Friday, August 17, 2012

I.10d - Take two - military expenditure


My previous post was lengthy and I was wrong about a very critical fact about mitigration scaling. Thanks to Phreak for setting me straight.

TL;DR Both health and mitigation scale linearly; prefer health early on because it helps against all incoming damage sources. Itemize not only based on your opponent's comp but also which of your opponents is doing well.

In depth:

I previously stated that the mitigation becomes less efficient over time, but I was wrong. Both health and mitigation scale linearly. 100 armor provides you with 50% mitigation. If you have 1000hp it will take 2000 physical damage to kill you. 200 armor gives you 66% mitigation, meaning it will take 3000 physical damage to kill you. Every 100 points of mitigation gives you 100% more effective health from that source.

I maintain that health should be favored over mitigation because it allows you to absorb all 3 types of damage (physical, magic, true). Purchasing 100 armor does nothing to protect you from an ap carry, so building out a solid base of hp is a great way to soak damage from both sides and adding mitigation contextually.

To demonstrate the principles that effect your build order, I'll discuss jungle Amumu and cover two scenarios. The enemy team has a balanced comp so I will account for both types of incoming damage. I always start with Philosopher's Stone and Heart of Gold. The passive gold helps my late game itemization, the Philo allows me to give blue to the ap, and the HoG establishes early health for tanking. I use 0-21-9 masteries, armor yellows, and mr/level blues, move speed quints and magic pen reds.

Ad carry is doing awesome, ap is meh:
  • Randuins or Frozen Heart. The Frozen Heart is very popular right now, mana, cooldown reduction, armor, and a debuff aura are very compelling. Note however that these stats will do nothing to protect me from the enemy ap. If I expect incoming magic damage, the Randuins will protect me from both sides and give me the aoe slow.
Ap carry is doing well:
  • Aegis/Negatron. The aegis gives me everything a tanks need and is extremely cost efficient. With  health from the HoG and Aegis, a Negatron will substantially mitigate incoming magic damage. I would likely go back to buying armor at this point because the utility of the big armor items tends to beat out the negatron items. 

I've found it very challenging to succinctly describe the driving forces that optimize your itemization because there are so many variables at play. Watching pros is a good way to start building a body of knowledge around when to build what. Pay attention to what's happening in the game, watch the order in which items are picked up, and actually talk through the decisions with yourself (or a buddy) to determine how wise the choice was. It's worth noting that the mathematically you can calculate the maximum effective health for a given amount of gold, however this is of limited value because in reality you are limited to six item slots and items give you specific stats that may not follow this curve. Additionally for legendary items you are paying for things other than the base health and mitigation stats so it's difficult to say what item will treat you best based on simple models.

Sunday, February 19, 2012

I.10d - Method and Discipline - military expenditure

10. By METHOD AND DISCIPLINE are to be understood the control of military expenditure


Spending your gold wisely is probably the most discussed topic in League of Legends. Guide quality has improved drastically because the pros have gotten involved. This post will discuss the mindset behind making the most out of your builds rather than answering specific questions about what to build in a particular role. While some champions have fairly stable build paths, builds should always be adjusted according to the context of your current game. On the flip side, there are core principles that should inform your decisions and not be compromised, regardless of the situation you are in.


Guides have been improving in quality and the greatest improvement is that builds often consist of core items as well as situational items. One topic that gets less attention however is what order you buy items, even while trying to finish the same core build.


Tanks need health and mitigation to be effective. The most common mistakes players make is to build mitigation before health. As a quick disclaimer, while I am a numbers guy, I am not a huge fan of number crunching in LoL because so many factors are related to human behavior, which is very hard to quantify. 

Now I'll illustrate how you can maximize the value of your build even when building the same core items. 
[edit] This portion of the post has been rewritten because it was too long and I had some bad data. The followup is linked below.


A few core principles that you should consider while you adjust your build.

