Why everyone is playing mid Ezreal wrong

It's been a while since I wrote anything here, what with being busy coaching a team and all, but what better subject to resume with than one I've ranted about on twitter quite frequently, mid Ezreal. The execution of mid Ezreal comes down to the usage of the ultimate and how well the player understands the sheer versatility contained in that one spell. Unfortunately most players don't understand the tool in their hands, whether you're watching Bjergsen, Coco, Pawn, or any other pro, their understanding of Ezreal's ultimate is incomplete, except for Faker. The brilliance in Faker's play is in knowing not just the varied possible uses of Ezreal ultimate, but also in knowing when to use it selflessly vs selfishly.

To really illustrate the differences I searched all over for someone to compare against Faker's Ezreal and found that pretty much everyone is equally terrible, so I picked Bjergsen because I could. I'm going to use this game against Team 8 but you could easily do this with any other Bjergsen Ezreal game. In this game Bjergsen fires 22 ults, averaging about 1:20 between each one. On top of the frequency of his ults, the other really important thing to look at is where and for what Bjergsen fires the ult. Watching this game you can easily see his use is predominantly mid and many of his ults are simply to push the wave.

Here's what Bjergsen's first 8 ults, those before he loses his turret, looks like:

Rough illustration of Bjergsen's first 8 ultimate paths

As you can see 7 of his 8 ultimates fire down his own lane. Of those 7, only 2 were for anything other than to push the wave and both were defensive moves to try and make T8 back away from his turret.

Now for comparison, we'll look at a game of Faker's mid Ezreal. For the comparison we're going to use this game against KT. In it Faker ults 20 times averaging 1:09 between them.

Rough illustration of Faker's first 8 ultimate paths

To make it an even comparison we're tracking the first 8 ultimates. Unlike Bjergsen only 3 of Faker's first 8 ultimates rattle down the mid lane and only 1 was used to shove the wave. The other 2 own lane ultimates helped Faker pick up kills. Of the other 5 ultimates 1 was used selfishly in Faker's standard attempt to steal the enemy 2nd blue, the rest were fired bot to chunk the enemy duo and help put pressure on them to allow Bang to keep even in farm in a somewhat disadvantageous match up.

After this point the two player's ultimate usage starts to look quite similar if you're not looking too closely. Both mix pushing their own lane with using the ult for fights or objective takes, but even here Faker's use of Ezreal's Ult is a step above. Specifically in the way he pushes the lane with the ultimate, Faker uses it as efficiently as possible. Where Bjergsen will fire the ult just to gobble up a few extra CS as he backs regardless if his teammates are in the area to soak the wave, Faker will clear the wave with his ultimate while farming wolves or raptors, and he is perfectly willing to leave the farm to a teammate that needs it and happens to be in the lane freeing up the ult to check a buff or chunk a side lane instead.

So it sounds like Bjergsen just needs to use his ult a little more selflessly in the early game, and be willing to give up farm to people who can cover his lane (a problem he has on every champion) and he'll be as good as Faker right? Well... this is the part where I admit I've been a devious little sneak. You see, while that Bjergsen game is pretty typical of everyone I've watched on Ezreal, and I watched dozens of players before getting bored of more of the same, the Faker game was the one closest to other people's Ezreal play I could find. I rigged it in favor Bjergsen and other mid Ezreal players.

This game vs Sbenu is what typical Faker Ezreal actually looks like.

When Faker's ahead his ults get everywhere!

He pushes his own wave once, not once in the first 8 ults, once all game. That is his first ult and it's because he's being shoved in and needs to reduce the pressure, after that he never again ults simply to push the wave despite firing another 14 ults this game. What's the big difference here? Faker does not lose his tower. With his outer turret still alive and his team the one with the pressure advantage Faker never needs to relieve pressure on his lane with a quick ult to clear it. Unlike other players, when Faker isn't forced to use his ult to relieve pressure in his own lane, he simply stops considering that a valuable use of the Ezreal ultimate. It doesn't mean Faker ults less though, he still averages an ult every 1:16 this game. Faker simply transfers the ults that other players use to shove the wave into firing to other lanes to chunk enemies or help out in a fight while he's split pushing. Very rarely he'll push a wave with the ult at all and it's always an empty lane to keep minion waves pressuring the enemy while SKT pushes elsewhere.

Most players consider Ezreal's ultimate simply another of their spells meant to help get themselves ahead, but Faker has realized that using a global ranged, vision granting, nuke to farm is a terrific waste. With Ezreal ult you can pressure bot without having to roam, you can force an enemy to surrender farm for fear of a tower dive without leaving your lane, you can let your team 4-1 with the pressure of a 1-3-1, and more. While many players touch on some or all aspects of what it can do occasionally, everyone inevitably wastes ult after ult pushing their own lane, trying to get themselves further ahead, trying to put the game on their back.

Faker isn't playing to get himself ahead, he's playing for SKT to win the game. This is the real key to what sets Faker apart from other Ezreal players. Because Faker is playing for his team to win, not himself, and he has a full grasp of the versatility of the Ezreal ultimate he uses the spell to maintain or increase SKT's ability to pressure the enemy team at all times. As seen in the game against KT, Faker's play to maintain pressure means that when he is pressed in and has lost his turret he is willing to ult the wave to keep the lane stable instead of pushing against SKT. This use makes perfect sense, by keeping the lane from progressing forward into SKT's side of the map he hinders the enemy's ability to press their vision forward, stops roams from mid to other lanes, and prevents the enemy using the lane as a counter-pressure point, all while gobbling up CS to steadily close the pressure gap. However,  as seen in the Sbenu game, when Faker is ahead and able to play forward and farm without needing the ultimate he, unlike every other mid Ez player on the planet, finds any other way he can see to pressure the map. Whether it's ulting an empty lane so it pushes in SKT's favor, adding damage to a fight while he split pushes on the opposite side of the map, or trying to snipe the enemy mid's blue buff, when ahead Faker looks for ways to increase SKT's pressure on the map more effectively than getting 6 more CS slightly faster than he otherwise would. And that is why everyone, except Faker, is playing mid Ezreal Wrong.

Learning From Failure: Team 8's Repeated Baron Throws

Team 8 struggled all game long with vision control and wave management in their first meeting with Team Impulse (TIP). Despite their early problems Team 8 never had reason to worry, their comp was sledge hammer to the face simplistic and at 28 minutes sledge hammer and face met in a perfect storm of press R to win. By layering Malphite, Orianna, Jarvan, and Graves ultimates together Team 8 won a fight 4 for 2 in an instant. The remaining 3 members of Team 8 were more than enough to clean up the bottom inhibitor of TIP after the fight, and while it did cost them an additional 4 deaths being chased after securing the inhibitor Team 8 had still in all but the most literal interpretations won the game. Because of the simplistic power of Team 8's 5v5 team comp, the fact they'd taken the bottom inhibitor, and TIP's inability to reply with an objective of their own, Team 8 needed was well positioned to wrap up this game.

The Problem

To their credit Team 8 shows they understand the basics of the situation, bot inhibitor dead plus superior 5v5 team comp means TIP will have to give up Baron to Team 8. To top it off Team 8 catches TIP's jungler Rush a little too close to the Baron pit and kill him, taking TIP's chance of a smite steal off the board. At this point Team 8 thinks Baron is theirs for free and they try to rush it down.

In their haste Team 8 has failed to account for a couple of important factors. First, Malphite ultimate, their biggest initiation tool, was just expended to secure the kill on Rush. Second, Impact on Gnar was clearing the minion waves in bottom lane before Rush died so Team 8's super minions are not yet in position to force a Baron or Nexus decision from TIP. Finally, Impact left the bottom lane as soon as Rush died, signalling TIP intends to contest during their small window before the super minions begin hammering away at their nexus turrets. 

Team 8 rush to attempt baron, lose 2 people, a tower, and then lose an additional baron buffed member trying to chase stragglers. While it's not the worst result ever it still means Team 8 can not safely split up to get all 3 waves pushing with baron buff, nor do they have an impenetrable back line while Janna is dead. Instead of having an insurmountable advantage and the ability to wrap up the game in the next few minutes, they are relatively easy to deal with, a fact TIP exploits to keep the game alive. 

Fast forward to 38 minutes into the game, Baron has worn off and respawned with Team 8 getting very little for it. Luckily bot inhibitor has been destroyed again, 1 of TIP's nexus turrets is gone, and Impact on Gnar has just recently used his teleport. We are again looking at an unrecoverable situation for TIP, they can't take a straight up 4v5, they can't leave the super minions alone in their base for long, and Impact has just walked away from bottom lane for TIP's last ditch attempt to save the game.
This time instead of starting Baron it's an even simpler mistake. Calitrlolz is up aggressively clearing an unimportant ward in TIP's top side jungle with limited vision and his teammates too far away to back him up. TIP pick off the Malphite and cascade into the rest of Team 8 before they can react. Slooshi whiffs a shockwave intended to halt the engage, Gnar gets a great stun on 3 remaining T8 members, and TIP races down mid to win the game. 

