League of Legends: What we don’t know about the Meta-game

11 min readAug 19, 2017

Knowledge in League of Legends

League of Legends players, as a whole, actually know a lot about the game. There are quite a few community driven websites that centralize knowledge and allow players to learn more about the game. We can distinguish these websites depending on the nature of the knowledge they provide: game knowledge or meta-game knowledge.

You may be unclear about the differences between the two terms. After all, meta-game isn’t exactly the clearest of terms. So, what does “meta-game” mean? Well, meta-game refers to the way the game is played. Thus, meta-game knowledge can be summed up as knowledge about the way people play the game, as opposed to game knowledge, which is knowledge about how the game works. If you are still unclear about what the differences are, imagine the game is poker. Poker game knowledge is knowing that a flush beats a pair, and poker meta-game knowledge is knowing that players fold 70% of the time at the flop if they have anything less than ace and 7 (this is completely made up, I actually have no clue about how to play poker).

These are two fundamental aspects of knowledge about any game and while knowing how the game works is essential to be good at the game, knowing how players actually play the game is essential to be great at the game. For instance, take poker again. With perfect game knowledge, you’d be able to correctly estimate how strong your hand is. This, in turn, would allow you to bet the best amount based on how likely you are to win. However, while this would be enough to beat inexperienced players, pro-players would be able to take advantage of your betting pattern to tell whether you have a good hand and if it’s worse than theirs, allowing them to easily trounce you. That knowledge of how players behave is meta-game knowledge. The same holds true for League of Legends. Knowing that you shouldn’t buy a Rabadon’s Deathcap on Zed is some basic game knowledge that will allow to win over newbie players, but knowing when Zed players usually roam will allow you to win against much more experienced players, just because you can read them and close their opportunities.

What we already know

Now that we’ve clearly established the importance of game knowledge and meta-game knowledge. Let’s a quick horizon tour of what is currently available to the community.

Game Knowledge

Nearly all of the inner mechanics that compose League of Legends has already been decomposed, analyzed and centralized into a few knowledge sources.

If you want to learn more about how the game works, the best website to learn is, by far, the League of Legends wiki. This community maintained website covers almost all the possible game knowledge you could ever think of, and is nearly always up-to-date, even including information about future patches. You can find information about the following bits of knowledge over there:

It’s basically a gold mine of game knowledge. However, if you are more interested in future changes to the game, surrender@20 is probably your best friend, with precise information about PBE changes, future skins, and balance changes.

Finally, if you want a more under the hood look at how things work inside League of Legends, the developer platform for League of Legends also contains quite a few tidbits of knowledge that can be useful, like how big the map is, or how spell tooltips are actually written, and what the actual internal statistics of items and champions are.

Meta-game knowledge

So the knowledge base about the game is massive and nearly complete. Does the same hold true about meta-game knowledge? Well, not really. There are a few sites that provide some knowledge about what is played or how it’s played, but the overall information is fairly limited as we’ll see.

Of course, some basic meta-game knowledge that can interest the general player base is available. For instance, you can find information about which champions are most played, and their success rate on champion.gg, which is a website I cannot recommend enough. It is a true treasure of analytics geared towards champions. There, you’ll find information about win rate, play rate, temperament matrix, damage composition, player base experience distribution, which summoner spells players usually take, skill orders, build orders, masteries, runes, and even matchups for each champion in each role. It truly is a great learning tool. A competing tool would be op.gg, which is great for other reasons. Op.gg provides players with information that is more centered about themselves, by highlighting their recent performances and their best champions. Naturally, as they gather a lot of data about players, they also do aggregated stats about champions, and the statistics page of op.gg is in many ways complementary to the ones found in champion.gg, by sometimes providing more detailed information, in particular in the match up section. However, this game knowledge still revolves around champions, and what people do with them.

An attempt to break away from just champion analysis is heralded by LoLKing, which also provides players with a look into which items are the most performant at the moment, but also by doing some general analysis about spells, duo lane synergy and lane matchups in general.

Finally, Oracle’s Elixir gathers a great deal of data about pro games. There you can find a ton of information about many, many elements of the game, like when do pro players take herald, or the first turret, etc. This analysis is, however, restricted to pro players.

Missing Pieces in Meta-Game Knowledge

Most of the meta-game knowledge seems to be centered around which champions to play, when to play them, and how to play them, with some partial analyses of other elements that revolve around champions, like summoner spells, or items.

However, once you move away from champions, the meta-game knowledge available online practically vanishes. Even the most basic questions, like how much gold does a player earn every minute is not readily available. And there are so many questions that could use an answer. Let’s take a tour of interesting questions and why they matter.

Gold Per Minute (GPM)

GPM tells us how much everyone earning over time. This metric is fundamental, as it depends on how certain roles are played. For instance, take the support role. Supports sacrifice their personal income so that AD Carries can safely earn more gold per minute. Consequently, you’ll see clear differences between the GPM of supports and ADCs, which are due to how players value each role. Now imagine that in patch 18.0, the impact that ADCs can have on the game is drastically reduced, this could yield significant changes the priority that ADCs take over supports, and we’d maybe see both an increase in support GPM and a decrease in ADCs’ GPM. So looking into GPM can tell us how players value each role.

