League of Legends - F2P MOBA (like DOTA) part 2

[Taejin]

Irelia squashes him top solo

Kayle does a TON of damage with her range skill now. I think it was buffed hard since Reckoning's damage amplification was nerfed. I played a game where the Kayle semi-fed as the carry bottom lane, but got ALL the kills in teamfights. Ended up 14-4, while I was 3-1-15 as Irelia =P.

 

[TheUnk]

Monkey players suck so far. He is good at escaping death, but isn't of much threat. His ult knocks you up once and does laughable dmg.

The champ spotlight has him levels ahead of the enemy. I guess it was the only way phreak was able to make the champ look useful.

 

[douglasb]

Anyone have impressions on Monkey King, new Tryndamere, new Kayle?

Monkey King's invis/double can be pretty annoying when trying to finish him off (I suppose that's the point) but the two I played against tonight didn't seem to be dishing out all that much damage.

Played against a Trynd and he was hitting HARD early on but seemed to taper off as we entered late game.

Played against several Kayle's, couldn't really make up my mind about her.

I haven't played against a Corki in forever so I can't comment on his changes.

I played a bot game with a Wukong, and he rocked pretty hard. I played Kayle that particular game and he was in lane with me, and he got quite a few double kills (although I had a lot to do with it; he just finished them off). He does seem a little OP based on what I saw, but let me remind you that this was a bot game.

New Kayle seems a little better to me, although I heard she was nerfed. Maybe I just learned to build her better, I don't know. My core build was:

Meki Pendant > Fiendish Codex
Boots of Speed > Boots of Swiftness
Stinger > Nashor's Tooth
Guinsoo's Rageblade (sometimes I completed this before Stinger/Nashor's)
Madred's Razor

I never got past Madred's but Banshee's Veil probably would have been next. At this point, Kayle was hitting pretty hard, healing for about 285, and no issues with cooldowns or mana. It seems like you are best off maxing out her R and E first, then her W, and only keeping Q at level 1 until you absolutely have to level it up. I will say that she requires a lot of farm to be successful, though.

I played a couple games with revised Tryndamere, and he seemed to hit harder early game and weaker late game than before. It looks like people won't be able to rely on turret diving, using his ult before they die, doing insane damage/getting a kill, and then spinning out. I think that was one of the things that was broken with him; his ult allows you to turret dive or go in 3v1, get a kill, and then escape without getting killed. Now, he doesn't seem to do quite as much damage later on, so it really isn't worth it to kamikaze like that.

As far as the new models/animations, other than Janna, I can't really tell any difference. Seems like a huge waste of time for them to update that when there are so many legitimate problems with the game.

 

[crownjules]

Monkey players suck so far. He is good at escaping death, but isn't of much threat. His ult knocks you up once and does laughable dmg.

The champ spotlight has him levels ahead of the enemy. I guess it was the only way phreak was able to make the champ look useful.

It's the same in every Champion Spotlight video. Obviously done on purpose to make the champion look better than they really are and get more people to buy them.

 

[AyashiKaibutsu]

The champ spotlight has him levels ahead of the enemy. I guess it was the only way phreak was able to make the champ look useful.

The champion spotlights are a joke. Yorick's was particularly funny in how set up it was.

 

[HumblePie]

Trynd right now is a monster in my opinion. He always scaled really well at the late game. If he could get his items he can decimate as long as the other team doesn't have many hard counters like multiple stuns, knockups, exhausts, or suppression. Those are still his bane, but then again they are the bane of almost all champs. The trick before was shutting down his early game so he was late in scaling up. This was easy to do before because he was so erratic on both healing and damage. This is no longer the case. He's now actually GOOD pre level 6 and a monster after level 6. He's way to OP in my opinion now.

Gap closer + escape
Slow + armor debuff
Healing + invulnerability
Massive reliable damage

There isn't much Trynd doesn't have going for him now. He only lacks a hard CC.



As for the other chars. I can't comment on Wukong yet. Haven't played him but every game last night I was in I got stuck with a Wukong player who was using the champ for the very first time. No one could be bothered to practice with Wukong in a bot or custom game. That would be too easy! Instead they invade my games and then feed and suck balls because they haven't a clue how to play the character. Then I lose. I end up with scores like 8/1/28 and I still lose. Won only 1 game in 8 last night. It was effing retarded.

