLAYERS IN UNRANKED GAMES? 

MMRs between queues aren’t connected. This means occasionally high-MMR ranked players have lower MMRs in normals. 

Our perspective on the queues is that ranked should continue to be where you put on your tryhard pants. In contrast, unranked queues should be a place to play and experiment in a less intense competitive setting. The reason the MMRs aren’t linked is that introducing a consistent connection could change the competitive intent of each queue. 

For example, if having a higher ranked MMR meant you would be placed higher in normals, it could put more pressure on you to have a “ranked mentality” in all queues. Beyond this, we want to provide an outlet for experimentation and lower stakes through unranked modes. A player who tries super hard in ranked may play with their casual friends and try goofy strats in normals in a way that means they perform at a lower level. 

That said, we think there’s room for improvement here. We’re looking into some options, such as having better algorithms to get players to more accurate MMRs in a shorter time, or potentially “splashing” small amounts of MMR from ranked into normals without a direct link between the two. 

— Riot Socrates, Designer, Meta Game Systems 

WHY DO I GET MATCHED WITH UNRANKED PLAYERS IF I’M BRONZE/SILVER/GOLD? 

When players first start playing ranked, they have to go through a period of placement games where we try to figure out how good they are. We have to place them against players we already have a lot of information about to get the most accurate placement. Since most players are Silver and below, we tend to start looking in Silver and move them up or down based on their performance. 

— Riot Gortok, Designer, Get in Game 

WHY DO YOU CONTINUE TO MATCH ME WITH PLAYERS I BLOCK/MUTE/REPORT? 

Short answer: It would literally break matchmaking if we didn’t.
Imagine you’re Challenger in KR (congrats) and you come up against Faker in midlane. He dumpsters you, and your plays show up in a TOP FAILS montage on the Inven frontpage. Next time you see him in solo queue, you rage-check a box that says “don’t match me with this player.” Now imagine 100 more people do that this week.

Matchmaking would have to look at players much further from your MMR because it’s unable to fill a lobby, your queue time would increase, and you’d end up being on consistently less-good teams. Players end up like a bunch of magnets constantly repelling each other—if enough of them misuse a feature like this (and it wouldn’t take many!), then fair matches would end up next-to-impossible to find. 

— Riot Draggles, Comms Strategist, Meta Game Systems 

WHY DOES THE SYSTEM NOT USE KDA AND OTHER SIMILAR STATS INSTEAD OF ELO/MMR SYSTEMS? 

League of Legends is a team game and teams win or lose games together. We don’t want to impose an arbitrary system of rating players on performance because then the best way to beat that system would probably be to get good scorelines, rather than helping the team win. 

We want to reward good play that ultimately leads to a win, no matter how small their impact may seem. Supports sacrificing themselves to save their ADC, tanks zoning three enemies in a teamfight, or assassins diving the backline to blow up the carry are all examples of plays that may not make a great statline but help get the W. Some champions have unique playstyles (think Singed/Nunu) that would be tough to measure, and their mains may not be graded appropriately based on their play. 

You should be rewarded for different types of achievement and impact, so we’re always looking at new systems to surface cool in-game plays (like vision score or unique missions) that aren’t necessarily reflected in KDA or win/loss ratios. We also want to try and give you ways to express mastery across different positions. What sorts of different ways would you like to see progression in League? Let us know! 

— Riot Gortok, Designer, Get in Game 

WHY CAN’T I SEE MY MMR? 

We talked about this when we first released the Leagues system (holy crap, five years ago now). tl;dr: Showing MMR has a lot of downsides in a team-based game like League. 

On the plus side, MMR is a more accurate summation of where you are in relation to other players across the entire server, and showing it can be more reassuring that the match you’re in is fair when scouting your opponents out before a game.

Using MMR as the sole mark of achievement in League punishes half of the playerbase as their MMR will decline over the course of the season, which sucks because most of them are gradually getting better at the game—but so is everyone else around them. 

Ranked tiers also provide contextual progression and status. Knowing you’re “a Gold player” as opposed to “a 1650 MMR player” or “120,353 on the server” gives you clearer targets to work towards. Moving from 1595 to 1600 MMR is probably not that compelling, but promoting from Silver I to Gold V should give you the knowledge that you’re truly improving.

