top of page
  • Writer's pictureBruin Sports Analytics

Debunking the Ranked System: Is Ranked Based on Luck or Skill?

By: Krithik Jatavallabhula


Source: leagueoflegends.com

There have been a lot of talks about changing the ranked system in the popular game, League of Legends. Popular LoL players such as Tyler1 have made complaints about this, especially in a game where his teammate purposefully cost their team the game and Tyler1’s rank went down despite having a good performance. While someone may argue that League is a team game, thus success should be determined by the team result, the particular gamemode I am talking about is Ranked Solo/Duo, meaning that every game, you will have 3-4 new teammates. If this were Ranked Flex, where you can have the same 5-person time every time (each LoL team is 5 players), then it is fair to have team based ranking, since you’re playing with the same people every time, but it doesn’t make sense for solo or duo ranked. As a hardstuck bronze/silver, I like to think that my teammates are the reason I can’t climb, but maybe I’m just not that good. In this article, I’d like to develop a model that determines whether your individual performance or your team’s performance is more important for your rank.


For this article, we will be focusing on four ranks: bronze, silver, gold, platinum. The only rank that is lower is iron, and I decided not to include it because when it comes to a rank that low, a lot of it just comes down to you not performing up to standards. When it comes to higher ranks, there are plenty of higher ranks, such as Diamond, Master, and Challenger. However, for those ranks, it takes a long time to rank up to that level, so it is fair to assume that they deserve it if they get that far. For context, approximately 24% of League players are bronze, 30% are silver, 22% are gold, and 9% are platinum, meaning we are covering approximately 85% of people who play League.


Before going into the data, I will briefly explain what League of Legends is. League is a 5 v 5 MOBA (Multiplayer Online Battle Arena), where the end goal is to get past all the enemy teams’ towers and destroy their nexus. During this game, you will be fighting the enemy team, and your performance against them is recorded in the form of KDA score (kills plus assists divided by deaths) and the damage score (how much damage you do to opponents). While this doesn’t always determine who wins, the team with the better KDA and the team that does more damage most likely ends up winning the game. Once again, there are other factors (towers and dragons, which can’t be found using Mobalytics match history. In order to track more advanced stats, you will have to watch the games themselves) that influence the result of the game. One of these factors that are relevant is Creep Score, or CS. Throughout the game, numerous creeps spawn, and anytime you kill them, you get a certain amount of gold, which can be spent on items that make you stronger (kills and other objectives also increase the amount of gold you get, but the data site I used did not track gold earned). Seeing as CS can make you stronger, it is also fair to say that the team with more CS is more likely to win.





This pie chart shows how many more games are won when your team has a higher CS and Damage differential than the other team. As you can see, more than 75% games are won when your team has the better CS and Damage differential.


Now that there’s a basic understanding of the game, it’s time to go back to the ranked system mentioned in the first paragraph. Essentially, what this article is trying to prove is that your success in mid-tier ranks is mainly determined by your teammates. On top of that, another common complaint in League has been the consistency of LP gains (think of LP as the points that you win or lose after playing a ranked game. Having a certain amount of LP will get you into the next rank or demote you a rank).


Before analyzing the data that I collected from 500 ranked games, I will explain the variables used. The Y variable is the amount of LP gained or lost. The X variables are divided into My KDA, Team KDA, My CS Diff, Team CS Diff, My Damage Diff, and Team Damage Diff. In League, you have a direct opponent, and My CS/Damage Diff is the difference between your CS/Damage and their CS/Damage. Team CS/Damage Diff is the difference between your teams’ CS/Damage not including yours minus the other teams’ CS/Damage not including your direct opponents’.


To start, we’ll specifically look at bronze. In order to figure out the impact of KDA and other factors on your teams’ performance, I ran a regression analysis to see whether your teams’ performance in said factor is more important than your own. For the sake of simplicity, for all the graphs, a blue line will indicate the performance of your team, and a red line will indicate your individual performance. To start, we will see the difference in KDA’s.



As we can see, your team’s KDA is significantly more important than your own in bronze elo, already showing major red flags in the ranked system. In fact, when looking at the slopes, your teams’ KDA is approximately 7-8 times more significant than your own, and both the p values are significant in this case.


When it comes to CS, the graph is as shows:



Here we find something interesting. Despite KDA having a overwhelming preference towards your team, this graph shows that your ability to out-CS your opponent is more important than your teammates’ ability to out-CS their opponents. At a lower rank, this makes sense since many players struggle with having good CS, so being able to get a lot of CS will significantly impact you more compared to other ranks where everybody is able to CS well. As for higher ranks, we have yet to see.


Finally, we have damage in bronze elo.



Similar to CS, we see that it is actually your own ability to out-damage your opponent that matters more. This is essentially saying that if you can out-damage your opponent, that will matter more than if you get out-damaged but your team can out-damage their opponents. Once again, in order to see if there are any differences between ranks, we will have to see how the graphs look for the other ranks.