The Solution

Patience. In both cases this game Team 8 has an inhibitor down and super minions pushing on the bottom side, meaning for both of these baron attempts there is a very limited window where TIP can contest without losing their base and the game. Both times Team 8 starts baron with the super minions still outside of TIP's base and Impact having obviously left the lane to join the baron fight.
Team 8 simply needs a bit of patience in both cases. Stay together, keep up the baron dance, but don't start the baron. This would let the super minions push into TIP's base again and force impact to return to deal with them. Without the tanky front line of Gnar there's nothing to lock Team 8 in the baron pit when they do it and no way for TIP to contest without disastrous results.
Patience can be a hard thing for new teams to learn, especially as it's often reinforced that they should capitalize on a mistake the instant they recognize it. Once the lesson is learned though, patience is an invaluable tool that will sometimes mean the difference between a horrible throw and a well controlled finish to a game. 

The Ins and Outs of Contesting Dragon

It's been a little while since I've had the time to make a post so I figured I'd go all out on one. In this I'm going to break down the surprisingly large list of factors involved in a dragon contest, from the basics of vision and numbers advantage, to the more advanced mid lane advantage and wave management. While this post did get a little larger than I'd originally planned I really wanted to make sure i got all the moving parts of a dragon contest packed into here because I so often see some overlooked to disastrous consequence.

The Basics

Dude Potential:

A term blatantly stolen from Phreak, Dude Potential is the absolute basics of contesting a dragon and the first thing most people learn about taking dragon. Dragon hurts, and eats up some of your cool downs trying to rush it, so when you contest a dragon you want to balance out that power loss by being able to bring more people to the fight than the other team can. Advantageous dude potential is gained one of three ways. 

The most basic is to force an opponent to back or even kill them. So long as the opponent you forced away does not have Teleport you have the better dude potential by creating a 5v4 at the dragon. 

The second way in the current meta is to force a situation where the enemy top laner uses his teleport without your top laner doing the same. Once done you avoid fighting long enough for the enemy top to retreat back to his lane to deal with your top's push and then start dragon, again having a 5v4 dude potential advantage. 

Finally you can let the enemy create the dude potential for you when they roam to kill a low health top laner. Generally the enemy jungler, and sometimes their mid, will move top to dive your own top and take your outer turret, if you're paying attention you can respond with a dragon. Since the enemy jungler and/or mid can't teleport to the dragon you will have the numbers advantage and should be able to secure it.

Vision:

Everyone knows having vision on dragon can help you see where enemies are, if they've started it, and avoid baits. However knowing the efficient places to ward, where to expect enemy wards, and how to play around them can give a significant edge when contesting the dragon. 

1. These spots are where you can generally expect blue team to ward when their lanes are even or behind. Clearing the dragon pit and pixel bush are important for maintaining control of the area and making enemy smite steals difficult.
2. These are wards blue team will often place when lanes are ahead. The deep wards on the bottom half the map help in making decisions on burning the dragon or turning to fight. Many people don't recognize the importance of top half deep wards, but a well placed ward can spot an enemy jungler or mid roaming top giving your team the advantage in dude potential.
3. These are the wards generally dropped by red team when lanes are even or behind. The one directly in front of the pit is often overlooked since the introduction of scuttle crab but can make the difference in a close fight. 
4. Deep wards placed by red team when ahead. Again the deep ward in the enemy top half of the map can spot the enemy jungler out of position to give dude potential. Also important to note are the wards in the enemy tri-bush and behind their red buff which can be integral to stopping an enemy smite steal attempt. 

Advanced Tactics

Team Composition and Power Spikes:

This is something that effects the entire game, but as I see even LCS teams sometimes forget to consider their team comp when contesting a dragon it's important to add to this list. What your team comp and the enemy's excel at is important in deciding how to go about each dragon. 

AoE team fight teams love having the enemy team clumped up in the dragon pit. Disengage teams are great at using champs like Azir, Rumble, and Janna to keep the enemy away while they burn down dragon. Poke teams have no business ever being the team to start a dragon when the enemy has any dude potential to contest with. How both your and the enemy's specific team comps want, or don't, to fight dictates much of how you position around the pit, if you start the dragon, if you're just baiting the fight, if you're rushing the dragon and running away, etc...
Similarly knowing when specific champions are at their strongest is important to deciding if you can contest a dragon, and often something forgotten by teams. Even SKT in the first match of LCK(formerly better named OGN) this season got into a disastrous mid-game dragon fight when they didn't consider their Mundo and Tristana need significant time to ramp up. Where your team composition dictates much of how you fight, knowing champion power spikes helps to dictate when you fight. If you, like SKT, have champions who need to ramp up you want to avoid contesting mid game dragons. Or if you're a team who's Rumble and Corki just hit their item power spikes (liandry's and Trinity respectively) then you want to force a dragon soon. 

Knowing the strengths and weaknesses of your champions and how your team comp plays a team fight will tell you when, how, and even if, you should be contesting dragons. With that knowledge you then can extrapolate timing windows for when you need to set up each of the other important factors of a dragon contest to make sure it goes in your favor.

Cool Down Management:

This has already been touched on a bit with your top laner's teleport but it extends to other summoner spells and important abilities as well. Knowing your Rumble's just power spiked with a Liandry's is not very useful if you opt to fight while is ultimate is on cool down. Timing enemy cool downs and communicating your own to your team is another major part of the equation with power spikes and team composition that will dictate how and when to contest a dragon. 

Forcing, Turn or Burn:

If you've compared power spikes and team comps for the two teams and decided you're capable of winning a 5v5 if it comes on your terms you can "force" the dragon, which means you're the team who starts the dragon forcing the enemy into reacting or giving it up. There are two types of forcing which depend on what your team is better at. 

If your team is really good at controlling an area and zoning your opponents away, perhaps you have Azir, Rumble, and Janna, you're likely best suited to "burn" the dragon. Burning means your carries kill the dragon as quickly as possible while those champions skilled at zoning keep the enemy team away. If your zoning team is far enough ahead you can probably engage a fight afterwards but typically you'll disengage after securing the dragon as you'll have spent most of your powerful skills keeping the enemy away. 

The other good forcing team comp is the "turn" comp which is generally a strong 5v5 aoe team comp, something like Orianna, Jarvan, and Maokai together. This team is suited to fighting straight 5v5s and as a result the dragon can be used as bait to create the fight you're looking for. The only real difference between a bait and a turn is that in a turn you've actually started doing the objective with intent to finish if your enemy doesn't contest, where a bait you don't start and are just hoping the enemy runs at you blind. 

Expert

Both of the items in this section, Mid Lane Advantage and Wave Management are often overlooked as the power comes from areas away from the dragon pit and even when done correctly can amount to nothing if a team loses the dragon fight due to a little bit of impatience. I have seen teams at all levels from OCE challenger scene all the way up to the height of OGN be punished by overlooking the importance of Mid Lane Advantage and Wave Management in a contested dragon.  When properly done though, they can compound the lead gained from a successful dragon, or even reduce the deficit caused by a lost dragon contest.

Mid Lane Advantage:

Mid lane advantage is pretty much exactly what the name gives you, it's the advantage gained by positioning your team between the enemy team and the middle lane. This tactic is most important for poke comps, or those who're behind/unsure if they can win a fight, but any team can still get the benefit from it.

1. This is the general position you want to hold to gain mid lane advantage while pushing the other team down beyond the red line towards position 2. From here you have more choices than the other team, you can sit and poke hoping to dissuade the dragon, you can engage the team fight, you can disengage if there's no advantage for you, or you can move up to take mid turret (3) without an easy reply from the other team. This is why you will often see a player shoving the mid wave at the start of a dragon contest, they have the basic understanding of mid advantage and want the wave in position to be used.
   To properly use this positional advantage you must know what your team does best and be decisive in your calls. If you're going to poke you want to lay it on heavy before the other team starts the dragon and be wary of them hard engaging on you. If you're going to turn and rush a turret instead of the fight you need to be patient and let the enemy start the dragon, but decisive enough to turn up the mid lane the instant they commit to the dragon attempt. 
2. This is the more disadvantageous position as it has fewer escapes and attempting to rush position 4 is incredibly dangerous to do. Still being here is not the end of the world when your team comp is suited to it. A strong team fight comp can actually find this position advantageous, using the dragon to bait the other team into mispositioning and a subsequent team fight you'll hopefully win.
   The advantages of this position are heavily team comp dependent, and the mistake I see most often when it comes to mid lane advantage is a team with a weaker 5v5 comp voluntarily taking this position for the dragon contest. Usually it's a red side team making the mistake by trying to rush to dragon and pathing through their own blue buff, but teams on both sides have fallen prey to the mistake. If your team is not better in a straight up 5v5, and you have a choice, you have no business being in this position for a dragon contest. 
3. This is obviously the mid turret, you will often see teams who've successfully figured out mid lane advantage choose to rush down mid instead of fighting a dragon or a baron. These are either teams who do not expect to win a 5v5, and are taking something in exchange for the objective they're giving up, or they've decided the value of the turret is higher than the objective they're giving up. For example a siege team generally will get more value from knocking down mid inner/inhibitor turret than they would securing 3rd or 4th dragon buff. 
4. This isn't actually a good option, I only added it to the map to illustrate why mid lane gives the advantage it does, a question I've been asked by players I've worked with in the past. Sometimes when you decide to rush the turret instead of taking the dragon fight the enemy team will stop dragon and pursue you. If you've rushed up mid (3) to their turret and the enemy team moves to pursue you there's lots of escape paths. As you can see from the lines marked in yellow if your vision is good you can just run straight back to the safety of your own mid turret, you can run into the river headed towards your top turret, or because all the enemy team are near the dragon you can run up into their top side jungle to escape because no one will be there to cut you off. Now compare the escape routes in black from the bottom tower (4). Because the enemy has just started the dragon (2) there are no actual escape routes from the bot turret, you can run straight up the lane but the enemy has a shorter distance to cut you off from the safety of your turret forcing the fight you were trying to avoid, you can run straight at the enemy again forcing the fight you didn't want to take, or deep into their jungle except unlike in mid they're in position to cut you off long before you reach safety. The superior escape routes and extra decision allowed by mid lane advantage is what makes it a powerful tool. 