Also the kayles are getting annoying. They are still trying to go all out dps with the char. They only use their ults and healing on themselves. They don't bother with any team support. Just stand back and swoop in at the end for some massive ksing. They just want to get the penta or quadra kills if they can at the end because of their splash damage. It's getting effing annoying with them for them to not do anything for half the team fight at least and let their team mates have to shoulder the burden with fighting 4v5 until they deign to join in at the end to claim the kills. It's one thing to hold back a second to make sure the tank gets the focus first, but it is something else what these kayles I've witnessed do. And it wasn't just one but 4 different ones that played the same way.


It's also damn near instant lock every game I've been playing to have a kayle, trynd, and wukong almost immediately. So I was stuck playing tanks or support most of the night. Not too big of a deal, but it's not exactly the best team comp that can be mustered with those 3 auto locked in. One of which last night was always a guaranteed feeder because the wukong player never played wukong and would suck balls.

 

[douglasb]

ANOTHER patch. So annoying.

 

[RPD]

So after a slight upgrade to my PC (new SSD's, HDD and fans) I finally got to play a game in about a week. Played Leona and I must say, not impressed.

Like others have said, she's weak as shit (damage wise, even for a "tank"), needs mana regen items (money taken away from tank items) and, this was the one that pissed me off the most, I think she may have the WORST move speed in the game. Holy shit she's slow. I know she has a gap closer, but if you ever need to run away, don't bother because you ain't getting away with her slow ass.

 

[Fenixgoon]

So after a slight upgrade to my PC (new SSD's, HDD and fans) I finally got to play a game in about a week. Played Leona and I must say, not impressed.

Like others have said, she's weak as shit (damage wise, even for a "tank"), needs mana regen items (money taken away from tank items) and, this was the one that pissed me off the most, I think she may have the WORST move speed in the game. Holy shit she's slow. I know she has a gap closer, but if you ever need to run away, don't bother because you ain't getting away with her slow ass.

i played a 75min game last night and she became tanky as shit (my team had much less late game, dominant mid/mid-late). still managed to pull off a win, but good god was it a PITA to get that nexus down. the final score was like......56 (us) to 58 (them). obviously this game is outside the norm

 

[HumblePie]

So after a slight upgrade to my PC (new SSD's, HDD and fans) I finally got to play a game in about a week. Played Leona and I must say, not impressed.

Like others have said, she's weak as shit (damage wise, even for a "tank"), needs mana regen items (money taken away from tank items) and, this was the one that pissed me off the most, I think she may have the WORST move speed in the game. Holy shit she's slow. I know she has a gap closer, but if you ever need to run away, don't bother because you ain't getting away with her slow ass.

I actually do alright with her. I use her as a counter pick to certain heroes in ranked. She is great for interrupting channeling ults with her double stuns. I do this for her build everytime and it works for me.

Regrowth Pendant + HP Pot
Complete Philo stone (takes care of almost all your mana problems and gets you some gold)
Boots
Heart Of Gold (add some HP and more gold which she needs cause her farming sucks)
Complete Merc Treads
Ruby + Cloth Armor + Null Cloak -> Aegis of the Legion
Phage (Hps and a chance for a slow and some AD)
Zeal (Attack speed, crit, and move speed)
Complete Tri Force
Complete Randuins

This nets me at level 18 around 3K hps, 160 AD, 190 base armor, 130 MR, and 420 movement. With the gap closer, a couple of well timed stuns, slows from tri force and randiuns, plus a decent movement speed will mean that no one gets away without flash. When I use Leona's W for her "AoE shield" it adds about 50 armor and MR if I remember. That puts you into a nice tanky place for team fights with that much HP and resistances.

Usually it is game over by this point but depending upon the circumstances of the game I'll pick up one of these items next:
Complete Shureliya's Reverie from Philo stone (which nets another 300 hps as well for more tankiness and speed!)
Thornmail versus AD team
Banshee's veil against MR team with lots of CC
FoN versus MR team with little CC


Her attacks are fairly weak. Her auto attack is nothing spectacular at all, but with Tri Force it does more than tickle at least. Her best asset is her ability to close the gap on the squishies in the back of a team fight and take them out of commission. The trick is to prioritize her stuns as much as possible. Her gap closer roots them for a second as well.

