League of Legends esports as we know it turned the corner this year on its first decade of existence. As the scene matures, historical datasets will grow and analysts will streamline their methods. In the long run, statistics should become more an integral part of how the community discusses and understands the game. My hope is that any LoL esports fan who reads this, no matter their math/statistical background, should have an improved picture of how they can use stats on this handy site to help inform their perceptions of player and team performance.

A Long Time Coming: Statistics in Traditional Sports

The year was 1966. Davey Johnson, then-second baseman for the Baltimore Orioles, had a disagreement with his manager Earl Weaver. The subject: Johnson had written a computer program called “Optimizing the Orioles Lineup” to try and convince Weaver that, statistically, he should bat second. Johnson lost that argument, but a later iteration of that same program ended up playing a major role in Johnson’s prolific managing career in both the major and minor leagues. “Optimizing the Orioles Lineup” is an early example of what most people think of as “analytics” in baseball, i.e. a system that takes in-game data and uses them to inform coaching and management decisions. Despite the fact that baseball had been recording statistics to various degrees since the founding of the sport, most of the major innovations in how analytics are used in baseball today happened during the past fifty years. Nowadays, statistics play a major role in how people discuss and analyze the game, and a major reason for that is simply the game’s incredibly long lifespan.

Why am I talking about baseball? Well, it’s mainly to answer the question:

Why should we care about statistics in League?

Now, League isn’t a traditional sport- the game itself constantly changes, and the competitive scene and infrastructure thereof is still in its infancy relative to games like baseball and basketball. However, League has the benefit of beginning in the 21st century- it’ll mature at a much faster pace. Our datasets will get much bigger with another 5-10 years. More players with long, historic careers will retire. Analysts will continue developing better stats to use. We’ll one day have a Hall of Fame, and we’ll need informed ways to decide which players should be in it.

In short, statistics are weapons that grow more potent with time. As the scene matures, they’re likely to become an integral part of the esport, and as such we as fans should begin practicing how to use statistics to help inform our opinions, even at the most basic level. The rest of this guide will include some key tips to getting started, and supplementing those with examples from the past LCS summer split.

The General Categories of Player Stats and What They Tell Us

• The D10’s (not the D&D kind)- GD10, XPD10, and CSD10 are generally used to evaluate laning prowess, especially in solo laners of course. The first two are mostly functions of the third, but they all say slightly different things, so if they’re radically different that can tell you something about a player’s playstyle. For instance, if a player has a high GD10 but a low XPD10, that probably means they make a lot of successful early-game roams (an example of this would be Jiizuke).
• FB% isn’t a terribly important stat for laners, but for junglers it’s a good tool for measuring how effective they are at setting up early snowballs. Note that this is harder to do for junglers on teams with weaker laners.
• DPM and DMG% are good stats for measuring how impactful a player is in teamfights since that’s where the majority of champion damage happens, particularly for bot laners. Keep in mind that you should double-check a player’s played champions over the games you’re looking at- I would not expect a player with seven Ornn games in a season to have a particularly high DPM, for instance, and if they did that would be impressive.
• GOLD% is generally used as a measure for how much “help” a player gets, i.e. what portion of resources they are allocated. Although this is usually seen as a “bad” stat (i.e. you look better if it’s low but your other numbers are high), many of the best players in any league will have relatively high GOLD%. Is that because those players are being propped up by receiving more resources, or because they’re so good that their team wants to play around them? That depends on the player and is best decided by the eye test, but more often than not it’s the latter.
• A note on KDA- This is obviously the most popular and well-known statistic in League of Legends, and a major reason for that is how digestible it is. However, for our purposes, it’s a largely useless stat that doesn’t say very much about a player’s performance. In pro play, it’s largely an indication of how well that player’s team is doing rather than how well that player in particular is doing.

Use Stats to Tell You What to Watch For

Viewers who casually followed the results of LCS 2021 Summer games might not have paid much attention to Golden Guardians given their mediocre record through the first chunk of the split. But about halfway through, Ablazeolive started receiving a lot of eyes when people noticed that he was in the top three among midlaners in almost every category despite being on a struggling team (with the exception being that he had a high death count). After watching his laning performances more closely, most analysts started heaping praise upon Olive after the eye test confirmed that he lived up to the hype.