Wave Management:

This skill is equally important whether you're contesting dragon, baron, trying to take an inhibitor, or just baiting a team fight. It's also one of the harder things to wrap your carries heads around. They've spent plenty of time chasing every possible bit of farm and listening to casters and analysts talk about how someone won the lane because they were up a few dozen cs, and now you've just told them to only kill half a wave and then let a bunch of farm go to no one at all. I'm of course talking about the most basic form of wave management, the slow push. You kill just the enemy caster minions and then leave the lane alone and since your minions die slower than the other team's the wave will gradually grow in your favor until it's an unruly mass that someone on the other team has to deal with. 

A slow push is just one part of it though, equally important are the who and when of building that slow push. If you send your ADC bot lane to build a slow push when baron is a legitimate possibility you very well may have thrown the game off that one single tiny mistake. Similarly mistiming your slow push can result in it actually being worse than if you'd just cleared the whole wave. Set up a slow push too early and an enemy can easily go and collect the massive free farm you've sent them. Start the slow push too late and not only can they collect the free farm but there's nothing pulling their attention away from simply forcing whatever objective you were hoping to contest.
The slow push is not the only form of wave management of course. Things like fast pushing to deny an absent enemy the wave, rebounding the wave so it'll push to you, and freezing the minion wave are also important aspects of wave management throughout the game. However those have far less impact on a dragon contest than the slow push.
Done right proper wave management amounts to having a 6th or even 7th member of the team split pushing for you while you have your team grouped as 5 for the dragon fight. It can force weaker enemies to surrender objectives to you by making them decide between contesting a dragon or backing to save their turrets. It can delay stronger teams long enough for you to scale back into a game by giving them only tiny windows to force a fight before they've got to back and save their own base. Proper wave management, combined with mid lane advantage, can turn any dragon contest into a win-win scenario for your team where you either get the dragon, or you secure multiple mid and top lane turrets.

There are many facets to a dragon contest. While not all of them need to be used to be successful the more advantages you can apply to your contest the more likely you are to come out ahead.

What IEM San Jose taught us about next season.

1. Unicorns Of Love should definitely not be dismissed as a contender in EU.
While they were not able to win the tournament and their play left some things to be desired UOL impressed far beyond expectations. PowerOfEvil and Vizicsacsi had incredible performances beating up on some of the best players NA has to offer. Kikis also impressed with his vast, though a little misguided (see below), array of champions he was able to showcase in the tournament. UOL still has some work to do, especially in terms of quickly turning a large lead into a victory, but given the track record of EU teams surging out of the promotion tournament it's definitely worth keeping eyes on how UOL progresses.

2. Dragon and Baron are far more valuable, but far more dangerous to take.
With the patch 4.20 changes Baron and Dragon went from helpful advantages to out right overpowering in importance. Baron's new buff gives pushing power on par with a downed inhibitor. Dragon gives a myriad of powerful buffs that are generally more valuable than the advantage gained by destroying outer, and in some cases inner, turrets and the 5th dragon buff basically wins the game for you.
On top of being more valuable Dragon and Baron have also had their damage increase making them harder to pick up. Dragon is hard, nearly impossible next patch, to juggle and does significant damage making early dragon plays or contested dragons far more dangerous than last season. Attempting the new Baron with low health or in the face of too many enemies is a good way to not just lose the buff but get your team aced to an enemy all in.
Interestingly you can see how these changes shifted things like champion picks to cope. Because the dragon and baron objectives are harder to secure but more worth it when you do most teams shifted into variations of safe AoE poke comps. Jayce and Corki who were already strong on 4.19 ascended to near god tier, Ezreal rejoined the meta as a popular marksmen, and mid marksmen picks showed up as easy ways to get to ban Jayce without giving up on the Jayce style team comp.

3. The jungle possibilities hype was a bit premature.
This was something I was already pretty sure of watching early tournaments on 4.20 such as Brazil's Razer Challenge or the Black Monster Cup, but IEM San Jose was really the final nail in the coffin that confirmed it. The reign of early game pressure junglers like Lee Sin remains firmly intact. We all enjoyed watching Kikis' unconventional picks like Twisted Fate and Kayle, and we've seen some Warwicks go big, but Meteos ran a clinic on viable jungle picks during the tournament.
Without doubt Meteos was the most influential player at IEM and the MVP by far and it's largely due to his better understanding of viable junglers on the new patch. Instead of being fooled by soloqueue success of champions like Warwick, Meteos tore apart teams with junglers who could get into the lanes and the enemy jungle earlier and cause chaos before his opponent had time to reach their much needed power spike. The dominance of Meteos at IEM San Jose coupled with the continued popularity in tournaments around the world make it clear the reign of Lee Sin King of the Jungle is far from over.

The Myth of the 5 Man Roster

Since it's inception at the start of season 3 the LCS has required teams to list at minimum two substitute players, and Riot currently requires a sub be brought to the World Championships with teams. Subs have seen play numerous times in the two years of LCS. From GGU's multiple sub attempts before settling on Jintae to replace Shiphtur in the inaugural spring split, to SK being forced to use Gilius to replace Svenskeren for their first three matches of the season 4 World Championships, substitutes see play quite frequently in the LCS. Despite the dozens of substitutions required in only two years of the LCS only one team has brought a sub into their gaming house and trained with them as an actual part of their roster.

Most teams in the LCS have a very careless attitude when it comes to substitute players. They treat subs as a check box on the LCS entry forms at worst, and as RoI (return on investment) streambots for their sponsors at best. TSM listed popular streamer Nightblue3 as a substitute player while he posted their sponsor's ads on his stream, but both parties admit he was never going to play in the LCS for TSM and when TSM made a change in the jungle they imported former Copenhagen Wolves jungler Amazing rather than pull up a listed substitute. Speaking of CW, they were last split forced to forfeit a LCS match when mid laner CowTard became ill and CW failed to get the paperwork for their substitute ready in time. When it came time for CW to make roster changes they again ignored their listed substitute Mazzerin, who unlike most had actually made it in for a game, and brought in a new player SorenXD as their new mid laner. When Gambit had visa issues that prevented most of the roster from attending the LCS road trip in England they scouted 4 new players to use as substitutes instead of pulling up either of their listed subs. Examples like this can be found all over the LCS where listed substitutes being used, much less becoming a starter, are an oddity rather than the norm.

Contrast that to other regions where deep rosters are used quite often. Samsung White's Looper is a well known examples to western players as he made his debut with the team at the Season 3 World Championships. LPL's OMG had trained enough with both their support players Dada777 and Cloud to allow them to switch between the players going from group stage to bracket stage during the World Championship and actually see improvement. There's also Chawy for TPA who played multiple games with the team during the GPL season and was reportedly on hand for the Season 4 Championships though not used.

Despite it being quite a common practice in the other World Championship regions the LCS regions have only had 1 single team really attempt the deep roster. That team was Evil Geniuses (EG) back when it was a European team. They brought in Shacker to train with the team and rotate with Snoopeh. The results were a clear success for the team. Snoopeh was fierce in pursuing his starting position again and played some of the best games seen from him in season 3, while Shacker also drove the team forward. Shacker's addition to the team in the summer split of EU LCS helped EG secure a spot in the 2nd-5th tie breaker matches. EG was able to go 2-1 taking 3rd for the regular season. Though EG did fall in playoffs to Gambit landing 4th place just outside a ticket to worlds the success of the experiment is still plain to see in their results.

Though the western scene has for the most part chosen to ignore deep rosters the benefits are plain to see in examples like OMG, and EG. A deep roster of live in substitutes gives teams insurance against whatever random illness or unfortunate life events might sideline a player, with live in subs a game is not simply a write off just because your mid laner can't show up this week. It gives greater flexibility to a team's play style. Subs with different champion pools can rotate in and out depending on opponents to maximize the team's strength, capitalize on a particular enemy weakness, or simply avoid a slump. More than anything deep rosters give more effective training, having another player there to bounce ideas off of, to train specific match ups, and test builds against is an incredible resource many teams go without.

While LCS teams may have mostly ignored the results of the EG experiment change is coming. The next season of OGN will see them move to single team 10 man rosters, a change that if handled properly could strengthen the Korean scene. Instead of splitting focus between two teams with two play styles they'd be able to assemble a ferocious group of 10 that fit seamlessly together and constantly vie for the starting position. Closer to home Team Fusion has commenced another grand experiment like EG's with a 7 man roster living in the house which so far has net them 5th seed in the expansion tournament and a strong chance to make LCS.

How Samsung White built advantages out of nothing.

It's no secret Samsung White is unparalleled as the best team in the world right now. They demolished competition on their way to being Season 4 World Champions this weekend. I and many others could talk at length about the incredible subtleties in their play, in fact I have a few more ideas beyond this one already. I chose this to talk about specifically because SSW were kind enough to do it both from a lead and from behind against SHRC this weekend to really demonstrate how they can create advantages from nothing.