Leona's main source of damage in team fights come from her passive really. Her W popping is also decent as well. Her passive stacks damage on top of what your team mates do. So after Leona hits, then they need to hit. Leona's can not be passive in team fights as they only hamper their team and their damage output. Don't be afraid to press W for your light shield, tap E in the direction of an enemy you want to close the gap on, and then Q to stun then as soon as they move after the root wears off. On a squishie, if you have ANYONE on your team hitting them while you do those three abilities you WILL kill a squishie. Even on bruisers it only takes your Ult right after the Q stun wears off to finish of most mildly tanky bruiser characters. Just remember to lead the enemy with your ult so they get hit dead in the center of the ult for max effect.

Again the trick with Leona is she NEEDS someone working with her. He can cause pretty damn decent damage if you have someone who is on the ball by hitting their abilities on the same target as her.

Still the majority of the kills I get for her are early and mid game. Her damage is sill decent, but it really tapers off late game comparatively to how much more survivable enemy champs get late game. Still if you aren't getting a kill with her at level 3 or 4 in the laning phase then you are doing something wrong. Her ability to dominate the bottom lane, even against other early game bruisers like Jarvan, Garen, and Rumble make her great at slowing those characters down to the point they become non factors late game usually.

 

[bart1975]

Who is the best ranged character? I switch between Annie and Ashe currently.

 

[HumblePie]

Who is the best ranged character? I switch between Annie and Ashe currently.

Those are two very different characters. There is no "best" ranged either.

For the better ranged AD carries here is the list:
Ashe
Vayne
Caitlyn
Corki
Miss Fortune
Kog Maw
Twisted Fate

For the better ranged AP carries here is the list:
Annie
Malzahar
Morganna
Anivia
LeBlanc
Kennen
Brand
Karthus
Veigar
Tristana (she is devastating right now as AP build)
Twisted Fate
Vladamir


for the better support who CAN be almost a carry:
Lux
Orianna
Janna



As you can see there are more viable ranged AP characters than there are ranged AD. However, there are more ranged AP characters period. Still the other ranged AD characters such as Sivir, Ezreal, Teemo, and Twitch just don't cut it anymore. Tristana actually makes a better AP carry right now than an AD one. The dot damage + her jump + buster shot will 3 shot combo kill any squishy and most mildly tanky characters damn near instantly I've seen. But the same can be said for most of the AP characters. Which is why they are more popular meta game wise right now. You have people going AP Poppy, and AP Trist, AP Twisted Fate, and other characters that have traditionally been AD built. This is because AP is all about the burst damage and not sustained damage. AD has some burst, but only from crit based characters now like Trynd and Gangplank. Otherwise the AD carries are about sustains DPS during a fight. Also AP is actually easier to get to higher levels than AD. So it's not unheard of to see AP based characters with AP scores from 600-1000. Then again, AP characters don't crit either. So AD characters rely on crits to give them those big burst hits. But crits are fickle and some characters are better than others with them.

 

[Taejin]

Are we playing the same game? Orianna is on track to be the best AP carry in the game...

 

[zebano]

First off, Elo was designed for chess players. It was designed for 1 v 1 interaction and not team play. It was designed for the long haul over several seasons. It was designed in mind that all participants are trying their best to their abilities to win. It was designed based on a game that only has one real outcome. It was designed based around players being there from the start of a match until the end.


League of Legends does not fit into the Elo model because of that. It is a team based game. Some players are out to troll others or goof off. Some people have internet or computer problems for various reasons. Players and accounts come and go constantly.

So how would I fix it?

First is the team aspect. Sometimes you do really well as a player and still lose. I've had games where I've gone 17/5/20 and still lost. Why? Usually because someone on my team was an idiot all game. Sometimes it was just because it was a close game, but that is the rare exception. However I get penalized the same as everyone on my team regardless of my performance versus theirs. This should not be the case. If your KD sucks, your minion kills suck, and all the other meta statistics suck then you deserve to lose more elo for that loss than others on the team that don't. Same thing for a win. Those that play well should get a bigger elo boost than those that were carried along for the ride to victory. This would go a long way to helping the current elo system.

