“There are three kinds of lies: lies, damned lies and statistics”.

This quote, popularized by Mark Twain (albeit of unknown origin), is something to keep in mind whenever statistical analysis is consumed. While numbers can often point us in the right direction when properly used, they can just as easily mislead us into unreasonable conclusions. Not only are statistics never a valid substitute for actual analysis, but they can often tempt us with oversimplified narratives that rarely hold up to scrutiny.

This might very likely be the case of Rick Fox’s League of Legends squad. Even with a 12-17 record, Echo Fox holds the highest gold lead at fifteen minutes of the entire NA LCS. On the whole of the LCS, this is only beaten by the much more successful H2K that has, in Jankos, one of the best early game junglers in the world.

It is easy to see how this feeds into the narrative we’ve been hearing about Echo Fox’s early game. Quoting prominent Riot caster Phreak, Echo Fox is “the best early game team in North America”. By focusing on average gold at fifteen as a measure, the idea that the team’s lack of success was a result of them not knowing how to abuse advantages arose in the community. However, the numbers do not agree with this idea.

To better understand whether this claim is valid, it seems reasonable to break the games between wins and loses, in order for us to get a better insight into what is really causing these loses.

On won games, Echo Fox looks nothing short of fantastic. Averaging a humongous 2.83 thousand gold lead at the fifteen minute mark, Looper’s new team dominates every aspect of the early game. What’s likely the biggest gold lead in wins on the entire LCS is not enough for them, as they seize more than 3 out of 4 early game dragons, more than 4 out of 5 first towers and average almost half a tower over their opponents.

We are starting to get an idea on how the narrative of this team having the early game figured out can quickly permeate within the community. By not separating analysis between wins and loses, we get an overinflated perspective of their average advantages simply because of how dominantly they win.

This, tied to how fantastic they look in their wins, makes it so that little analysis seems to be left in order to declare Fox masters of the early game. However, Fox wins overpoweringly and loses moderately. By simply staying in the surface of averages, we deny ourselves deeper understanding.

On their loses, Echo Fox sports no statistical measurable advantage as the fifteen minutes close. With an average deficit of 0.23 thousand gold, they reach only for 1 in every 3 dragons and land exactly at 1 in 4 first turrets. They even lag slightly behind on turrets taken. No numeric value suggests them being set up for success on their loses.

Where is the narrative of closing problems coming from?

The reader, paying as he is close attention to this piece, knows that statistics are not to be trusted so easily. Everything in the previous paragraph is true, yet there are reasons to believe Echo Fox has problems closing.

On their first week of LCS, each and every one of their loses fits the narrative perfectly. Reaching every mid game ahead, Fox still managed to lose four out of five matches, resulting in two series lost and the birth of a myth. Nevertheless, it would not be honest analysis to leave it at that, as there is a very strong case for the first week being a statistical outlier.

The squad, keeping only two members from the previous split, had more than reasonable causes for this performance. A new Korean member, combined with the fact that teams were refusing Echo Fox scrims based on behind the scenes drama, is quite the perfect recipe for a bad mid to late game.

Not only is this the part of the game most based around communication and coordination, it is also the biggest quality difference between challenger rosters (Fox’s forceful scrim partners) and LCS teams. Unpunished mistakes and poor communication amount to a quite disastrous mid to late game practice, but this seems to have been patched in the following weeks’ matches.

When keeping the first week out of our data, a clearer picture of the actual early game of losses is obtained. Each and every one of the previous stats becomes strictly worse, with gold expectancy at fifteen going as low as the negative 0.75 thousand mark. Closing hardly seems to be the problem here.

In this new set of games, Fox still throws a game were they held a significant lead, but also comes back from another were they were significantly behind.

The narrative of reaching mid game ahead and not knowing what to do with leads vanishes, yet a more interesting question arises.

How can Echo Fox win so dominantly yet lose so frequently? A clue hides in this next graphic, waiting for the impatient reader in search for answers. Unfortunately, the larger point is still in need of development.

If there is one pattern to the North American squad’s early game, it's that they do love ganking. While their opponents sit barely over 2.4 in the combined of their series, Echo Fox averages 4 ganks per played game. Win or lose, ahead or behind, Fox is always on the hunt for overextended enemies and poorly lit lanes.

This result is not only a consequence of the magic of averages, as only Immortals hold a positive gank record over them and only by one additional gank. Outside of Immortals, Fox strictly and absolutely dominates every other team in the league on amount of ganks in the combination of the games they’ve played.

Their predilection towards heavy ganking is shown not only by numbers, but also by their choice of supports. Heavily favoring strong roamers and gank assisters, Gate’s four most picked champions are, in this order: Malzahar, Tahm Kench, Zyra and Camille. Gate’s roams to the mid lane plays a huge part in the total of Fox’s ganks. Fans of the team have become used to seeing Keith on his own, barely holding the lane.

This choice of playstyle is definitely not unfounded. Out of every gank they perform, Echo Fox expect to obtain 0.42 kills, while their opponents average an expectancy of 0.37. This speaks volumes of their gank prowess. Not only are they performing 70% more ganks per game, but they are also get more out of each one.

