Pros and Cons of New Champion Select

As someone who has played close to ten thousand ranked solo queue games across four seasons, I think that the champion select changes fundamentally undermine the concept of a solo rating in two ways: Being able to queue up to play only 1-2 roles Alright, first: this is a GREAT idea for normals games. It's also a good idea in a bottle - as in, people should be able to see what roles others want right from the get-go in solo queue. However, if the matchmaking for ranked automatically matches 'compatible' teams, then the very concept of a player rating is thrown out the window. What you have instead are five separate ratings with minimal coupling for each of the five pre-determined positions of the metagame. This is because the same people will be playing mid against the same group of people, and so on for each individual role - there will not be any motivation to learn new champions or roles, and there is actually a disincentive to 'fill' now: if you select 'fill', you may end up playing your worst role, while if you simply select your two best (even if you can play the others), you're guaranteed to play one of your two best roles. It's a no-brainer and there's never a reason to select 'fill', even if you're already an incredibly well-rounded player.

In other words, you may as well create separate ladders for each individual position if you do this, because the effect will be exactly the same. People will only be competing with others who main the same role(s) as themselves, and there will be no point of comparison between players who play different roles because the distribution may be larger or smaller depending on the role (for example, the player pool of support players is probably much smaller than, say, mid lane, and so the statistical variance of the ratings of support players will be artificially increased in order to compensate for this).

What you'll end up with is a fundamentally different rating system from the current one, which in my opinion is a bad thing. The current system is quite similar to that adopted by chess, and is one of the main points keeping veteran players coming back year after year. You will no longer be measuring an individual's overall skill but rather their ability to play one or two specific roles and/or champions. This also has a few other broad-ranging applications: People will have no incentive to 'play nice' in champion select, or in general work with their team towards team compositions in champion select. The system has already done all the work and, in fact, validated the opinion of the player who currently types 'mid talon or afk/feed' in the ladder today. On today's ladder, such a player will see their rating be decreased as a result of their toxicity and inability to play other roles; meeting other players with similar champion pools results in near-guaranteed losses. However, with the new system, they don't even have to say it the system has already allowed them to say it in advance with no negative consequences to their rating. As mentioned previously, there will no longer be one ratings ladder but rather five separate ladders thrown together, which will result in five normal distributions overlaid on top of one another. Each ladder distribution will have different widths and peaks as a result of having different numbers of players who can handle each current ELO bracket. For example, let's suppose that the average ELO of an ADC main is higher by some amount X than the average ELO of a top lane main. Under the new system, the top lane main will artificially have his/her ELO boosted by X relative to his/her ADC main counterparts, as there will no longer be any real interaction effect between their skill levels, since they main different positions. Additionally, roles such as support, with relatively fewer players, will see much larger variances between the top-rated players and the next tier of rated players, because their playerbase being smaller means that that must necessarily be reflected across the entire player distribution.

In summary on point 1: Implementing this for normals is a no-brainer, and implementing a view-only version for ranked makes perfect sense. However, matchmaking based on these preferences will re-define what a 'rating' in league of legends means to what I personally feel is a much worse standard (sending the message that people should only be playing 1-2 champions to achieve their highest possible rating).