Second of all, do NOT start new players out at 1200 elo, which is the middle. By definition of elo, a new player starts near the top. A loss counts BIG for them at first. New players in rank should start at 2000+ elo. A loss at first can be from 100-400 elo points depending on how they lose. That is how chess Elo works. You start out player the top players and as you get your ass handed to you then you sink down the elo rating until you reach the level of your skill. You don't start in the middle or bottom and have to work your way up. That's just dumb. More so in a team base elo system where your win is very rarely dependent upon the actions of a single player but losses can very well be dependent upon a single player.

Null out games where teams have players disconnect OR allow players to come in during the middle of a game to take over where a disconnected player left off. One of the two needs to happen. The player that actually disconnects needs to have a "loss" counted against them, but not the rest of the team.

Allow X amount of queue dodges to occur per day. I think 2 or 3 is fine. Those don't count as losses or leaves or hurt elo. Anything more would. I also should state that "leaves" occurred after the limit for queue dodges are reached need to have sizable elo loss afterwards.

Last, remove duo-ing from solo queuing. You either enter a game solo and everyone is solo, or it is all 5 man pre-mades. There is no in between for ranked nor should there be. That is what normal games are for. Keep the playing field as level as possible. One team in the "solo" queue should not have that decisive of an advantage because they have 2 duo's on a single versus the other team with 5 singles. I've seen that happen. It's stupid.


Just those fixes would go a LONG way to solving the problems with the current elo system and leaving some good players to get stuck in "elo hell" and prevent crappy players from being

Ok you said tons of stuff that I don't agree with but I'm going to skip most of it because we've had the argument before about good players (1800+ creating smurfs and easily getting their smurfs up to 1800+ .. if that doesn't convince you that elo works, I don't know what will)

FACT: Chess players do not start at 2k / top elo. Provisional/unrated players are assumed to have a 1200 rating for rating purposes. Often times what you will see in chess tournaments are that they are not allowed to play in under 1400 sections because they may just blow away the competition, but tournaments big enough to have multiple sections usually have a specific unrated section for the new players (who may also play the open section if they wish). Their first 25ish (it's changed recently) games, elo is calculated as follows

game elo result = opponents elo if game drawn, opponents elo +400 if won, opponents elo - 400 if lost.

They then average the game elo for all games until they have played 25 games, after that they gain and lose elo similar to how we do in LoL (difference in rating, Kfactor and result). The main differences are that the K in chess is constant until you hit a very high rating (2200 iirc) at which point it is lowered. K is lowered in LoL based on how many games you have played and raised or lowered based on your deviation in rating from the rest of your team. Regardless, you are still matching average team elos and adjusting based on the result of the game this just means that after your first 50 games you have to play better over more games to make a significant difference in rating.

The only good argument to say that elo does not work for LoL over a large sample of games that I have heard would be to state that players effective ratings are so incredibly variable that they cannot be estimated reliably at which point they can't use any rating system.

Shit got real in this thread haha.

Regarding the skill cap in LoL. I only started playing LoL a few months ago. But I played a lot of Dota at a decent non-bnet level for quite a while. I bring that up because most people will tell you Dota has a higher skill cap then LoL. This could possibly be true. But I've honestly never even come close to hitting skill caps in either game. This may be because I'm drawn towards heros with high skill caps (Puck, Potm in Dota, Lee Sin in LoL). But it's more likely because as good as I am, I'm not that good. I'm not high elo because I don't belong there. I do feel my elo should be higher then it is, but that's fine because I haven't played a lot of ranked and elo is games above .500 pretty much so the more you play the higher you go, or lower haha. Honestly if you feel you've hit the game's skillcap and you aren't high elo you're probably doing it wrong. You can make all the elo hell arguments you want, but people carry to high elo all day long. I think it's pretty safe to say most of the best players play ranked, and most of them are 1800+. If you aren't challenged by normals that's fine. But to say you don't find the game challenging is a little ridiculous. Unless you're like jiji or someone who can carry an account to 2.5k elo in 140 games, then maybe you can say you've pretty much reached the ranked skillcap. He still needs to work on his tourney level play though, so he can still improve.

I mean we played with that one guy in a few inhouses several weeks ago, Zirze? Anyway he straight carried them all no problem. He deserves his 2k elo. People who are high elo typically deserve it, people who are low elo also typically deserve it also. Elo hell is a bit of a crutch, like raging at people for not calling miss when in reality it was you who should have been watching the map while being overextended.