 While their drafts often favor ganking and roaming more than those of their opponents, they are still getting more average success with far less time to set up each play. Both data and observation support the fact that they are extremely proficient gankers.

Unfortunately, this might very well be all there is to their early game.

Fox’s boys fall flat on their asses whenever their draft does not allow for heavy roaming and unpredictable ganking. Recalling the earlier graphic and tying it to this new one, we have solid reasons to suspect that both gold and gank efficiency at fifteen minutes see a slow but steady drop as weeks go by.

This is suggestive of a flat style that is getting figured out by opponents, as it is plain to the eye that their mechanics are not dropping. When ganking is all there is to your early game threat, it does not take long for players and coaching staff to take notice.

Even more, the speed at which Echo Fox is getting figured out might be increased by their problems with other teams. When people are denying you scrims, it is not out of the picture to imagine teams sharing pieces of information to keep Fox from succeeding, rather than saving them for the competitive edge.

Perhaps the strongest indicator of their gank-relying engine is Froggen’s winrate with Anivia, his most played champion this split. The best Anivia in the world is shamefully sitting at only one win and seven loses with his key champion on the 2017 Spring Split.

As a team, Echo Fox clearly struggles to play around the cryophoenix’s strengths, or rather, her weaknesses. Her reduced mobility heavily impairs both her roaming and snowball potential from the mid lane. With an immobile midlaner, Echo Fox find themselves with two of their most common gank paths completely shut out. Because we are in a tank meta, Froggen is the only superstar on the team with snowball potential, and his bird is clearly not playing Fox’s game.

The queen of stalling has her strongest cooling effect on Echo Fox’s kill expectancy per gank, which heavily drops from 0.42 to 0.24 whenever Anivia hits the Rift, almost halving itself. Perhaps trying to make up for the lost efficiency in ganks, Froggen and his team increase the average amount of ganks performed, but this is far from a good idea.  

It is self-evident that an immobile mid laner makes ganking into a pricier choice. It either results in less resources for Froggen or more time drained from teammates, all to make up for the midlaner’s sluggishness. Even when they reach advantages, snowballing is much harder for the Freljord native than it is for most meta midlaners. The result is usually the same and unavoidable -- less gank efficiency combined with more time invested in ganks results in heavily impaired midgame results.

We are, however, forced to note that Froggen’s Anivia heavily shapes their opponents game as well. Being that Anivia has low mobility and no escape mechanisms, ganks by opponents increase heavily, almost matching Echo Fox’s aggressiveness. While Froggen’s prowess and years of expertise having his bird camped drop the opponent’s expected kills to an average of 0.29, the effect is not nearly as strong in their opponents as it is on Fox’s own players.

As our conclusion approaches, the reader should recall -- or perhaps revisit -- the average gold graphic presented at the beginning of this exploration. There is a sort of beauty to how the conclusions were hiding in plain sight, only waiting to be discovered.

We’ve already established that Echo Fox is unadaptive and can only play a gank heavy style. This is backed both by the graphics showing a steady decrease both in average gold and gank effectiveness as well as observation, but the most important question is still hanging.

Why are their wins so incredibly dominant?

“There are three kinds of lies: lies, damned lies and statistics”. We never accounted for variance! As we sat here talking about averages and exploring expectancies, we never once addressed the huge difference existing between each week’s data.

Reality cares very little about expected values. If you are ganking like a madman, you will unavoidably have games were every gank lands. In these situations, your team will look like a world beater and get heavily ahead. Little expertise is needed to take a drake or tower if the opposing laners are dead.

The opposite is also true; some games, you will miss every gank, and your team will surely and quickly lag behind in farm. The enemy team can snowball this advantage as well and you might see yourself never recovering from your deficit.

The picture is now clear and complete. Fox dominates the games it wins and lags moderately behind when it loses because when ganks are not working, the team can simply stop ganking and play at a disadvantage. This serves to control the bleeding and since almost nobody ganks as heavily as they do, the snowball will be consistent, but with smaller peaks.

Because stopping ganks is not a possibility for the enemy team, when Fox gets ahead they absolutely smash their opponents. While we don’t have a clear measure on how good Echo Fox’s early game is relative to their peers, we can assure with certainty that they have the most feast or famine early game on the entire LCS.

A complete understanding of Echo Fox’s early game will never be achievable for us as viewers, yet our understanding grew with the exploration each layer we’ve dug into. Even if not apparent, many wrong turns were also avoided, and perhaps an imaginative reader can recall forming a picture in their head that later data in the piece disproved.

Whenever someone uses one or two stats to prove a point, always remember that there is a correlation between the decreasing number of pirates and global warming. Numbers and statistics are not themselves a path, they are a flashlight, and with every new item lit another shadow is cast.

If you enjoyed this piece, follow the author on Twitter at @Cabramaravilla.

Images courtesy of Leaguepedia, League of Legends Wiki, Dot Esports, Live.echofox.gg, LoLEsports Flickr and Openculture.com