We will now be moving on to silver. Silver is probably the most important rank to look at seeing as how the vast majority of League players are in silver (including me before I got demoted). The three silver graphs are as follows:



When looking at KDA, there is a major difference between the bronze and silver data set. While the graphs looked kind of similar, the colors of the lines were reversed. This is interesting to look at because in bronze, it seemed that your performance had nearly no impact on whether you win or lose, but in silver, even if your teammates do great, if you don’t play well, it can hurt your LP gains. When it comes to the CS differential, things are a little more interesting. Not only is your teams’ ability to out-CS more important, but according to the p value tests, it seems that your CS differential isn’t even significant. This means that in silver elo, your ability to out-CS your opponent isn’t really that important (this doesn’t mean it’s okay to just not care about your CS, but it’s not the end of the world if you don’t out-CS your opponent). This makes sense, as people in silver can CS better than people in bronze and they are also better at getting objectives (towers, dragons, etc), so CS may not have as big of an impact. When it comes to damage, your individual damage is more important than your teams’ damage differential. With that said, the lines are very close to being the same and it is fair to say that it is nearly an equal effort between your and your teammates’ abilities to out-damage their opponent.



Now we will be looking at the stats from Gold. Gold is interesting, because at this point in your League journey, you are a higher rank than the average player but still have a long way before you reach the top. Let’s take a look at the graphs for gold.



When looking at these graphs, the KDA graph immediately stands out. Given that the silver KDA graph favored your own KDA over your teams’, it was to be expected that gold would follow suit. Instead, the pattern reverted back to bronze, where your teammates’ KDA was a lot more important than yours. This is something interesting to see when looking at platinum. Is silver more like

platinum, or is silver some weird anomaly and your teams’ KDA is more important in general? When it comes to CS and Damage difference, Gold once again was the opposite from silver and matched with bronze, which is definitely a little unusual. However, it seems that the difference between the slopes for CS and Damage are a lot closer than the difference for KDA.



Finally, it is time to look at Platinum. At this point in your League career, you’re probably better than everyone in your friend group, but you still have a ways to go if you’re trying to join an ESports team. The graphs for Platinum are as follows:




Going off the KDA graph, it seems that even in Platinum, your teams’ KDA is still a lot more important than your own. When going off CS and Damage, your ability to out-damage your opponent is more important while your team’s ability to out-CS their direct opponents is more important.






After looking at graphs, some final conclusions can be made. After looking at the data, it is clear that your teams’ KDA is the most important factor to you climbing in ranked. However, CS and Damage differences are very variable in terms of your performance versus your teams’ performance.


Before I make the final conclusions on ranked, I should note the flaws in this study. To start with KDA, what typically happens in a game where your teammates let you down is that you start off with a good KDA, but as your team continues to throw the game, you can’t keep up with their team, causing your KDA to take a hit as a result. There is also another important stat, gold earned that was not used in this study due to it not being available on Mobalytics. In order to get a better understanding of how much of ranked is based on circumstances outside of your control, one would have to collect the stats as the game is going on. Another major aspect that is hard to track is the choice of champion. I used Damage as one of the big 3 factors to determine ranked success, but not every champion is built to deal damage. For example, there are champions that have abilities designed to heal their teammates and there are other champions that have a lot of health and shields and are meant to absorb damage rather than deal damage.


Now to answer the question? Is ranked truly based on luck or is it your own skill? The answer I’ll give is that it does prefer luck, but there definitely is skill involved. Seeing as your team’s KDA is the most important metric, it would make sense that if you want to maximize your teams’ KDA, it is important to have a playstyle where you can help them to get a better KDA. This is where the roles come into play. There are 5 roles: Top, Mid, Bottom, Support, Jungle. Of these, the best ones to play to climb in ranked would be Mid, Jungle, and Support because these positions have the most map impact, meaning you will be able to help more teammates. However, unless you really enjoy playing Top and not helping your teammates (me), it is best to not play Top. Top is a pretty isolated lane where you won’t have much ability to help out your teammates, so it’s up to your team to do well without your help. The important thing is that you don’t stat-pad. Stat-padding is essentially when you play to boost your own stats instead of helping out the team. On paper, it may look like you had a very good performance, but in reality, you probably would’ve been better off letting your teammates get involved and share some of the gold instead of taking it all for yourself (there is a strategy known as funneling, where you do stat-pad, but it is a strategy used in Diamond Plus, as people below that don’t know how to properly funnel). While gold sharing wasn’t a metric I focused on, it is indirectly related to damage dealt as for most champions, the more gold you get, the better items you can get, leading to more damage dealt. This reminds me of a game I played recently, where I was playing Jungle and I was doing really well, but we were losing. The enemy Jungler called me out for stat-padding, and like the toxic League player I am, I made sure to clap back and call out my teammates. But deep inside, I knew they were right: if I just wasn’t playing selfishly, maybe we would’ve had a chance of winning. If I hadn’t stolen kills and instead settled for the assist, my teammates would’ve been able to carry me (you get more gold for a kill than an assist). Based on the KDA and Damage graphs, it shows that it is better for two or more players to split the gold from a kill rather than one person taking it all.


Going back to whether ranked is fair or not, it depends on the role you're playing. If you are playing Mid, Jungle, or Support, then the ranked system is fair, it is just a matter of how you play (play to help your teammates rather than yourself). Especially if you are playing Support, you should not be taking kills, and instead letting your teammates have them. Bottom is more or less fair, and Top is probably the worst position to play to climb in ranked. Maybe I should find a new role.


433 views

Comments


bottom of page