That said, ranked games are much easier if you duo them haha. I know that most of us here are in the 1000-1400 range, I would be willing to wager that if some people here duo'd up they could get higher. Most of the AT people are pretty solid. Duo'ing is a lot more fun too then just straight solo play.

Quoted for humble to re-read. There is a reason Zirze often (not always) carries and Humble does not. Note that the more people we queue with the more likely we are to lose even with zirze. I believe this is because we end up matched with more premades who can deal with losing one lane by winning the other two. We're often on vent but we don't do much to control buffs, ward or adjust to the enemy.

Another simple adjustment: If they're blowing Xin up is it still worth it to initiate if you're only going to get e and r off? Sometimes the answer is yes, sometimes you should pretend that you're master yi and wait for the enemy to blow their spells on someone else and then clean up. Sometimes your teammates have already lost the game (which you seem to claim is every single loss of yours...).

The intarwebs are full of stories of people playing slightly better and raising their rating even after hundreds of ranked games which combined with the fact that I always graviate back to the same rating after good/bad streaks means that elo works.

Oddly enough duo-queues seem to be worse for me than solo queuing. I believe this is because they artificially inflate your elo when you duo queue to compensate. Pipi and I have queued a bunch lately with very mixed success.


/my elo thoughts

So after a slight upgrade to my PC (new SSD's, HDD and fans) I finally got to play a game in about a week. Played Leona and I must say, not impressed.

Like others have said, she's weak as shit (damage wise, even for a "tank"), needs mana regen items (money taken away from tank items) and, this was the one that pissed me off the most, I think she may have the WORST move speed in the game. Holy shit she's slow. I know she has a gap closer, but if you ever need to run away, don't bother because you ain't getting away with her slow ass.

I really like Leona. One of my gripes lately has been Alistar's combination of control and damage. He can build sheen or AP and dish out significant damage while still tanking like a boss. Someone with that level of CC should not be doing that much damage and be that tanky (ult is too good). Leona is a nice balance of innate tankyness, utility & minimal damage to offset that. I personally like to build her CDR + Aura.

Typically I may go:
start with regrowth + pot (35g)
Boots of Lucidity (1050g)
P. Stone (800g) or HoG (825g) for G/10
Glacial Shroud (1525g)
Negatron Cloak (740g)

For a total cost of no more than 4175g allowing you to buy wards and items like Aegis of the Legion, Frozen Heart, Will of the Ancients and Starks Ferver for you team as needed.

From there I like to

 

[nanobreath]

I actually do alright with her. I use her as a counter pick to certain heroes in ranked. She is great for interrupting channeling ults with her double stuns. I do this for her build everytime and it works for me.

*snip*

As somebody who plays Leona, maybe you can answer this question for me. I was playing on a team with leona, me on Udyr. Leona was solo top and getting her ass handed to her by xin and katarina (yeah probably the two WORST characters in the game to feed). So I of course go up to her to try and gank. She kinda pulled them to the tower, and used the gap closer on them when they got close, and I charged in from the bushes. Xin is level 3 or 4, and has 800 HP at this point. As soon as she initiated I went in and attacked xin as that was who she was on and I swear he died in 2 hits. His HP literally MELTED. When they first started attacking, leona lost a good bit of HP early, and I was afraid they were going to kill her first, but the second we hit Xin he was dead from full. WTH gave us that much burst that quickly. I'm used to udry taking down people quick, but this was absurd.

 

[Firsttime]

I'm not a Leona fan at all. She's incredibly dull to play. Her skill shots require little skill and her damage is pretty lame. She's pretty much a stun bot. She mostly serves the anti fun purpose of being a "dive punisher" in lane, in that she's there to discourage trading hits with her partner. As far as tanks go Alistar and Amumu both outclass her and are more fun to play. In addition to only being 1350 ip haha.

 

[HumblePie]

Ok you said tons of stuff that I don't agree with but I'm going to skip most of it because we've had the argument before about good players (1800+ creating smurfs and easily getting their smurfs up to 1800+ .. if that doesn't convince you that elo works, I don't know what will)

Because even when they create smurf accounts, they are NOT starting at the bottom or alone. They duo with other good players, which means they know 2 out of the 5 are good, and usually duo with a higher ranked player. Which means their average elo for the "duo" is higher than the new smurf account. So a new smurf account starting at 1200 duoing with someone that has 2400 elo are going to have games with people in the 1800 elo range. Meaning they are going to be playing with people who are usually decent and are TRYING to win a ranked game. Not people playing a ranked game for the hell of it before dinner only to go AFK during the middle of the game. Or people just out to troll. Or even people just typically starting ranked for the first time.