The Leagues system also gives you a bit of protection from losing a bunch of games in a row and having your MMR plummet as a result. Using demotion protection and promo helper, you can get a few extra lives in rare cases of not getting your preferred position for a few games or just being in a slump. Ranked anxiety is real, and we know there can be a lot of pressure, so having meaningful progression that feels good should hopefully break down a few of those barriers. 

— Riot Gortok, Designer, Get in Game 

I’M IN MASTER/CHALLENGER AND MY MMR SEEMS TO BE GOING DOWN. ONE DAY I HAVE FAVOURABLE LP GAINS AND THE NEXT I DON’T. WHAT’S GOING ON HERE? 

LP gains and losses are also based on your skill estimation compared to the players around you. This becomes most apparent at the Master/Challenger level, where there are only a few players and the competition for spots becomes very contentious. If you take a break from playing for a bit, the MMRs of the people around you are still changing, and that can result in notably different LP gains when you return. 

For example, let’s say you grind to Challenger and bank games to lock your LP for 10 days, then go on vacation until your banked games run out. While you’re away, your MMR/LP is safe and static, but all those other Challengers (and Masters, and high Diamonds) are still playing games. This tends to slowly increase the average MMR at the high end of the ranked system. So when you come back from your hiatus, your MMR is right where you left it—but now the competition around you is even higher. When you play again, you’ll get lower LP gains and higher losses, as you are now slightly below the average for your ranking, but as you win a few games, those LP changes will return to normal. 

— RiotIAmWalrus, Designer, Competitive 

We got pretty technical here, but we hope this provided a little more context into matchmaking in League. What other questions do you want answered? Where do you feel we can improve our systems? Let us know in the comments! 

This is the second post of a three-part update on matchmaking in League! Read the rest below: 

Part 1: How does matchmaking work? – A much-needed update to the Riot Support article to better represent how matchmaking works in League in 2018. 

Part 2: /dev: Matchmaking Real Talk – Answering some of the tough questions around matchmaking, and hopefully dispelling some myths along the way. 

Part 3: /dev: Making Matchmaking Better (releases next week) – Take a look behind the scenes at some of the improvements we’ve been making for the last couple of years."

Quick Hits

• Here's a video from Riot KR having an interview with Riot Jag, the designer on Kai'Sa! [Note, the video is in Korean but Jag's answers are in English!]

• Riot Sparkle talked more about the buff bar and runes panel when it came to Runes buffs and tooltips:

"O hai, I'm Riot Sparkle and I'm going to write a long answer because I never get to talk shop about stuff like this and I'm really excited. :D 

So maybe make that panel more useful and put cooldowns and other things there where applicable. 

So, this suggestion is a pretty dang good one - I know because it was actually originally how we had the in game rune panel set up tbh. We had different rune cooldowns and changing damage values on there, and the idea was to provide tactical "in the moment" information on the in game panel. 

However, in practice (i.e in testing) this didn't work out so well. Players needed the information on the buff bar for many of their existing abilities and buffs, and attention was divided between the two spaces. This meant that players often missed important rune info because they were looking at that teamfight action - and the rune panel is farther from the action than the buff bar. Hence, we decided to move keystone cooldowns etc. to the buff bar instead and group all of that "need to know now" info together. 

Additionally, we found that players were actually most often using the runes panel in a "moment of rest", while dead or in the base, for example. And, generally, having a cooldown or value of "0" for something wasn't very useful to them in that use case. Stats like you mention for the quest runes, for example, were far more useful. 

Now, this isn't to say I think we've perfected the formula of showing information on the runes panel - we still have an upcoming task to do a giant sweep of both in-game and end of game rune stats to make sure we're showing the most useful things we can and to push the two rune info spaces further apart (players tend to want different stats during the game than after it from an evaluation standpoint). So, basically, I think there's more work to be done around rune stats for sure. 

I've also talked about how we're going to be evaluating individual runes more often now (ex: Manaflow Band) to see if they need to display more tactical information on the buff bar to be used effectively. The danger there is indeed in overloading that space, which is why we're pretty carefully weighing each rune that goes up there, but I think that runes like Manaflow or Null Orb, despite being normal runes instead of keystones, certainly earn that spot. 

TL;DR We are concerned about having too many things on the buff bar BUT at the same time it made sense to group all of the tactical information you need together in one spot visually, or else it was actually really hard to use that information when you needed it. Also we have more work to do in the runes stat space. "