Advertisement

Checking Riot's lane swap assumptions: Part 2

by theScore Staff Jul 25 2016
Thumbnail image courtesy of Riot Games/lolesports / NA LCS Summer 2016 / Riot Games

So far, I’ve mainly dealt with statistics and fairly small, straightforward statements from Riot Games’ original statement discussing the changes coming to turrets and minions on Patch 6.15. Addressing the rest of Riot Games’ claims for why they believe the changes they are implementing are necessary requires more interpretation and subjective reasoning, including a discussion of whether or not lane swaps are dynamic. As a result, I've given it its own separate treatment.

Formulaic with few strategic tradeoffs

We have now entered the territory where statistics can no longer assist us. Other than the statistics-based assumption that teams internationally find it attractive to both lane swap and play standard lanes, there's no hard statistic that can prove or disprove the statement that there are few strategic tradeoffs in lane swaps.

The common assumption is that teams will use lane swaps to avoid very skewed matchups for top and bottom lane. But over time, lane swaps have evolved in such a way as to prioritize fast pushing. When teams go into lane swap scenarios, top laners risk getting dove in 2v1s, so changes to turret resistances and strengths made it increasingly optimal for top laners to double jungle. Then, changes to jungle creep experience sent top laners firmly into 3v0 scenarios with bottom laners to fast push the first turret. If teams chose not to fast push the first turret, they would fall behind in gold and map control, giving the opposing team advantages they could pressure and snowball.

Since matchups in standard lane scenarios might seem like they favor certain champions frequently, it’s easy to see why someone might think that lane swapping is always optimal. The decision to lane swap, however, is rarely this simple.

In many cases, the jungle matchup may influence the decision to swap. Part of the reason first bloods occur later in lane swap scenarios is that having 3v0 scenarios incentivizes the jungler to continue farming rather than ganking. A jungler with high pressure early like Elise can take more advantage of a standard lane scenario than a jungler like Graves who might prefer to farm until Level 6. In this case, a team with a scaling jungler and a strong top lane matchup may have to choose between assuming the top can get ahead on his own with his jungler in the area for a countergank in standard lanes, or allowing the jungler to scale and not using a strong lane matchup. This is when the decision to swap becomes more subjective.

Other examples include the case for Ashe as a champion. Ashe has perhaps the strongest level 1 of any AD carry, but she can also clear waves fast in a lane swap, and her ultimate benefits considerably from opening up the map early. When Ashe is hidden in fog of war with multiple turrets down, and the opponent duo lane doesn’t have eyes on her, she can send deadly ultimates across the map with less notice, potentially resulting in more effective catches. Teams then choose between lane swapping with Ashe to take advantage of this — perhaps if they want to get an advantage in the mid lane, or maybe they're trying to use Ashe to get an advantage in a 2v2.

Assuming a team opts into a lane swap, they still have several dynamic choices to make. For example, the LPL teams found ways to utilize a strong jungler, even in lane swap scenarios, to get leads early by ganking the mid lane at two minutes while the rest of the team were tied up in lane swaps. One well known example was Royal Never Give Up at the Mid-Season Invitational.

In the example above against Counter-Logic Gaming, RNG made the decision to delay the double jungle, setting Jang "Looper" Hyeongseok behind. He secures three small raptors on his own before being chased from his jungle by CLG’s support, Zaqueri "Aphromoo" Black. In return, Liu “Mlxg” Shiyu ganked the mid lane at around two minutes and fifteen seconds for first blood. CLG top laner Darshan "Darshan" Upadhyaya accompanied his own jungler for extra experience.

Another LPL quirk of the lane swap is often to push out a wave and back to match (this happens in many regions when teams want to force a 2v2 after starting in a swap, but quite often in the LPL, and frequently multiple times after an initial back). Darshan initially started Teleport when RNG’s bottom lane started their backs to catch the wave, but canceled it, either because he believed CLG were ahead enough in the push to get the turret or because he had no knowledge of Mlxg’s location.

RNG sent Looper to the bottom lane for one wave of creeps, waited for him to finish farming it, and then used his Teleport to prevent CLG from finishing their push on the top turret with their duo lane having rotated to the top side. In this way, RNG managed to come out ahead. This case included a wide variety of decision-making and tradeoffs that was far from formulaic.

Perhaps one of the most interesting games this split that I’ve watched in any region is the first game of the second round match between Immortals and Cloud9 in the NA LCS. Within the first nine minutes of the game, Immortals took Cloud9’s inhibitor and started on the Nexus turrets. Cloud9 answered with an inhibitor of their own before backing and answering Immortals’ game of chicken.