Ad Carry: You need carry damage, finish two damage items (iedge + pd) before building an expensive mitigation item (banshee's). If you are struggling, go dorans blades and wriggles to add some damage as well as survive.
Ap Carry: You need carry damage, if you're forced to buy mitigation consider an early Null Mantle. Go for RoA or Rylai's rather than doing something crazy like building a Banshee's before you get damage. A Catalyst and Negatron would be the absolute tankiest I would ever go, this still allows you to finish a RoA and keep working towards real damage.
Tank: Health is king, a 1000hp singed with FoN isn't really doing that much for you, really.
Tanky DPS: Aim for items that give you damage and mitigation, Wriggles, Wit's End, Frozen Mallet and Atma's are excellent at giving you great value for tanking and dps'ing.
Support: gp/10 is critical, always. Shurelya's, Aegis, and other aura items are legit.


Maximizing the value of your build is complicated. Experiment, study, watch the pros, and you'll increase your itemization efficiency in no time. For the follow up on this post, check out Method And Discipline Take Two.

Saturday, February 18, 2012

MiG vs Startale - Game 2 - OGN Invitational

My second pro meta post will look at a game played at the OGN Invitational in Korea between Maximum impact Gaming (MiG) and Startale in late January. The Korean teams have come out of the gate strong in the competitive scene, especially MiG who beat CLG.na and won the invitational.

MiG Bans:
-Ahri
-Nocturne
-Shaco

Startale Bans:
-Rammus
-Kassadin
-Graves

Because I don't know these teams well, it's tough to say whether any of these bans are targeted towards particular players. Rammus, Ahri, and Kassadin, and Shaco were all standard bans in high elo solo queue at this time. Locodoco for MiG called out Shaco specifically as a must ban, which is interesting because he was not considered a threat at Kiev. NA pros have stated hes meh in competitive because he's a snowball champ,  careful teams will starve him a leave him useless late game. I'll speculate that the renowned Korean mechanics are responsible for Shaco being feared even in a pro setting. Nocturne is another champ who wasn't picked or banned in any of the later stages of IEM Kiev, demonstrating further distinction in the Korean jungle meta. Graves was a strong pick/ban during Kiev, matching well with that ban.

Note: Picks are listed based on lane position, not pick order. The lane comparison will discuss matchups and cover counterpick situations.

MiG Picks:
Kennen
Brand
Skarner
Corki
Sona

Attributes: Poke, Ap, Squishy, Aoe, 4x cc
Tons of damage from Kennen, Brand, and Corki with good bruiser damage from Skarner. Double ap in solo lanes plus ap from Skarner, leaving Corki to bring the ad. Good range with Brand, Corki, Kennen, and even Sona. Also great engage with Kennen, Skarner, and Sona plus Brand/Corki followup, huge aoe. Great all around composition, fairly squishy but because of great engage and cc they should be fine, Kennen could pick up Zhonya's to guarantee a full ult. Strong lategame scale.

Startale Picks:
Morgana
Ryze
Maokai
Vayne
Alistar

Attributes: Ap, Tanky, Single Target, 5x cc
Solid damage from Ryze, Morgana, and hypercarry potential from Vayne. Also running double ap solo but Vayne dishes out a lot of ad and true damage. Okay range, Morg's puddle can clear waves and Ryze q is decent. No champs can poke, requiring a committed engage. Tanky with Mao, Alistar, and Ryze plus Morgana's shield. Excellent single target damage with Ryze and Vayne plus a huge amount of cc, team can leverage this by catching champions out of position. Very strong lategame scale. Comp has all the right pieces, however without any poke it's difficult to control space. This makes positioning for map objectives and sieging towers difficult.

Top Lane: Kennen manages to beat Morgana in cs because she's not puddling carefully and roams a bit too much. Poor execution by Morg.
Summoners: flash/teleport for both
Builds:
-Kennen goes WoTA/Rylai's, sustain and health with solid damage but not pure hate
-Morgana goes tanky with early Catalyst into RoA. Goes for Rabadon after this but is very far behind.

Mid Lane: Brand and Ryze tie on cs, good job by Ryze to keep up.
Summoners: flash/ignite for both
Ryze: Standard tanky mana build, Tear of the Goddess, Catalyst, Glacial Shroud. Adds Negatron for more mitigation.
Brand: Double Dorans into Rabadons, had to get Mercs.