In both games one and four White had their level 1 lane swap matched by Royal Club, matching up 2v2 in the top lane. Both times White used that fact alone to create and advantage by having their duo be the first ones to swap back to the bottom side. Each time Mata bought wards and used the fact they were the first to swap back to safely invade SHRC's jungle for vision. 

In game 1 Mata and Dandy invade SHRC's red buff placing a combined total of 5 wards, 3 greens and 2 pinks. Almost immediately after they finish placing the wards Insec shows himself on the one at double golems and Dandy uses that knowledge to dive mid lane and snowball Pawn's Jayce. Pawn used the mid gank to push down his tower and roam top, Dandy used the vision to out maneuver Insec, and bot used the oppressive vision to safely pressure Uzi in the botlane. 

By the time those 5 wards were gone SSW had a 1k gold lead, significant damage on the mid turret which would fall soon after and had used the vision of Insec to gain vision control over the top side of SHRC's jungle as well setting up another tower dive. Still it's easy to downplay the importance of the early swap from Mata by talking about SSW's advantageous lane matchups and the level 1 first blood that helped them afford so many wards. Thankfully White provided us with an example of the same thing but from a position of weakness in game 4. 

This time instead of a first blood advantage and having pressured the lane before the swap it comes after Looper's Kassadin died to a gank bottom while Imp died in the 2v2 top. Off the back of the 2v2 death Mata backs and again he uses the fact that by being first to switch there's no threat of the enemy bot lane collapsing to invade for vision. He's not able to illuminate it as completely as the previous example but he still gets a strategically placed ward. 

Moments later that ward spots a level 6 Insec passing by to setup for his first Pantheon gank. When he walks past SSW ping him and Pawn cheats down from the mid lane while Dandy clears red. A few seconds after they spotted level 6 Insec heading towards bot White signals clear understanding of what's happening as they ping the bot lane where SHRC's duo is pushing the minions into the tower as hard as they can. Thanks to the early warning of Mata's deep ward Insec is barely able to get off his combo before Dandy's Rengar snarls out of the jungle to turn the gank into a disaster for SHRC, and shortly after Pawn arrives to clean up Uzi and take the dragon with his teammates. 

In game 1 Samsung White used the early swap back to slowly build one small advantage onto another. In game 4 they used the same move to turn an early deficit into a punishing advantage.In both scenarios Mata knew that by swapping first he was safe to invade for deep vision. That vision gave Samsung White the knowledge needed to make plays from ahead or behind, and build advantages from nothing. 

How OMG used champion select to sweep Najin Shield

Despite being a clean 3-0 sweep the quarterfinal series between OMG and Najin White Shield was an absolute treat to watch. Both teams had high points in the games from Shield's early dominance game 1 and Watch's incredible smite steals to Cloud's ridiculous Janna play, Cool's dominating performance against Ggoong, and OMG's decisive calls. Above all else though the series really showed how important a good champion select phase is to victory. Every single game OMG was able to out maneuver Shield in champion select, securing themselves comfort and counter picks while also punishing members of Shield who had smaller champion pools.

Game 1:

Game 1 picks and bans. Source: LoLesports.

Looking at just the bans we can get a clear idea of what each team is focusing on for champion select. NWS bans Maokai, a common meta pick and something Save rarely plays, Jayce a target ban against Cool that will end up being used later in the series, and a priority pick for both teams Ryze. We can see that Shield is trying to make sure they make good use of their first pick advantage. Ryze and Zed have 100% pick or ban rate in Shield's games during the Korean regionals and group stages. Ryze was also a 100% pick or ban for OMG's games in the Chinese regionals and group stages. By taking Ryze of the table Shield intends to win themselves the only contested pick. Considering both Zed and Ryze have been hotly contested by most teams in the tournament it seems like a sound move to ban Ryze and secure Zed, it's also Shield's first big mistake.

Unlike for Shield Zed has a sub 50% pick or ban rate for OMG's games with OMG only playing it twice in 12 games between their regional qualifiers and Worlds group stage. Not only that but when OMG played it they did not first pick zed, they saved it for much later in the rotation. This means Zed was not only not contested, but also something OMG feels comfortable playing against.

On OMG's side they ban two meta bans Zilean and Alistar that no one wants to deal with, and then use their final ban to target Gorilla's Thresh. They then use their own first rotation to secure Janna away from Gorilla. This does a couple of major things for OMG, first it pushes Gorilla off his comfort champions. Gorilla played 11 of the 16 games during Korea regionals and groups on either Janna or Thresh. Secondly, it also counter picks the Zed for team fights, Janna's shield and Ultimate are great tools to help teammates survive a Zed all in and Cloud would make full use of them throughout the series.

Shield knowing they're going to have issues team fighting with the Zed against a Janna build a reasonable 1-3-1 split pushing comp. The idea being that Zed and Kassadin get rolling early and split while the mobility of Kha'Zix and Corki with the peel of Morgana keep them safe. Unfortunately for them OMG also knows this and respond with a counter pick Ahri who can shut down Zed early and get picks on any Shield member foolish enough to try split pushing, and their own powerful split pusher with a much better early game Irelia. This gave OMG enough early power to stop Shield snowballing and enough pick pressure to force Shield into a difficult choice between dying trying to split push or struggling to team fight effectively against a stronger 5v5 comp. The game was actually quite close in large part to some incredible early pressure in the jungle from Watch but OMG's decisive moves and Cloud's vision control kept them in it long enough to overcome the early setbacks and take the victory their champion select phase had crafted for them.

Game 2:

Game 2 picks and bans. Source: LoLesports

In game two we can see clear adaptation in bans from Shield, they realized that while OMG has played Maokai it's not an overly likely pick for either team and removed it from their bans. They've also correctly identified Gogoing and Loveling as major players in OMG's first victory and removed the champions used so effectively by them. Finally they adopted the meta Zilean ban that's basically mandatory if you don't want to deal with a first pick Zilean. This time Shield is trying to stop OMG from getting the advantage with first pick by leaving up a lot of contested champions. They intend for OMG to pick Ryze, Kha'Zix, or Zed first and leave the other two for Shield. It clearly works as they secure Zed and Kha'Zix but again they've failed to account for how much priority OMG gives to those champions. Loveling had already shown he's willing to run other champions into Kha'Zix, and again OMG doesn't put much stock in Zed, the result is OMG gets one of the most contested picks for both teams, Ryze, almost for free.

OMG's bans don't change much switching to blue side. They keep the Alistar meta ban, and continue to target Gorilla's champion pool with Thresh. The one major change is moving Lulu into their bans. mostly done thanks to Shield banning away the Zilean that OMG had banned on red side, but also a reasonable target ban against Ggoong. The Lulu is not an integral part of OMG's champ select here but removing her as a potential flex pick made the Zed even more predictable and removed a counter to the burst they'd eventually build into their team comp.

This time around OMG gets the biggest contested pick of the series in Ryze. Again they counter the Zed pick with Janna and due to Shield's bans this time they bring Rengar for pick and engage potential instead of the Lee Sin. OMG builds a really good all around counter to what Shield is trying to do, countering the Zed all-in, having good siege, incredible burst potential, plenty of disengage, and even a late game Ryze that will make it impossible for Zed to split push. Shield again doesn't have an answer they put together a sort of awkward 1-3-1 comp but it's more of a weird poke siege comp due to the fact Save builds Nidalee as an AP top. The game ends up dragging on longer than it should have due to some incredible smite steals from Watch, but apart from the stolen Baron buffs the entirety of game 2 is as heavily in OMG's favor as the champ select.

Game 3:

Game 3 picks and bans. Source: LoLesports

Game 3 is an oddball honestly, it's the closest Shield comes to actually winning champion select, but ends up in a comp filled with stuff they're not comfortable on. The bans from Shield are good, they remove a lane dominant comfort pick from each of OMG's carries. They also assemble a strong 5v5 team comp with decent laning and excellent late game scaling. If it were any other team I'd say Shield had the edge in champion select here. The problem is, all but Lee Sin are champions Shield rarely if ever plays. Nami is Gorilla's 4th support behind game 2's braum, Ggoong and Zefa rarely played their picks in the regionals or group stages, and Save has neither played Maokai before nor is he known for his tank play. Shield has also again made some champion priority mistakes here, they take Lee Sin as the first pick giving Gogoing Ryze for the second time and failing to recognize Loveling is content to play Kha'Zix into Lee Sin. 

OMG reverted back to their red side bans from game 1 and used the same champion select strategy. This time they weren't able to bait Ggoong into a 3rd straight Zed vs Janna match, but they were able to secure Gogoing's Ryze again and steal away Zefa's Corki he'd played much of worlds on. This resulted in OMG mostly assembling the same style of comp they'd had the first two games despite the Lucian ban. Lane dominant ADC, Gogoing on a comfort scaling champion, play making jungler, and Cloud being an absolute nuisance on Janna. The one big weakness in OMG's game 3 lineup was picking up Zed last pick. The Zed didn't really do anything for their team comp. Something like a Yasuo which Cool had shown before likely would have fit the comp better but Cool seems hesitant to play it after his poor performance on Yasuo in groups.

Despite the fact Shield had the more well rounded comp it was still clear that OMG won the champion select due to their superior familiarity with the strategy and champions their composition involved. OMG were able to secure multiple contested picks, several of which were comfort champions for their star players, as well as force Shield into not just unfamiliar champion picks but also a strategy far different from that they'd attempted to build throughout the tournament. The relative strength of Shield's well rounded team did allow them to stage a bit of a comeback in the middle of the game but it was simply not enough to contest with the confidence of OMG on so many comfort champions. 