The lesson here is that stats can be used as a tool to help identify interesting trends to follow when watching games. If you see some kind of statistical trend that surprises you, that’s probably a good sign to pay closer attention to that in game and see whether the eye test confirms your hypothesis.

Use Stats to Double-Check and Contextualize Narratives

On the other side of the coin, we should also use community sentiment to help inform hypotheses that we can then try to verify. A prevailing narrative among the community at the beginning of the summer split was that Jiizuke was a coinflip player, and that he would play too risky and often cost his team games. That seemed to be somewhat confirmed by the eye test, but if you take a look at the stats for summer split he had by far the highest GD10 among any midlaner in the league, as well as having the highest DPM despite spending a good amount of his time splitpushing. The broadcast’s narrative changed as a result of his good performances, and gradually Jiizuke’s community perception improved to the point where he finished 1st team All-Pro.

The lesson here is that while stats aren’t everything, it can still be really valuable to double-check prevailing narratives and see if they line up with the data.

Everything is Connected; Context is Everything

Another important mantra to keep in mind here is that every stat in League is somewhat influenced by many other stats. We can pick stats that try to eliminate as many variables as possible when attacking a certain hypothesis, but at the same time League is an incredibly team-based game and players’ performances are always going to be highly dependent on how well their teammates are performing.

If you notice an outlying statistic or a particular trend, it’s prudent to think critically about what the possible reasons are for that outcome and try to assess the likelihood of each of those reasons using other stats. For instance, let’s take Alphari, who had by far the highest GD10, XPD10, and CSD10 of any top laner in 2021 Summer. Yet TL only won half of their regular season games with him, and at the end of the season (before the playoffs) there was a lot of speculation about why the team wasn’t translating his leads to wins. One hypothesis might be that Alphari wasn’t actually creating his own leads, and instead the team was sacrificing a lot to put him ahead and falling behind as a result with constant CoreJJ roams. However, Tactical still had the third-highest GD10 among bot laners despite CoreJJ being out of lane so much, so that hypothesis isn’t really supported by the data and we’re back to the more nebulous (and perfectly valid) explanation of “the team didn’t execute well in teamfights.” Even though this isn’t some shocking revelation, it’s still worth checking for the sake of due diligence. Imagine if Tactical’s GD10 was bottom two in the league- then we might actually have a case that TL playing around Alphari was costing the team games!

Pitfalls/Disclaimers

There are a lot of ways to mishandle stats, so I’ll just cover some common ones to try and avoid.

• Averages are tricky. If you see a player who on average has +10 CSD10, you might think they’re consistently winning lane. In reality, that same player could be absolutely smashing some lanes and getting destroyed in others, averaging out to a +10. That’s by no means a consistent player, even though at first glance they might look like one.
• Stats don’t always correlate one-to-one with the same trends in game. Take gold share, for instance. Although it’s often used as a negative stat to indicate how many resources a player gets on average, what if you have a player with a high amount of solo kills, or someone with an abnormally high KP? That could actually mean in this case this player just has a higher gold share because they’re straight-up earning more gold for their team than the team would otherwise get, and that’s very worth keeping in mind.
• Statistics can be massaged and presented to paint many different stories. It’s important to always second-guess yourself a bit- how much am I reaching for this stat to support my argument? Am I weighing all the compounding factors appropriately?

Conclusion

The main takeaway here is that careful and methodical use of statistics should always go hand-in-hand with carefully watching games and applying the eye test. Data are great at tracing patterns over time, but they don’t and shouldn’t tell the whole story. Stats are an excellent way to help inform viewers and analysts on what players to keep a close eye on in game, and also an excellent way to double-check and provide additional context towards common narratives that arise in the community- it’s a two-way street. Oracle’s Elixir can be a wonderful tool to help invested LCS viewers grow their understanding of competitive League of Legends, its teams, and its players, and hopefully with this guide doing so will be a little easier.