In this case, I’m not going to break down the lane swap, but rather some considerations that have to go into a situation like this one. Jung "Impact" Eonyeong played Trundle, a champion that doesn’t clear super minions well and doesn’t perform well in lane swaps. As a result, opting into this scenario may better favor Immortals in that respect because Rumble may be better equipped to clear super minions.

Immortals’ pick composition may also benefit much more from a scenario that requires them to focus the mid lane earlier. In a Karma vs. Twisted Fate matchup, both mid laners keep shoving, while Twisted Fate has an advantage in roaming. With top laners pushed into their bases, Immortals can restrict the lanes Nicolaj "Jensen" Jensen can influence early and snowball, as Jensen wouldn’t want to port into Immortals’ base. Immortals are also now pressured to push mid with the next priority target being dragon and the mid lane Tier 1 turret, which keeps Jensen in lane.

Over time, however, Immortals have to be aware of the fact that their decision accelerated the pace of the game. The split-pushing threat of Trundle and Twisted Fate with an exposed inhibitor with dragon in play would be massive for Cloud9. To an extent, they could also take advantage of Baron, though the bottom lane turret would have been more optimal. If C9 could stall to force these trades, they would happen earlier than if the inhibitor had not been opened up. Immortals acknowledged the tradeoff they made and did their best to force the game to end earlier. Cloud9 willingly opted into earlier fights, misidentifying their win conditions and losing the match.

Pushing past the inhibitor was also a strategy favored not infrequently in the Spring by Flash Wolves. If one watches a Flash Wolves game in which this occurred, one can notice the team pulling the wave of enemy super minions to the mid lane. This denied minions coming from the team’s own base down the mid lane to the enemy mid laner to allow Huang "Maple" Yitang to gain massive advantages. Flash Wolves have also experimented with proxying waves before they pass the outer turrets this summer. These types of continuing innovations from one of the world’s leading teams suggest lane swaps are far from formulaic.

So far, one may argue that I’ve looked at particularly interesting variations in lane swaps and have avoided talking about the kind that most think of when discussing lane swaps. Specifically, 3v0 trades of all outer turrets in the side lanes prior to 10 minutes, followed by the team in the bottom lane after the second set of turrets fall, taking the first dragon if they’re ahead. These have their own nuances and tradeoffs.

In the first game between Splyce and Fnatic in Round 2, there was a much higher incentive for Fnatic to lane swap. Neither Braum nor Trundle are high impact supports in lane, but Jhin has more early game power than Ezreal. Fiora and Gnar is a complicated 1v1 matchup, but Gnar typically does better in lane swap scenarios due to range and his ability to build lower cost items. Both Hecarim and Gragas have incentive to farm for a Level 6 power spike, but it’s generally agreed upon that Gragas’ capability for pre-six ganks is higher.

Martin “Rekkles” Larsson and Bora “YellOwStaR” Kim went bottom to encounter Kasper “Kobbe” Kobberup and Mihael “Mikyx” Mehle. Fnatic’s bottom lane then warded, backed off, and walked to the top side of the map for the lane swap. Splyce’s bottom lane didn’t notice they were swapping until they started to invade Fnatic’s enemy krugs, but then saw Rekkles and YellOwStaR on the top side of the map forcing Martin "Wunder" Hansen and Jonas "Trashy" Andersen off their own krugs. Kobbe immediately reacted to return to bottom lane.

Ultimately, Fnatic got the lane swap advantage with more experience on Mateusz "Kikis" Szkudlarek after Wunder was forced to Teleport to get to the bottom lane initially and more farm on Lee “Spirit” Dayoon when Trashy had to back after encountering Rekkles and YellOwStaR. Fnatic also managed to secure the first dragon off their tempo advantage.

In this particular example, the lane swap naturally favors Fnatic, but it was risky to take the lanes they did. Fnatic’s early encounter in the bottom lane was a red herring because they wanted the lane swap more than Splyce. No matter what, there’s no way to completely guarantee you get the lane start you want. If Splyce had gotten standard lanes, this would likely have gone the other way, especially around jungle and top matchups.

This means that Fnatic took this draft as a risk, the tradeoff being the possible early loss if they were forced to play standard lanes. Splyce similarly accepted the risk of a lane swap, hoping for a standard lane scenario they could more easily pressure. This reflects both strategic tradeoffs and interactivity in several instances where Fnatic players and Splyce players try to force encounters to get advantages, even in non-standard scenarios. Yet it’s unlikely Fnatic would draft like this if they knew they could not get a lane swap, which will limit the types of compositions they might run in the future and make counter-picking more fundamental, actually reducing the number of strategic tradeoffs that can be made.