Throughout the series OMG showed superior champion select. It's clear OMG did their research understanding that Shield will first pick Zed given the opportunity, and exploited that fact as well as the limited champion pools of Watch and Gorilla to build advantages in their picks. They also made key adaptations like taking Zefa's Corki in response to Shield's Lucian ban to further push Shield out of their comfort zone. It's clear OMG's champion select was much better researched and prepared than that of NWS, and it was a key factor in OMG sweeping the series. 

Worlds Quarterfinal Predictions

Samsung White vs Team Solo Mid

I know I got bit once before claiming TSM had no chance, but I really feel like saying anything else here would just be hopeless NA fanboyism. TSM may have managed to squeak through to first seed in NA but they struggled in their group ending 1-1 with SHR and an SK that had to play one of their matchups with a substitute. Meanwhile Samsung White dominated their group even against EDG who've had SHR's number all LPL long, and SSW are the heavy favorites for the entire tournament. White has demonstrated better team work and better champion select than TSM. White's bot lane out classes TSM's by a monstrous margin, their Jungler is far and away better both at understanding and working from the jungle than Amazing, and solo laners Pawn and Looper are world class players who are unlikely to allow Dyrus or Bjergsen to put the rest of TSM on their backs.

Predicting Samsung White 3-0 the match is the only realistic option to those of us without access to time machines.

Samsung Blue vs Cloud9

Where Samsung White is incredibly strong, Samsung Blue has shown some weaknesses including dropping a group stage game to Fnatic. Some people mistake Blue for the better of the Samsung teams because they've managed to beat White in the team kill matches but they have clear weaknesses in champ select and their early game play. They also clearly have issues dealing with chaotic games as Fnatic and KTA have both proven.

Cloud 9 has managed to earn the honor of being the first NA team to take a game off a Korean team in high level play since 2012 with their victory over Najin White Shield to secure a ticket to the quarterfinals. In their games against Shield C9 demonstrated a clear understanding of the sort of chaotic split push pressure that Blue has proved weak against in the past. Cloud9's players are also not especially out matched by Blue's, Deft is not particularly aggressive in the bot lane, Dade's champion pool is small enough he can be pushed out of his comfort zone, and Balls has successfully proven he's a world class top laner. The big problem for C9 is they've been a bit terrible on red side lately, their Champ select suffers and it negatively effects the rest of their play.

I'm still not willing to call this series in C9's favor, but I think they've got the tools to give Blue trouble and take at least 1 game. 3-1 Samsung Blue.

Star Horn Royal Club vs Edward Gaming

This one is a bit hard to call as there's not a lot of direct confrontations in the LPL to draw from. SHR did have better group stage results but they didn't face an opponent as strong as SSW nor did they openly admit to concentrating on the bracket stage instead of groups. Namei seemed to struggle a bit in group stages as some positioning errors led to a tie breaker game against AHQ, he shaped up in the final game though which is a good sign for EDG.

EDG in the group stage seemed to have better champ select than SHR and they're less reliant on Namei getting ahead than SHR seems to be on Uzi. Also important to remember EDG did beat out SHR in their last meeting to take the 1st seed from China.

I think this one has some potential to go either way but EDG are the favorites given past results and the relative strengths of the teams. 3-1 EDG.

Najin White Shield vs OMG

Shield looked great in their gauntlet run to qualify as the 3rd seed from Korea, but their groups performance got a little spotty. They completely misprepared for more than one game letting major comfort picks like Irelia, Lee Sin, and Rumble through against Alliance and C9. Their champ select disaster resulted in two dropped games, and the one against Alliance ended up being a perfect game at that. Shield also looked very uncomfortable dealing with chaotic split push when they faced Cloud9.

OMG also suffered problems during their group stage games. Cool and Dada777 have not looked very good against their competition. Dada777 especially has become a glaring weakness for OMG with his limited effective champion pool which unfortunately lines up with Shield's star support Gorilla. OMG won't be able to exploit the weaknesses in Shield as well thanks ot their own weak support needing the same picks Gorilla uses to be effective. While OMG did show some ability to create chaos against their opponents they also quite literally escaped groups by a single auto attack.

OMG should be able to put up some fight against NWS but so long as they continue using Dada777 their bot lane will be an ever present liability against Gorilla. 3-1 NWS.

Mid Hecarim, it's not just horsing around

When Keane locked in Hecarim for the mid lane against CLG in game 5 I like many others was shocked. Hecarim was an out of favor melee jungler known for his speedy ganks facing the ranged wave clear machine of Ziggs in lane. After watching the game multiple times and reading up on some lane Hecarim guides though I not only know what Keane saw in the pick, but also where things went wrong for Curse Academy that game.

THE GOOD:

Hecarim has a shockingly strong laning phase. Where most melee champions struggle early on in lane without some form of ranged harass/farming spell Hecarim doesn't seem to suffer the same problems. Keane's starting item choice of boots combined with Hecarim's high base move speed and his passive that ignores minion unit collision made it shockingly easy for Hecarim to avoid most of the ranged harass coming out of Link's Ziggs. On top of his ability to dodge harass Hecarim brings a spammable aoe damage tool to both shove the wave and harass his opponent, as well as self sustain from his W. When put all together Hecarim not only survives the lane but actually does quite well managing to put harass back onto Ziggs and only falling behind in farm from jungler pressure and when roaming.

Lane Hecarim solves one of the champion's biggest issues, gold income. One of the major issues with Hecarim as a jungler that saw him fall out of favor was the gold flow of the jungle. Hecarim needs expensive items like Trinity Force to be fully effective and it takes forever to farm up that much gold as a jungler. Dedicated farm from a lane, helped along by his surprisingly good lane phase means Hecarim can hit those item break points quick enough to be relevant in the mid game. Hitting his big damage items quicker means opponents have less defenses and Hecarim can do the ridiculous 50% hp burst hits like we saw Keane do to Dexter 17 minutes into their game 5.

Even as a laner Hecarim's speed makes for impressive roams. As a jungler Hecarim's speed was a prized part of his kit, allowing him to zip through enemy wards and be on someone before they could react. Though moving from a lane makes him a bit more predictable that speed for Hecarim is still there allowing impressive freedom for the post 6 roams to put the rest of his lanes ahead. In this way Hecarim works a lot like Twisted Fate pick, he farms his way to level 6 as quickly as possible then roams to build advantages elsewhere on the map.

Hecarim is a strong split pusher. His spammable Q and high move speed allows him to shove waves into towers quickly and safely a must for any split pusher. Not settling for just shoving around minions though Hecarim is also capable of putting down incredible burst and chase onto someone caught alone trying to shove back his minion waves. We saw the ridiculous burst Keane put onto Dexter before he'd even finished his Trinity Force. That instant burst potential means Hecarim's split pushing forces a lot of enemy resources to come deal with him.

THE BAD:

Hecarim is about as melee as you can get with only small AoEs on his abilities all of which are targeted around himself. This means Hecarim is not very good at sieging objectives. That can be a problem as the game goes late and you're forced into sieging the enemy's inhibitor turrets, as well as it changes the dynamic of the team comp you build around him. Being very melee means it's easier for defending opponents to land poke and ranged CC to kill you under turret. So Hecarim needs to fight away from enemy turrets, and a team that helps him catch the enemy when they venture out for neutral objectives like dragon or baron.

His late game is nothing to brag about. Hecarim's big power spike is the Trinity Force which sets up his mid game nicely and helps his mixed AP and AD scaling. It also means that he will not be as effective in the late game. He can continue to stack glass cannon and hope to finish out the game quickly while he's still far enough ahead the other team can't build defensive items, or become sort of a utility semi-tanky split pusher and flanking initiator, but either way he'll pale in comparison to late game monsters like Ryze.

Head on team fights are terrible for Hecarim. He's too  fast, his big initiation tool sends the enemies running away from him, and no matter how fed he is he'll never be as tanky as someone like Mundo. That means if Hecarim attempts to simply fight a head on team fight by running straight at the enemy he'll both run the enemies away from his team, and out run his own team, leaving him stranded in the middle of the other team with no backup and not nearly enough effective health to get out. You can deal with that by getting flanking team fights, but you need vision denial and room to move about to get a proper flanking team fight, which leads to the next problem.

When Hecarim falls behind it's hard to recover. It's hard for him to engage in heads up team fights, he needs vision control to either team fight properly or split push, his biggest power spike is his Trinity Force completion which typically comes first, he can't siege very well, and is about as Melee as you can get. All this adds up to Hecarim being mostly useless when he falls behind. He won't be able to split push, engage fights, or be much use in siege defense, from behind.

WHERE IT WENT WRONG:

So we've got the good and bad of the pick out of the way and we can see Curse Academy definitely went to work making use of the good parts of mid Hecarim. By 18 minutes Keane had done reasonable in lane against Link's Ziggs, and used the impressive speed and burst of Hecarim to roam and rack up 5 early kills before even finishing his Trinity Force. The biggest missteps for Curse Academy were actually in champion select. We heard in an All Chat segment that they were determined to play the mid Hecarim since they'd prepared it but their composition didn't really set it up as well as they hoped.