Bot Lane: Fairly even on cs early, Corki's poke is effective at keeping the aggressive lane at bay.
Summoners: Corki/Sona have both aggressive summoners, while Vayne takes cleanse. Cleanse counters exhaust however this leaves her open to cc from Sona or Skarner.
-Vayne: flash/cleanse, Alistar: flash/ignite
-Corki: flash/ignite, Sona: flash/exhaust
Builds:
Vayne: Bloodthirster, Zeal, takes Zerks (Berserker Grieves)
Corki: Dorans, Vamp Sceptor, Triforce, adds QSS (Quicksilver Sash). Takes zerks

Neither team takes cv, preferring summoners to help in the lane. This leaves open level 1 bush ganks which startale takes advantage of, picking up first blood in the bot lane. This also leaves the top lane a lot more exposed to ganks which gets MiG in trouble a few times.

Jungle: Both junglers focused on ganking, made possible in part by the lack of cv. Very little counter jungling, Skarner helped control dragon and pushed lanes, picked up an early oracles.
Summoners: Both take flash/smite. Great for Skarner ult, allows Mao some extra aggression to chain with his root.
Builds:
Mao: Philo, HoG, Giants Belt and Negatron, takes Mercs. Pure tank, does what he does.
Skarner: Philo, HoG, Aegis, Sheen. Takes Mercs. Mostly tank with nice damage boost from Sheen.

Execution:
Game felt somewhat sloppy. Champions caught out a lot with many isolated kills during the laning phase. Part of this might be the mechanical skills of the players in maximizing damage. For example, watching Jarvan gank from the jungle in the first game it was surprising to see how much damage landed all at once. I think these players are superior in skill ordering and getting every last point of damage out very quickly. This plus the lack of cv and players are a lot more likely to die early game.

MiG got behind on kills but kept up on farm, capitalizing on every opportunity to grab map objectives. MiG establishes an early gold lead by taking drag and two towers. They used their range by grouping early to siege turrets. Skarner allowed them easy dive opportunities, and after first baron they picked apart towers by poking from range and chewing up tower hp. This allowed them to snowball the rest of the game, gaining a 10k gold advantage at 24 minutes which allows them to close out the game.

Sunday, February 12, 2012

v8 vs Epik - Game 1 - Curse Invitational

The Curse Invitational started this weekend, giving me a great chance to get a fresh look at some of the NA teams I haven't seen in a while.

Bans:
Unknown as the vod starts late. The tournament bracket doesn't list bans either but hopefully I can track this down.

v8 picks:
-Udyr
-Ahri
-Shyvana
-Sivir
-Leona

Plenty of damage from Sivir and Ahri, with great additional damage from Shyvana and Udyr. Damage mostly physical. Light on range meaning they need a hard engage. Tanky with Udyr, Shyv, and Leona. Hard cc from 3 champs (counting Ahri's taunt because it's insane) Leona has 3x cc with her ult. Decent amount of splash damage in close quarters but not a ton of large range aoe. All champs scale well into lategame.

Epik:
-Gankplank
-Orianna
-Amumu
-Corki
-Soraka

Great damage from Corki and Orianna, with solid additional damage from Pirate and Amumu (Amumu's sustained damage is substantial though not everyone realizes it). Decent range with Corki and Ori. Can poke with sustain from Soraka, Pirate steroid, Ori shield. Tankiness from Amumu and Pirate. Hard cc from Mummy and Ori ult, pretty light considering Corki and Soraka don't even have a slow. Good aoe with Amumu, Ori, and Pirate plus Corki's spam. Champs scale well into lategame, not 100% sure on Ori as I haven't seen her in so long.

Top Lane:
I haven't seen a lot of Udyr vs Pirate but Elementz pointed out that Udyr turtle counters a huge amount of Pirate harass so he expected Pirate to have a tough lane. Lane Udyr's tend to max Tiger first so Pirate shouldn't have trouble getting pushed.
Pirate took flash/teleport while Udyr took flash/ignite. This gives a lane advantage to Udyr but a Dragon control advantage to Epik.

Mid:
Orianna hasn't been played for a long time since she was nerfed, but still has plenty of damage. I believe she'd be a reasonable against Ahri because she has great range and can farm safely. Ahri is very strong right now, she has damage, cc, escapes, and sustain.
Both champs took flash/ignite, very standard.