As a hinted at briefly earlier, it’s rare for a case like Fnatic where lane swaps are heavily favored by one team over another. It’s more common to get a top lane matchup that may benefit from standard lanes, and a bottom lane or jungle matchup that would suffer as a tradeoff. One of Immortals’ analysts, Nick “Lufty” Luft, published a video addressing some of these ideas and how strategic tradeoff decisions can tell you a lot about a team, so I won’t go into it in depth here.

In general, however, I would argue that lane swapping in its current form is far from formulaic. Regions have different approaches to it, teams within regions have different approaches, and there is a great deal of interaction between teams even in lane swap scenarios. And if Riot want to create more strategic tradeoffs, lane swaps are currently one of the richest places to find them.

Making professional games look like ranked games

One valid criticism of Part 1 of this investigation is that, by focusing specifically on the initial statement Riot gave, I have ignored additional statements they’ve made on why they’re implementing these changes. Part of the reason I saved it for later is because this discussion will have far more to do with speculation and opinion on what the future will hold rather than what we can already see in the games, which I wanted to avoid upfront, simply presenting available evidence.

The other reason is that I think these additional comments are the main crux of why Riot want to drastically reduce the occurrence of lane swaps, so I find it slightly strange that they were not included in the original statement. In response to Part 1, North American LCS caster David "Phreak" Turley stated, “48.7% of games starting without any real PVP interaction is obnoxiously high.” To an extent, I have already addressed this concept pointing to instances of PVP impact in my lane swap examples.

Using the Splyce vs. Fnatic example specifically, the first interaction is Rekkles and YellOwStaR encountering Kobbe and Mikyx. The second is Fnatic’s bottom lane forcing Trashy back when they invade the top side at Level 1 and damage Trashy. Then, Wunder risked heading to lane and dealing with a collapse from YellOwStaR in hopes of making up extra experience before Teleporting bottom. Mikyx also went into Fnatic’s jungle to try to chase Spirit and Kikis off their camps. Kikis intentionally saved his Teleport after taking the first turret and waited until Splyce’s bottom lane appeared. He was forced to back off and Teleport from there, but could have gotten caught in this instance. All of these kinds of interactions don’t necessarily result in kills, but are definitely PvP interactions, and being mindful of where the enemy team is on the map is an implied PvP interaction as well.

In this way, the idea that players are just failing to last hit against an AI when they fall behind in lane swaps is also dismissive of the decisions Fnatic’s players had to make to get Kikis ahead, some of which came as the result of actual altercation with Splyce players.

Ultimately, the bottom line is that what is exciting, high tension, etc., is subjective, but lane swap games do feature multiple instances of interaction and complexities not assessed in some of these comments.

Beyond this, it still feels like the main driving motivation by Riot to try to minimize lane swaps is another comment made by Phreak. “We just want the game to look at least something like the game people play at home.”

Most would agree that lane swaps represent the easiest difference to spot between professional play and games that individuals can play on their own when they queue up with others in their homes. Ranked or normal League of Legends games hardly, if ever, feature lane swaps. It’s reasonable to assume that nearly eliminating lane swaps — which does seem to be the objective of these changes, rather than the “not making them the default start” statement originally provided — in professional play will make games look more like games at home.

But pro games won’t ever look exactly like games played at home. Current LCK caster and ex-League of Legends Coach Nick “LS” de Cesare made a video discussing how less risk-taking is an inevitably of competitive games as skill differences in players equalize and more support staff enters to add numbers and weight the value of possible actions to take. As a result, an increase in what others perceive as passive play isn’t something Riot or anyone else can change.

We can see at least some evidence of this in the game’s history which is directly relevant to some of the proposed turret changes. In Season 3, Riot noticed that early game snowballing was incredibly potent. Level 1s in Season 3 were very risky as teams sought to get early first blood bounties in invades and secure creep experience and gold. In the preseason before Season 4, Riot reduced the value of kills before four minutes.

They later removed this restriction, hoping to encourage more early interaction in the game, but Level 1s remained incredibly safe. At the time, members of the community even argued that increasing early game rewards discouraged invades later because teams didn’t want to take the risk that their invade attempt would backfire and give more gold to the enemy team. Of course, this is not the only explanation as many other changes were made to the game in the intervening time, so it should not be considered the strict cause, just a point of consideration.

Among Riot’s proposed changes to turrets, they want to add a “first blood bonus.” With a reduction of resistances to bottom lane turrets and bonus for securing the first one, one might reason that this will create more early game dive and gank setups in the bottom lane. More emphasis placed on Teleport, junglers pathing bottom, and skirmishes to remove opposition and take the turret early to snowball.