Being able to queue up with multiple players Firstly, I am aware that duo queues already exist, which is already an undermining of the idea of getting rated based on one's own skill level. I personally dislike duo queues for this reason, but their effect on any individual game is not so suffocating as to be a significant negative impact on the ladder. Triple or higher queues, however, would be. Let me explain why: One's rating will no longer be dependent on one's own skill level. This should go without saying, honestly. If you let players play for 'solo' rating with friends, then who they are friends with becomes a component of their rating. I personally feel this goes against the very concept of rating individuals, and does nothing to enhance social interaction between players. My reasoning is simple: players currently get matches with those of similar skill level, ie those who they are most likely to want to make a team with. Were players' ratings to fall less in line with their overall skill, suddenly you'll see players who were doing decently but not 'hard-carrying' be neglected in favor of said 'hard-carrying' players, as the difference between the two will be more extreme and the benefits to cozying up to the perceived 'strongest' players one plays with will be much more tangible. Suddenly if you can find and latch onto a strong player (or three) there will be direct benefits to one's own rating, which is not a healthy environment. Currently, if someone friends me after a game in which I did well I can presume that that person actually wants to be friends with me, either for the purposes of future team building or discussion with someone of a skill level that he/she respects. While this can certainly still happen, there will also be people who friend others with the sole intention of piggybacking as much ELO as possible off of others. Solo players will be at a distinct disadvantage. This also should be obvious. Matchmaking will not always be able to find triple queues of players in a reasonable amount of time, and so mismatches will happen with some frequency. The result will be that solo players will, say, 5 in a hundred games, simply be at a large disadvantage that is incredibly difficult to overcome. The result long-term on a player's rating will be substantial (probably ~two divisions or so), and this is assuming average performance from each player in a triple queue. Typically players can queue with those who they work particularly well with, which is also not supposed to be taken into account by a rating system (ie the ability to work well with certain select groups ought to be a marginal aspect of one's rating, as it would be in a true solo ladder system); an example would be a Yasuo main who queues with a Lulu main. Currently, the two could duo, yes, but they would have no guarantee of getting either (much less both) of their roles, and even when they did would still have to deal with the possibility that the enemy jungler applies a lot more early pressure than their (for the sake of argument) jungle Warwick. Now take that scenario and extend it to two players who only play Rek Sai and Alistar. Suddenly, those four players, unable to play any other champions, are allowed to pick those champions every game without fail and perfectly complement one another, while the other team will (sometimes) be composed of five random players who've never played together before. The result will be a gross overexaggeration of rating of the aforementioned four players, simply because they happened to find one thing that they do well with others and abused it in an environment where counterplay simply isn't likely or even possible in most cases.

Of course, the scenario I described above could happen in a 5s queue, but that is a completely different environment. In 5s, people can see team compositions unfold before their eyes and play counter matchups, as well as bringing their own plans to the table. Such countermeasures are not available in solo queue.

The current system allows large groups of players to test how well their skills line up in coordination with each other on the ranked teams ladder - in other words, there is already a ladder to test teamwork in existence!

One final problem with 3+ queues is that in such a queue, an individual player being matched with them could face the possibility of a triple (or larger) report if they simply disagreed with one of the mass-queuers. I shouldn't even need to describe what a toxic environment that is for a solo player, being constantly at risk of getting mass-reported if their teammates don't like them (or if they have a poor game while playing with a quadra queue). Of course this is simple to fix: if in a queue with X players in it, simply make each report worth a weight of (1/X) times a normal report. This can be solved easily from a report standpoint, but of course it would still be a horrible experience to play with four people who queue together and blame anything that goes wrong on their fifth, a solo player who will necessarily not be able to communicate as well with them as they can with each other (and so is less likely to do well in any case).

The purpose of solo queue is to test an individual's skill level in a bottle. This is the most pure form of it and it is what keeps competitive players coming back moreso than any other aspect of league of legends. Both of the aforementioned changes fundamentally undermine the very concept of rating individuals in a ladder, and the result would be that the solo queue ladder would effectively no longer exist. Instead, an entirely new set of parameters would be measured and represented as 'individual skill', parameters which have more to do with who you know and how specialized you are than your own general skillset.

Lest I sound too negative, let me conclude by saying that I love all the changes mentioned for the normal queue matchmaking ladders. Those tiers are intended to be as laid-back and enjoyable as possible in order to provide a good place to practice. Ranked, however, is a competitive environment which certainly is fun, but for which 'fun' is not the operative concept - competitiveness is. People play ranked because they want to prove and improve their individual skillsets.

The changes proposed for allowing 3-5 picks to do bans, for allowing people to hover champions, and for allowing people to select preferences (and have those preferences be shown in champ select) are universally beneficial and big upgrades to the current system. I would suggest thinking long and hard, however, about the ramifications of the above changes to solo queue.