Additionally, GPM is probably not constant over time. It is indeed very likely that each role earns less early in the game compared to later on, as it’s obviously harder to clear waves or camps at low-levels. Analyzing these changes in GPM over time could give us information about which role is more prominent at any given time. Who is the richest at minute 5 can also give information into which role is the strongest at the point in time.

Finally, knowing GPM gives you information into when players usually have enough to purchase certain items, which is an indirect way of looking at the timings during which a build is strong and weak.

So looking at GPM, in general, gives us information about how players value each role and into the power spikes of build orders, and looking into its evolution gives us information into the evolution of strengths and weaknesses of each role over time.

XP Per Minute (XPPM)

XPPM is the experience counterpart of GPM, and like GPM, it is essential to understand the strength and weaknesses of each role. There are a lot of asymmetries between roles when it comes to experience. the duo lane has to share the experience of a single lane, while solo laners get the totality of their lane. However, due to difference in the length of the lane, Mid and Top do not level up at the same time, because it takes more time for minions in the side lanes to reach the center of their lane compared to the midlane. Does this difference affect the time at which solo laners level up? Is this difference something that can be used? What about Junglers? Their source of XP is entirely different from the laners’. Does this change anything about who is strongest when? From observation, we know that junglers get to level 3 earlier than other roles. However, there’s no clear number about this. Does the jungler have a 30s window during which he’s ahead, or is it more 1 minute? Knowing the difference between the two can mean a lot in terms of success rate for junglers.

Looking into XPPM can also help us understand when each role reaches level 6, which is a critical level in terms of power spike. Knowing when a role is level 6 while others are not can give clear information into where that role should go to take advantage of its power spike, as well as how long that role has to act on that power spike.

So XPPM is a critical element when looking at early level strengths and weaknesses.

Jungle Kills and Deaths

Another element one can look into is where do Junglers get kills or die over time. Indeed, this can inform us about the general patterns junglers use when moving across the map. This, again, is fundamental knowledge. Not knowing that the enemy jungler is near you is likely to result in your death. Additionally, looking into where junglers die can give us a look into which lane look deceptively weak. This is also meta-game information that one can use, either as jungler — to avoid those death traps, or as a laner to bait opposing junglers.

Simply put, Jungle kills and deaths provide us with insights about where junglers usually apply pressure, both correctly and incorrectly. Knowing about this pressure is key to a successful game plan.

Objectives

And if kill pressure is a key element of a successful game plan, correct objective focus is probably even more critical. And there are so many unanswered questions that revolve around this. For instance, how strong is the focus shift when the first dragon is an Infernal Drake? What does that entail in terms of warding timings? Are dragons more heavily prioritized compared to turrets? What about the Rift Herald?

Knowing how players value different dragons and how the type of the first dragon changes their objective focus is knowledge that can be exploited to correctly read an enemy team by applying the correct amount of pressure in the right place at the right time.

Finally, knowing in which order players tend to prefer to focus objectives can be used to anticipate map movements and prepare traps.

Back Timers

Another type of map movement whose knowledge can be exploited to yield better results is back timers. If you know when a lane usually backs, you can adapt your own backs and come to the lane right before they attempt to back, which is the time where your item advantage will be the greatest. Knowing when each role backs is also helpful in understanding how much of gold they usually have converted into items at every point in the game, as well as how much they can spend on each back. This can be the key to understanding critical weak points in a build order. Indeed, suppose that some laner only need 900g for his next purchase, while his counterpart needs 1300g. This 400 gold discrepancy naturally transforms into a back-timers discrepancy, which can be abused by both the laner and the jungler to apply pressure until this extra 400 gold is obtained. Therefore, using this knowledge and by comparing the build order of two champions, you can find clear windows during which your ally is stronger than his opponent. And you can do this before the game even starts.

Vision Gaps

Knowing when and where players ward or when and where they don’t is also an element of gameplay that can be used to one’s advantage, for obvious reasons. And knowing how long control wards usually last can also be used to measure how costly of a purchase that really is when compared to other wards.

Filling the gaps

Overall, there are so many elements of meta-game knowledge that could benefit players that go unanswered that it’s almost frustrating. We most often have a general idea of what the answer is, thanks to our experience, but we rarely have hard numbers on which to rely to transform this intuition into real knowledge. Resolving this issue is the purpose of this series of article.

In order to do so, we intend to apply statistics and mathematics to a few thousands of matches in order to get valuable insights into each of these topics. Naturally, the first thing that we’ll need is to actually get the tens of thousands of matches that we require. For this, we’ll look into how the Riot’s developer APIs work and how we can use them; then we’ll look into how we can construct a tool to get the match data that interests us, if such data is accessible, and finally, we’ll explore these themes one at a time and give an in-depth answer to it.

This is why this series of article is structured as follows:

What we don’t know about the Meta-game (this article)
The Treasure Chest of League: Riot APIs
Building a League data scrapper
Gold, XP and Role
Jungling, Kills, and Deaths
Mid Game: Objective Control
Back Timers
Warding and Vision Gaps
Putting it all together: Jungling Paths

I hope that this series of article will interest you, and I recommend that you start directly with the “Gold, XP, and Role” article, as the articles preceding it are mostly technical information about where to get the data, and how to build the tool to get that data. Happy reading and bonus picture!