Bot:
Leona isn't a very common support but her cc can be devastating, combined with Sivir's damage I can see it working. Corki/Soraka will tend towards farming, Corki has a nice escape and with Soraka's silence plus sustain this lane shouldn't lose.
Leona: exhaust/heal, Sivir: flash/ignite.
Corki: flash/cleanse, Soraka: flash/cv
There is a lot of experimentation with summoners in bot lane right now. Chaox indicated that in EU most ad's take flash/heal so he has shifted to that. The Leona/Sivir lane is extremely aggressive with exhaust and ignite. The ignite on Sivir is nice against Soraka. Corki has very safe summoners and cv on Soraka is helpful for the whole team, including keeping her lane safe.

Jungle:
Shyvana has become very popular with M5's recent success, she clears fast, moves fast, great counter jungler, and doesn't need blue buff. Amumu has become a rare pick, he clears fast, ganks well, duels well (in my opinion, though this may not be shared), but doesn't counter jungle well because he has no spike damage. Both contribute tankiness and damage to a teamfight, Amumu has much better initiation and can devastate squishies.
Shyvana: smite/exhaust. Getting an extra aggressive summoner from your jungler is excellent, it also helps for ganks, especially Shyv who has no utility/cc of her own.
Amumu: smite/flash. Mummy has a lot of aggressive utility with flash (tower diving, initiation).

Execution:
v8 snowballed very hard in this game, making it challenging to assess the viability of Epik's comp.
Top:
Udyr dominated Pirate top, as Elementz predicted, turtle stance > Pirate q. Even after losing his turret Pirate couldn't safely farm because Udyr was 5-0 in teamfights and outfarmed him.

Mid:
Ahri charm plus Shyv gank earned Ahri the first kill mid, though Ori held up well and farmed evenly.

Jungle:
v8 took Amumu's blue buff and even though they left their own blue open, Mummy didn't take it. After getting a slow start he died early trying to defend blue buff and struggled for the rest of the game. After establishing some control mid, Shyv/Ahri took Mummy's wraiths several times.

Bot:
v8 got a quick kill against Corki bot, didn't see the fight but Corki still had flash and cleanse up indicating a pretty bad fail.

Builds:
Mummy built a WoTA (Will of the Ancients) after Philosophers and Heart of Gold. This is a weird build and he didn't tank at all, causing him to melt in every fight. He also added a blasting wand after his first negatron, the game was almost over at this point but an Amumu without any health is no Amumu at all. I've seen Dan Dinh build Amumu this way in solo queue but it won't cut it in competitive, especially when you are the primary tank for the team.
Only other thing to note is that Pirate got beat top which meant he also wasn't tanking very hard, leaving Epik extremely squishy.

Takeaways:
Poke comps generally do well against hard commit comps, however in this case the early game advantage allowed v8 to snowball very hard for a quick victory. Udyr definitely managed Pirate in the top lane, and Amumu getting so far behind meant Epik had absolutely no backbone during teamfights. One important part of a successful comp is that everyone needs to perform their role, having a tank on paper won't cut it. It's important when analyzing Epik's comp to note how badly their early game went based on a few bad plays/decisions. Epik had a winnable comp but v8 showed up and executed very well.

The VoD
Props: Elementz gave an excellent team comp overview while shoutcasting which helped me articulate this post. Thanks!

Pro Meta Analysis

I decided that in addition to my Art of LoL posts, I'm adding a series on analyzing the meta on pro matches. Most of my knowledge has been acquired by studying the pros and because I've read several forum requests for this topic I decided to go for it.

The challenge in meta analysis is composition versus execution. I have won a scrubby ranked 5s match with five ranged ad champs (ez jungle, kog mid, trist + vayne bot, cait top). The fact that we won speaks to the fact that my team was far more experienced than the other team, not that this comp is viable. Many pro games will snowball when the team with the early lead presses their advantage and doesn't make big mistakes. For this reason you can't determine what comps are viable based purely on which team wins.

The format look at a team based on picks, bans, and team characteristic, then touch on the in game performance, including item builds. For the sake of being concise I won't provide a play by play, just a few game changing highlights that effect the outcome. 