Seconds before going "Woah! what?" Source: Lolesports

Curse Academy built a cheesy speed buff composition to boost the Hecarim but few of their champs were set to spike at the same time as Hecarim and they weren't particularly good at the mid game picks he'd need to avoid his problematic sieging. Instead of trying to build a speed buff comp they'd have likely been better suited building a pick/disengage comp around it. So instead of Zilean we'd want Janna, Nami, or Thresh support for crowd control(CC) and disengage. An Elise pick instead of the Lee Sin for jungle would sync nicely with the pick comp goals as well as make up for Hecarim's sieging a bit, all while still keeping within the new trinity of competitive junglers. The Lulu top is actually decent for a Hecarim pick/disengage comp considering the meta at the time of this game, though I'd have liked a Corki for ADC to better synergize with Hecarim's power spike timing. With our new Lulu, Elise, Hecarim, Corki, Janna/Nami/Thresh comp we still use Hecarim's early roam power to help build everyone to their mid game spikes quicker but there's far more weaknesses than the one CA built. We've now got CC for picks, extra siege potential from Corki and Elise to make up for Hecarim's lack, and the potential for both a 1-3-1 or 4-1 split push where the core group has enough safety from Support ultimate and crowd control, Elise cocoon, and Corki's leap away that they can safely avoid an enemy engage onto them.

Curse Academy's team comp wasn't the only problem in champion select though, the team CLG had assembled was also very problematic for Hecarim. CLG had built a decent team fighting and sieging comp with both incredible late game scaling and plenty of ability to stall the game if it falls behind. With Ziggs, Ryze, and Tristana they had plenty of wave clear in case of falling behind to stall out Curse Academy long enough for those three to out scale the opposition. To make things worse Ziggs' near global range ultimate also meant CLG had strong answers to both the roaming and split pushing potential of Hecarim. Ziggs could use his ultimate to get involved with other lanes without having to keep up with Hecarim's impressive speed and shut down split push attempts with relative ease. When ahead CLG's comp is also excellent in team fights, with an impressive amount of AoE as well as Kha'Zix and Tristana's ability to reset their jumps and keep pursuing with each kill, and sieging turrets thanks to the high range and poke of Tristana, Ziggs, and Kha'Zix. Before Hecarim had even been hovered by Curse Academy CLG had built a comp that was good at countering Hecarim's strengths and stalling long enough to utilize their own.

There were some errors made in game as well by Curse Academy but those were far less relevant to their success in the game than their unfortunate champion select phase. CLG had a comp well built to counter Hecarim's strong points and weather the mid game storm of CA. On Curse Academy's side they not only let through problematic answers to Hecarim they also failed to build a comp that truly took advantage of the champion's strengths.

Worlds Group stage predictions

Group A: 1st Samsung White, 2nd Edward Gaming

Samsung White is the favorite of many for the entire tournament as one of the two best teams from Korea and EDG is the #1 seed from China a region that has placed 2nd behind Korea in most international competition for the last year. Their opponents in the group are the weaker of the two Wildcard teams and the 2nd seed from SEA the region that finished dead last in All Stars Paris invitational. I highly doubt anyone has different predictions for this group. 

Group B: 1st Team Solo Mid, 2nd SK Gaming 

This prediction is probably the one that will get me the most flak. After all I just admit that China has basically been 2nd behind Korea in all tournaments outside IEM for the last year, how could SHRC fail to escape this group with no Korean team? I've been watching some LPL games to prep for Worlds and what I saw gave me a lot of confidence for LCS teams. The Chinese teams have serious problems in their vision and map control and a particularly poor understanding of proper 2v1 starts. If the SK we saw all playoffs, and the TSM we saw against C9 in the finals show up for the group stage matches SHRC will be in a world of trouble. TSM and SK both demonstrated much better 2v1, vision, and map control than anything SHRC showed in their recent matches. SHRC might have a mechanical edge but LCS teams have been forged in a meta that considers kills mere tools to securing a bigger objective. Both SK and TSM have been trained all season to crush the kinds of teams coming out of the LPL, they have the tools in place all that's left is for them to make it happen. 

Group C: 1st Samsung Blue, 2nd Fnatic

Do I even need to explain SGB? #1 seed from Korea, heavily favored, the only team people think could stand up to sister team White. 

The real talk is about Fnatic and much like with SK in group B I believe Fnatic has an edge over LMQ and OMG due to watching recent LPL games. Fnatic has always had incredible team play and strong vision control. They have struggled some this season but that's even more reason to value them above OMG and LMQ as a Fnatic without summer split issues would have been just as favored to win this group due to their map and vision control advantage, without having been forced to improve by disappointing results in EU playoffs. Fnatic has made very smart moves since playoffs ended showing up to scrim with NA playoff teams at PAX and then moving on to Korea where they added Toyz to their coaching staff to prepare for the tournament. What we've seen suggest Fnatic is working out the kinks before the groups start on the 18th and if they do it's hard to see them landing anywhere but 2nd in group c. 

Group D: 1st Najin White Shield, 2nd Cloud9

Apart from Korea being the heavy favorites in this tournament it's hard to deny how good Shield looked as they ran the gauntlet defeating teams like OGN Summer champions KTA and S3 World champions SKT T1 K to secure the final seed from Korea. Shield is not as untouchable seeming as the Samsung teams in this tournament but their recent performances are enough to convince me and many others that they'll be able to hold off the LCS power house teams long enough to take the 1st seed out of group D. 

C9 is hardly a favorite for this group, many people admit they've got a decent shot against Alliance but Alliance is still the favorite of many due to the many times EU has come up ahead of NA in the past. On close inspection though it's clear Cloud9 should hold the edge in the matchup. Meteos and Balls are stronger players than their Alliance counterparts, C9 has had significantly more experience both regional and international as a unit, and Cloud9 is far more decisive in their wins often closing games they're clearly ahead in 5-10 minutes faster than Alliance in similar scenarios. It's these small edges that should be enough for Cloud 9 to make up for what Hai might lack in the head to head against Froggen and push them over into the 2nd place finish in group D. 

Riot's new more aggressive region lock is a mistake

Riot has just announced a new policy where 3/5ths of an LCS team needs to meet "residency requirements" in the region they wish to play. They haven't released the exact residency requirements yet but they make it quite clear it's intended to stop an overflow of foreign players taking over. However foreign players permanently overrunning the LCS is an incredibly unlikely scenario, and this rule to combat that worst case scenario could have negative impacts on potential salaries for the best league players in the world.

The new rule is primarily in place to stop full roster moves, like LMQ and the failed Quantic, as well as stop LCS from becoming what Starcraft 2's WCS has been in recent years with all the regions being filled nearly exclusively with Korean players. That unfortunate scenario for SC2 was caused largely by the fact Blizzard ran the majority of the WCS online. Players could live in Korea, play on the Korean server against other Korean opponents, and simply play the majority of their matches remotely. Because Korea was already established as the strongest region in SC2 the lack of LAN play for the majority of SC2 tournaments meant non-Korean players had to invest significantly more for the same chance to compete. Korean players scrimmed on the Korean servers even when they were participating in the North American region, meaning there was no chance for NA local players to get practice against the caliber of players they'd have to defeat to qualify without moving to Korea. NA players would have to move to Korea and play on the Korean servers just to practice against their potential NA opponents. Because of that disproportionate investment required the scene stagnated and Korea remained on top by virtue of simply having been at the top of the pile from the start.

Back in LCS games are not played over the internet with only a handful of LAN tournaments requiring people to show up, instead they are played live in front of an audience 2-3 days a week for 3 straight months at a time. The fact that the LCS is in its entirety a LAN event means teams and players must move to the LCS region they're participating in. Living locally in the LCS region means they scrim local teams and play on the local soloqueue ladder all against other players who live and play in the LCS region. This means just like how Hai and Froggen elevate their respective regions each time they play soloqueue or scrim another team, foreign players who come to compete in LCS like XiaoWeiXiao do the same for that LCS region. Because the LCS is on LAN when a player is imported to play in the LCS, while they do deny a slot to a local player, they provide more and better practice for local players for the duration of their career in the region.

The pre-existing soft region lock for LCS meant that the influx of foreign talent was a self correcting problem. By virtue of being forced to play against their future competition each foreign player added to an LCS region makes it that much less successful for the next foreign player to move over. In fact we've already hit an early milestone on this progression of the problem fixing itself. So called "C tier" players are no longer viable in NA due to the influx of foreign talent. While super star players like Bjergsen, and Dexter, and "B tier" Chinese players like Vasilii and XiaoWeiXiao have proven NA is not done correcting the problem, other attempts like the failed Quantic team, Amazing, Helios, and Seraph have proven there is verifiable progress towards foreign talents being more expensive than they're worth.

If NA had actually followed the worst case scenario people fear and gotten worse by the early importing of Edward, Bjergsen, and Dexter then subsequent imports Quantic, Amazing, Helios, and Seraph should have been slam dunks to only have to compete against other foreign imports in their lanes. Instead the best of them are middle of the pack NA players. Quantic a team of 5 "C tier" Korean players formed to dominate NA failed to even make the promotion tournament. Helios and Seraph, touted as the conquering Koreans come to dominate NA, shook out to be barely middle of the pack in their respective roles, both outclassed by more than one local NA player. Amazing, TSM's slam dunk don't even need to try him out European jungler import completely collapsed; dragging the entire TSM team down with him for the majority of a split before they could finally train him up to be a reliable asset in time for playoffs.