If both teams want to go to the bottom lane to take the first turret, however, this could create the opposite effect. Teams could set up to play ultra-defensively, placing vision and choosing anti-dive champions like Braum or Alistar or Trundle that can punish dives more effectively. The end result might be, because teams are too afraid of the opponent team getting the first blood turret gold, they farm passively and avoid confrontation until a point in the game when 400 gold has marginally less impact and the top lane turret defensive reductions are reduced so they can make a trade at that point, slowing action down. In this scenario, pro games will also hardly resemble ranked or normal queues.

Again, this is purely speculation, and I’m attempting to convey that both these results are possible as are results that have not been considered. If Riot want pro games to look more like ranked or normal queue games, it isn’t unreasonable to try to change the game to get to that result. These changes yielding that result is not completely implausible based on data or speculation.

Which means that there are only two questions left — why, and why now?

“We know the timing’s not great and ideally this would have landed sooner to give teams time before playoffs to adjust. With that said, we think it’s important for the state of the game and something worth implementing now versus waiting until the offseason. We also want teams to qualify for Worlds in the same general meta that they’ll be playing in this October, meaning now’s our only chance to make these changes. While we’re be iterating over the patches to come, these are the last major changes coming to the game before Worlds.”

This section in Riot’s original statement implies that these changes have been a somewhat recent decision. Almost all communication from Riot before this indicated that they wanted to avoid implementing major changes to the game right before this year’s World Championship. Riot’s Mid-Season follow-up communication outlined all patches leading up to Worlds, and patches from 6.15 on-wards only included plans for Worlds-focused tweaks, which doesn’t at all hint at a complete change to how laning phase works.

As a result, this decision feels somewhat rushed and reactive. A change like this would make perfect sense to me if it occurred in offseason, but a need to make it happen before Riot’s biggest professional event of the season makes me think that this abrupt change has a lot to do with recent viewership concerns.

If one has traveled to the League of Legends subreddit recently, one knows viewership decline panic has been all the rage. Some blame Overwatch, some blame Counter-Strike: Global Offensive, some blame too many games making die-hard fans feel somewhat bored, but still others blame lane swaps and the fact that casual fans don’t want to watch LoL because games don’t look like they do when they’re played at home.

Riot may want to test this assumption on the biggest stage, where viewership numbers have traditionally been monumental. They’re potentially afraid we’ll see lower numbers than ever before, and more standard laning games can change that.

Implementing changes focused on making the game exciting before Worlds isn’t new. As Worlds drew closer in 2014, though well in advance of qualification, Riot implemented several changes to incentivize 2v2 laning, in particular increasing dragon gold — as dragon gave flat gold bonuses at the time rather than unique buffs.

One can even argue that Riot’s juggernaut changes were intended to make Worlds exciting. In the World Champions press conference, Marc Merrill stated, “We think the changes to 5.18 had a big timing factor and something we want to reflect on for next year. It was an exciting meta change to give top lane champions a chance to carry, and that's healthy and fun for the game. But we want to think about the large patch from a timing standpoint more as we go into next year.”

Merrill’s statement implies that there are both positive and negative aspects to a large patch close to Worlds, and Patch 5.18 made Worlds seem exciting because a focus on top lane carries was a change of pace. Yet there are downsides Riot would want to reflect upon. This feels somewhat similar in that Riot are banking on standard lanes being more exciting and boosting viewers on the international stage, though I want to clarify that I think that introducing this change now as opposed to specifically for Worlds is a vast improvement for which I do credit Riot.

It just still doesn’t feel like this needs to happen now. I would prefer for these kinds of more experimental changes to occur in the offseason to maintain competitive integrity. That way it also doesn’t have to seem like a rushed endeavor, and a lot more testing can be done getting feedback from professional players at Riot Headquarters throughout the process.

Lane swaps may or may not drive down viewership. I haven’t seen data to support or destabilize this claim, but perhaps Riot have. The reality, though, is that viewership for League of Legends will eventually decline. It won’t always be the most popular esport, but it can have massive longevity, and it can do that by appealing to a core fanbase.

League of Legends has always appealed to many because it has a low barrier to entry for watching, but high levels of complexity. Perhaps eliminating lane swaps is in itself a strategic tradeoff, as the game loses some of its complexity to lower the barrier of entry. Maybe this is the fanbase to which Riot wants to appeal.

As Riot Games continue to balance for competitive, I think it’s time for them to start thinking, not about how their game can have the widest appeal, but about how it can have the longest-lasting appeal and truly strike a chord with its audience. If they believe these changes will do exactly that, then I completely support them, but the rushed nature of the statement and the way teams were reportedly consulted doesn't have me convinced that's the case.

Kelsey Moser is a staff writer for theScore esports. You can follow her on Twitter.

Advertisement