Wednesday, February 8, 2012

I.10c - METHOD AND DISCIPLINE - road maintenance

On Summoners Rift, road maintenance is about map awareness, which is predominately about warding. I attempted some tutorial posts in the new player forum and was surprised when my 'profound' advice was ignored and my post quickly died. It woke me up to the fact that every player already knows that warding is important, it's just that few players do it.

In the excellent series Ult in 10 by ButButButILY, the host interviews Wickd, solo top player for Clg.eu, formerly of SK. The format involves watching a replay of a pro player while BBILY interviews the pro. Wickd played Irelia top and was purchasing tons of wards, placing them throughout the enemy jungle and at key choke points in river. When asked about it, he explained that it's worthwhile to buy wards as a solo because even though a support defers farming and kills, it still helps the team when a support can finish his build. It's also worth noting that if the support needs to stay in lane, he has very limited capacity to ward the top half of the map.

Watch a competitive game and it's immediately clear which team is winning; the team with wards all over the enemy jungle has established dominance and will dictate the pace of the game. Starving the enemy team by taking their buffs and controlling the big objectives will ensure victory. Moscow5, the new powerhouse out of EU, have consistently demonstrated their map control and these guys ward like crazy. For example, at the Kiev finals against SoloMid, they regularly placed as many as five wards on the dragon side of river in preparation for a dragon fight. They had nearly every brush and entrance covered. This kind of awareness allows for big plays, Alex on Galio was able to get his ult off on three enemy champs hiding in the floating bush in mid river. If most teams place two wards to cover themselves for dragon, then m5 is investing an additional 225 gold for the chance at getting 2-3 kills plus the dragon. When it succeeds they come out 1350 gold ahead (600 for kills, 950 for drag). When it fails they at least have a very clear picture of the river for the next few minutes which should allow them to lick their wounds and farm safely.

In conclusion, wards aren't just great in theory, they are absolutely essential for success at the highest level. One interesting trend I've noticed is that supports are buying more pink wards early in the game. This allows them to control the side bushes in lane, clear enemy wards to allow for ganks, and vie for dragon control.

Wickd on Ult In Ten
m5 vs TSM Kiev Grandfinals

Monday, January 16, 2012

I.10b - Method and Discipline - promotions

10. By METHOD AND DISCIPLINE are to be understood the graduations of rank among the officers.

Acknowledging your teammate's success is critical to supporting your team's improvement. Take time to acknowledge your teammates achievements, particularly during losses. One practice I've tried is to ask my teammates to list one thing they did well when we lose, and what they can improve on after a win. This prevents the team from getting down after a loss. Taking a break and keeping the environment lighthearted after a loss can give everyone a chance to relax and reset their focus. The worst thing you can do when your team is down after a loss is to requeue. In his excellent post on improving your elo without improving your skill, a platinum player named Pat the Power advises you to introduce a stop loss, meaning that you stop queueing ranked after two losses in a row. You won't lose 300 elo in a day if you follow this very valuable advice.

Promotions may also mean giving your teammates a chance to play a different role. Having a flexible team can give you a lot more freedom during champ select, so if a player wants to try something new, support that. Give him a clear goal (master two champions at that role well enough to play them in solo queue) then give them a chance to try out the role with a team. Feedback is critical so you need to encourage your teammate to ask for feedback. If he's not interested in hearing from the team how he can improve then you can do little to help him grow. One example of where this principle should have been applied was Elementz on CLG. He's a very strong player but the team wasn't happy with his support role. Instead of giving him feedback, they kept silent about it until they chose to bench him out of the blue. They could have given him a chance to play another role or grow in his existing role but instead they let the problem fester until it meant breaking up the team.

Supporting your teammates applies to solo queue as well. Make sure you always tell your teammates gj after successful ganks or kills. When they mess up, encourage them to keep trying, offer a lane switch, buy them a ward. Rage loses more games than any other factor in solo queue. Often the players who are the angriest accuse their teammates of being trolls while they drag everybody down. Mute these players, you'll be more relaxed and better able to focus on the game. Unfortunately you can't communicate with that player but it's better than spending half of your game typing responses to the guy.

Pat the Power on Improving your ELO