That leaves Lustboy the newest acquisition for TSM who has certainly out performed former teammate Helios so far in his time in the LCS, and the full Chinese roster LMQ. LMQ actually helps further the point that "C tier" foreign talent is no longer viable as the team contains members from the 6th place LPL team, a member of the S3 Worlds runner up team, and several of the players were rising talents on the verge of breaking out in China. Most of LMQ would likely have found starting positions on teams in China had they stayed. On top of that LMQ while managing to stay near the top in NA has struggled to maintain that position all split long sharing top 2 with no less than 3 other NA teams prior to playoffs, and only barely made worlds as the 3rd seed. LMQ is another strong indication that NA is moving up in overall skill and becoming harder for foreign players to enter and find success. LMQ's move to NA LCS also had the added benefit of directly raising the strength of the challenger scene where most other imports only effected it tangentially through Soloqueue. The split LMQ spent competing in the challenger scene made the teams they faced like up and comers Team 8 that much stronger, and helped forge local rising talent Altec into a top ADCs in the NA LCS.

Recent imports make it readily apparent that high profile expensive moves to nab the best foreign talent will soon be the only way to make importing players more valuable than simply training up a rising star from the local challenger scene. In essence the imports would have to return to what was seen originally, pulling in players widely regarded as the best up and coming talents of their regions like Dexter and Bjergsen where when acquired, or competing away an integral part of one of the strongest teams the region has ever seen like with Edward. These big expensive moves due to the fact they're pulling the top tier highest profile players would be both rarer than the recent rash of imports, and far better for players in all regions. When teams from all over compete to secure a top player of a region like Froggen, not only does Froggen benefit but so do all other players. Froggen gets a huge check to secure him against theoretical offers from the likes of CLG or Curse, and players like Bjergsen get to say "hang on a minute, I'm nearly as good as Froggen and get payed 1/3rd what he does! Pay me more or I'll find someone who will." The expense of securing the best players in the world helps all the players get a better salary, and has the added benefit of further stunting foreign talent imports.

Now this is where the new rule moves from jumping the gun on a pointless rule into a full blown mistake. The drive and desire to secure super stars like Froggen, there by helping to increase theirs and other players salaries world wide, is now severely impeded by this rule. Because no LCS team can have more than 2 non-residents many of the teams with the deepest pockets are taken out of the bidding before it even begins when a must have like Froggen becomes available. Before this rule if Froggen's contract were to expire post worlds and he was still seen as the best in EU he could expect offers from most of the wealthy LCS teams in both regions. However after this rule is in place in the same scenario TSM, CLG, LMQ, and EG would all be incapable of putting forward offers due to their number of non-residents already on  the teams. This means competition for a star player's contract is lower, and as a result what he can expect to receive as a final offer is likely to also be lower than it would be if multiple additional teams were allowed to participate in the bidding.

It's understandable that Riot wants to calm people's fears of another SC2 situation arising but they are fighting against a non-problem with a rule that actively hinders the growth of LoLesports. Still Riot has decided to take action and lots of people are calling out for something to be done, so here's s way Riot could solve this problem without damaging the growth of the scene.

Simply make this rule apply to teams who enter LCS from the challenger scene. Rewrite the rule to require that any team participating in the promotion or expansion tournaments must have a minimum of 3 players who will meet the residency requirements by or before the start of the next split. The rule will function much like the age requirement for challenger teams, it halts full rosters trying to make it in with low investment into the scene, and it doesn't remove some of the richest teams in LCS as potential bidders for the biggest superstar players. It's still unnecessarily fixing a non-problem like the newly debuted rule, but it does it in a less damaging way.

The current iteration of Riot's new rule is a mistake that fixes a problem that will mostly self correct, while impeding growth of the scene. I am certain that if LCS continues it's current growth rate that within the next few seasons (not splits) Riot will be forced to significantly alter if not completely do away with this rule as it not only becomes obsolete but a more significant impediment to the growth of League of Legends as an esport.

LCS Playoff Predictions Review

Now that LCS playoffs are over I'm going to go over my EU and NA quarterfinal predictions as well as those from Gamescom and PAX. I'm going to discuss how I made those predictions, where they went right, where they went wrong, and thoughts on each region's Worlds qualified teams.

The first big thing to talk about when discussing predictions is that it's not something you can ever expect to get 100% correct without a whole lot of luck involved. Everyone from myself to Montecristo to players who'd been scrimming the teams in the finals got some predictions wrong. Because of the limited information we all have available, and the ever present question of player's mentality day of, as well as effectiveness of training and adapting to the patch it is very hard to make anything more than a semi-confident prediction.

While predictions can never be flawless, there is a lot of information available to help improve predictions. I looked at past history between the two teams, accounted for champion select tendencies, as well as strength of team rosters and their specific play styles when making my predictions. When it was a very difficult one to call, like Fnatic vs Alliance, or Dig vs TSM I gave preference to the team which in my opinion had a stronger peak performance with their current roster. In some cases peak performance failed, like in the Dig vs TSM, but in others it helped me properly weight the matchup, as in the case of ROC vs SHC.

EU

SHC vs ROC Prediction: 3-1 ROC, Actual Result: 3-1 ROC

This series went mostly as expected. The one thing I did not expect was SHC to be capable of rallying for game 3, rather than taking an earlier game off and falling to momentum. 

SK vs MIL Prediction: 3=0 MIL, Actual Result: 3=0 SK

I have to admit to a mistake in predicting this series. I didn't give enough weight to how MIL had performed in the final week, a mistake I made multiple times throughout my predictions. I undervalued just how badly MIL had performed in week 11 vs how well SK had performed that same week. Accounting for their difference in results should have adjusted my prediction to a 3-2 in MIL's favor. That prediction still would have been different from the results, but it would have better fallen in line with how the teams had performed throughout the season and accounted for their results only a week before their series. 

SHC vs MIL Prediction: 3-1 SHC, Actual Result: 3-0 SHC

I feel pretty good about this prediction. While I didn't get it exact I correctly noted that if either team went 2-0 up they'd sweep and correctly gave the edge to SHC. I didn't really adjust predictions going into this series, though I likely gave MIL too much credit in spite of their recent performances. 

Fnatic vs ROC Prediction: 3-1 FNC, Actual Result: 3-2 FNC

I'll be honest even as a Roccat fan I never expected them to put up such a wild fight against Fnatic in the semi-finals at gamescom. Overall I feel good about the prediction, Roccat showed up and pushed Fnatic further than most people thought they could and they deserve praise for it, but I think it would have been a bit fanboyish of me to have blindly predicted such a crazy series. 

ALL vs SK Prediction: 3-1 ALL, Actual Result: 3-1 ALL

Nailed it. 

ROC vs SK Prediction: 3-2 ROC, Actual Result: 3-0 SK

This is one of those series that makes it clear why predictions just don't end up being 100% correct without a lot of luck. Roccat has had better champion select than SK for a long time, the teams share a lot of the same problems avoiding either team having obvious significant edges, and ROC had just proved against Fnatic that they were on point at Gamescom and should have pushed SK to the limit. While ROC may not have taken it in the end this series really should have gone all 5 games. Instead Roccat appeared to have spent everything just trying to make it past Fnatic and when they failed they fell apart for the series against SK. 

ALL vs FNC Prediction: 3-2 FNC, Actual Result: 3-1 ALL

While Alliance did hold the head ot head 3-1 over Fnatic this split it felt in large part thanks to Fnatic's All Stars slump at the start. Fnatic was simply not showing up the first few weeks in a lot of peoples opinions. When Fnatic did finally turn it on near the end of the split they were hands down playing on par with Alliance, while having a better champion select phase. Fnatic's superior champion select and the fact they were defending 3 time EU LCS champions weighed it in Fnatic's favor. Instead just like Roccat, it seemed Fnatic was forced to burn everything they had just to make it to the finals against Alliance. I feel like this was another example where a prediction fell apart due to team mentality day of more than incorrectly assessing past results. 

NA

TSM vs Dig Prediction: 3-2 Dig, Actual Result: 3-1 TSM

I feel like not weighing Dig's final week results into this is less of a mistake than it was for MIL's games, but still a definite mistake. I think if I'd retained that lesson from the EU playoffs I might have weighted the series 3-2 for TSM instead of Dig. 

Curse vs CLG Prediction 3-1 CLG, Actual Result: 3-0 Curse

Unlike with Dig I don't think weighing in CLG's performance issues would have been a smart move for this series considering they hadn't been seen since week 10 and spent nearly a month in Korea. By all measure this series should not have gone the way it did. The last time CLG spent a few weeks with coach Montecristo in house they had a dominant performance and sat near the top of the NA LCS standings. Add to that the fact they were practicing in Korea against better scrim partners and it's difficult to reasonably argue that we should have expected CLG to lose, much less fall apart, this series.

Dig vs CLG Prediction: 3-1 CLG, Actual Result: 3-1 Dig

Again I think it would have been a bit unreasonable to predict differently here. Dig's singular working strategy against TSM succeeded thanks to a lack of aggression from TSM, something that CLG had not lacked against Curse even if it hadn't created a win. Dig's champion select also made progressively less sense against TSM where CLG seemed to actually be pulling themselves together a bit over the course of the series but were let down by their actual play and decision making. 

LMQ vs TSM Prediction: 3-0 LMQ, Actual Result 3-2 TSM

Another good example of day of performance making predictions a bit of a crap shoot. TSM did not look any better than they had all split long against Dignitas and LMQ had had their number every time they matched up. I'd hazard to say anyone who predicted this match in TSM's favor after watching TSM vs DIG is guilty of fanboying a bit. The only vaguely reasonable reason to predict difficulty for LMQ was the ownership drama they'd recently been through. Even that didn't seem enough to so heavily set them back though. 

C9 vs Curse Prediction: 3-2 C9, Actual Result: 3-0 C9

Cloud9 is probably the only team I didn't give enough credit to in playoffs this split. Curse had been surging like crazy, there were parallels to other similar team styles matching against each other. I thought that plus C9's continued struggles throughout the split was enough to make it a close match. I did not give the C9 dominance enough credit, peak C9 showed up this series when I didn't think it would and it rolled over Curse. 

Due to there being no break on the weekends and me writing my predictions before hand I didn't make actual predictions for TSM vs C9 or LMQ vs Curse. I didn't expect TSM to come out ahead of LMQ. On twitter though I predicted LMQ taking it over Curse and C9 3-0ing TSM again due to TSM still not looking better against LMQ. I was very happy to be wrong about C9 vs TSM. All credit must be given to TSM for finally looking like a team that could contest for 1st place against C9, both teams played very strong matches and TSM came out the victor.

So there you have it, how I made my predictions, where I think I went wrong, and where I think the unpredictable factors in League made a difference. I intend to make similar predictions for Worlds once more information on matchups are released and hopefully the lessons learned from looking over my predictions this way will mean more accurate predictions in the future.

NA LCS Playoff Preview: PAX

Fifth Place

Counter Logic Gaming vs Dignitas

This is definitely not the 5th place match I was expecting. I predicted both of these teams to win, because both have had the stronger peak performances this split and that is often an important factor when teams get ample time to prepare for a series like this. That same method is how I correctly predicted Roccat beating out SHC in europe, sadly Dig and CLG did not live up to expectations. I will say that Dig came much closer to what I predicted from them than CLG did though.

After watching the quarterfinals this game became rather hard to call. Dignitas had some good team comp ideas but their execution was abysmal. Even in game 1 their split push efforts worked more due to TSM's indecisive shot calling rather than anything specific Dig did, they lost most of the 4v4 split fights but weren't punished for that by TSM. All 4 games Dig had poor team fighting and abysmal vision control against TSM. The main redeeming quality for Dig, apart from Zion's singular adventure in split pushing, was Crumbzz's undeniable baron dominance even going so far as to win a smite war against a Nunu.

CLG on the other hand completely collapsed in their first outing back from Korea. Their game 1 team comp made no sense, running the binary, lane dependent Nunu with 3 losing lanes was doomed from the get go. It only looked vaguely competitive because Curse was struggling to hit their own stride in game 1. Game 2 CLG again seemed to have no idea what they were doing giving Curse repeats on the important parts of their team comp and then giving up multiple early game dragons to set Quas behind while Cop and Voyboy were allowed to become fed monsters with no contest. It wasn't until game 3 that CLG finally appeared to wake up finally banning and picking okay but it wasn't enough as Curse out maneuvered them lvl 1 to get a 3 buff start for Nunu and then compounded their advantage via a misstep from CLG around dragon to close out the series in dominant fashion.

Both teams leading into this are shadows of their former selves and absolutely deserve to land in the 5th place match from the way they played that series. I think objectively though, the edge goes to CLG just barely. While CLG was definitely not playing optimally in their series, they were definitely decisive in their movements. A lack of decisive calls was the only thing that allowed Dig to take a game off TSM and CLG is unlikely to make the mistake of letting Zion split push unanswered. I expect the series to go much like their regular season head to head, 3-1 in favor of CLG. 

Semi-Finals

LMQ vs Team Solo Mid

While my quarter finals preview may not have correctly picked the winner, what I was right about was just how insanely close the series would be between TSM and Dignitas and that to me spells disaster for TSM against second seeded LMQ. 

LMQ did not lose practice from someone renewing their visa, has had a better grasp on new patches, has shown much stronger champion select throughout the split, and they have not dropped a single game to TSM so far. Some people claim LMQ has little experience with best of series but that's also not true as they played several in the challenger scene on their road to the LCS and never once dropped a series. The worst LMQ has performed in a best of series is dropping singular games to the 2nd best challenger team in NA at the time, and it was always obvious over confidence that made them slip up in the one game they lost before tightening right back up and obliterating the remainder of a series. TSM's "historically strong playoff performances" are also utterly irrelevant to the conversation as Wildturtle and Dyrus are the only members still on the team that have seen more than one previous playoffs with TSM. All in all, after watching TSM's sloppy performance against Dignitas the only vague hope they have against LMQ is to pray the recent ownership kerfuffle around LMQ has caused the players significant issues in practice and mentality. 

I really don't think TSM has much hope outside of picking up a single game from LMQ over confidence. I'm calling 3-1 LMQ. 

Cloud9 vs Curse

I will admit despite their 3-1 finish to the season I gave Curse very little credit going into playoffs, but when you consider they had only 5 wins in the first half of the season (14 games!) I'm less crazy than I probably seem now that they've 3-0d into the semi-finals.

Before Curse showed up against CLG I'd have called this series 3-0 for Cloud9 and moved on. Even if C9 has lost a bit of their dominance this split they've retained the exquisite champion select and strong team play that keeps them on top. However C9 has exploitable weaknesses in their early game, a place Curse has been quite strong this split, and more than one of C9's losses have been them getting so far behind early that they simply couldn't catch back up.

We also have an interestingly similar situation recently happen in OGN. KTA a notoriously early game team came up against Samsung Blue a team with weaker early game made up for with strong team play. In the OGN version (SPOILERS!) KTA took the match by forcing it to a game 5 blind pick where they've always had an edge. The NA LCS is now giving us the Hollywood remake in C9 vs Curse, but where KTA pushed to a game 5 blind pick Curse can't so easily force an advantageous situation for them. Cloud9 also does not suffer from the occasional champion select weaknesses SGB has exhibited this season.

If Curse digs down deep and pulls out all their tricks I would not be surprised to see KTA vs SGB 2 Electric Boogaloo C9 vs Curse go 5 games, but I think in the end Cloud9 will still come out on top. I predict 3-2 C9.

Third Place

Team Solo Mid vs Curse

Despite a different team winning than what I predicted, TSM vs DIG still went very much how I expected, super close for no good reason. Meanwhile Curse played better than expected against CLG despite a shaky game 1 and were able to close it out 3-0 which gives me a lot more confidence in this match for Curse.

Throughout the split TSM had this match in their pocket. TSM was the one team Curse just couldn't seem to get a handle on, no matter what they brought to the table TSM walked away with the W. That was until the final week of the regular season. In their 4th and final match up of the regular season Curse absolutely demolished TSM in a 32 minute game. Curse moved from that result to shaping up fast against CLG 3-0 and are now moving into the final weekend of games at PAX looking absolutely on fire. With their strong early game now turning into commanding performances from Cop and Voyboy it seems likely that if Curse can keep their heads in it after losing to C9 that they will dominate a TSM that still hasn't changed after 32 games.

Where Curse looks on fire TSM looks stuck in a rut. Amazing still does not look comfortable outside of his Lee Sin and Elise, dropping a smite war as Nunu and securing zero of the first nine barons against Dignitas in their quarterfinal series. Dyrus seems mostly stuck in the same old champions from the middle of the split reverting to Lulu for most of the games against Dig, and the inclusion of Lustboy has not done much to get Wildturtle out of his slump.

If Curse can deal with the disappointment of losing to C9 it's hard to see how this series can end up in any result other than 3-0 (maybe 3-1) in Curse's favor.

Finals

Cloud9 vs LMQ

Despite Cloud9's struggles that started with Hai's hospitalization, and the hype around Chinese imports LMQ, Cloud9 actually holds the upper hand the head to head 3-1.

Both teams are top of the pile right now and it's easy to see why as their rosters hold a lot of the top 2 players in the region for each position. The real difference between the teams is in how they play. LMQ is unbridled aggression, they ward ahead of them, not behind, they dive your tower, they come at you and they kill you. That same aggression is also sometimes their down fall as it sometimes leads to tunneling on a poor dragon fight or a bad tower dive giving up what advantage LMQ may have gained. Cloud9 is a whole other beast, they like controlled purposeful strikes. Fights with a purpose are the only ones C9 desires, if they can't turn a team fight into a dragon, baron, or tower they're just as happy to fall back and let you pass. 

So far Cloud9's controlled style has won out giving them the lead in the head to head, which means in this series the real decider will be the team's ability to adapt in a best of 5. As of this writing C9 has never been pushed to a 4th game by any team in NA. They have gone to Playoffs #1 seed twice before and have dropped zero games in playoffs. Their record is not as spotless in international tournaments, but they have still proven their ability to change and adapt even outside NA.

LMQ has faced plenty of series before, unlike what many people claimed. However LMQ's multitude of series in the challenger scene is nothing compared to the level of competition they can expect to face against Cloud9. Even against challenger scene teams LMQ dropped some games due to tunneling on Baron or Dragon too hard. LMQ has proven that a bad loss does not rattle them, and that they can adapt in a best of series, but they will have to prove they can keep their play tight and champ select strong against the defending champions.

This series is difficult to call due to how strong both teams are and the differences in play style between single games and a best of 5 series. I think whoever no matter what whoever is the first to go 2 games up will take the series, and due to their head to head I think that team will be C9. I predict C9 3-1.