So no, that example doesn't float because I've watched the streams of those players from M1lkfat and on. They ALL do that. I have yet to see a single "high" level elo player start from scratch and completely SOLO reach higher elo ranks in the solo queue. Give me one example and I'll bow down to your assertion. But as of yet I haven't found a single case of this happening.

FACT: Chess players do not start at 2k / top elo. Provisional/unrated players are assumed to have a 1200 rating for rating purposes. Often times what you will see in chess tournaments are that they are not allowed to play in under 1400 sections because they may just blow away the competition, but tournaments big enough to have multiple sections usually have a specific unrated section for the new players (who may also play the open section if they wish). Their first 25ish (it's changed recently) games, elo is calculated as follows

game elo result = opponents elo if game drawn, opponents elo +400 if won, opponents elo - 400 if lost.

They then average the game elo for all games until they have played 25 games, after that they gain and lose elo similar to how we do in LoL (difference in rating, Kfactor and result). The main differences are that the K in chess is constant until you hit a very high rating (2200 iirc) at which point it is lowered. K is lowered in LoL based on how many games you have played and raised or lowered based on your deviation in rating from the rest of your team. Regardless, you are still matching average team elos and adjusting based on the result of the game this just means that after your first 50 games you have to play better over more games to make a significant difference in rating.

FACT: I KNOW they don't start out that high. But the fact is chess is SINGLE PLAYER and not multiplayer. I was trying to take into account the problems that a multi player ELO setup has over a single player.

I've read and understand this. http://en.wikipedia.org/wiki/Elo_rating_system

But one of the things with even a chess elo system is players at the higher ELO levels guard their elo as best they can. They choose when and where to play which skews the elo system there. It's a common concern. For LOL, the biggest problem is getting grouped with people who's skill level you have no control over. With the exception of 5 man pre-mades. Which I've tried to stay out of for now because I have yet to find 5 people I feel comfortable with to be on for a regularly scheduled ranked session of games. Not that I don't know good players, but getting them all together and working together with practice before hand is something different. Because team cohesiveness makes a huge impact in team based games regardless of skill. I know this from many, many, MANY years of team based sports. The underdogs with no stars on their team can beat the big dogs with all the stars for any sport so long as the underdogs play better as a team. Just a fact. There is a reason why team based sports do NOT used the ELO system. They use a regular win loss system. Then again, team based sports use the assumption that teams remain together for the most part through a season.

To use a quote from the Wiki link though.

Particular examples of 2800+ rated players choosing opponents with minimal risk and maximum possibility of rating gain include: choosing computers that they know they can beat with a certain strategy; choosing opponents that they think are over-rated; or avoiding playing strong players who are rated several hundred points below them, but may hold chess titles such as IM or GM. In the category of choosing over-rated opponents, new-entrants to the rating system who have played less than 50 games are in theory a convenient target as they may be overrated in their provisional rating. The ICC compensates for this issue by assigning a lower K-factor to the established player if they do win against a new rating entrant. The K-factor is actually a function of the number of rated games played by the new entrant.

Basically NEW chess players don't go to major tournies until have a high enough provisional rating. Even players with very HIGH elo recognize that there are players with lower scores who are probably better than they are, but get stuck with small gains due to the fact that no one above them in elo is willing to play them. The provisional ratings for new players start pretty high, and much higher than 1200. Here is another quote from the article talking about it.

Higher rating floors exist for experienced players who have achieved significant ratings. Such higher rating floors exist, starting at ratings of 1200 in 100 point increments up to 2100 (1200, 1300, 1400, ... , 2100). A player's rating floor is calculated by taking their peak rating, subtracting 200 points, and then rounding down to the nearest rating floor. For example, a player who has reached a peak rating of 1464 would have a rating floor of 1464 - 200 = 1264, which would be rounded down to 1200. Under this scheme, only Class C players and above are capable of having a higher rating floor than their absolute player rating. All other players would have a floor of at most 150.

So players can start higher, depending upon the circumstances, and would have bigger swings in ELO based on performance for their first set of games. LoL took the standard minimal approach to starting new players with ELO and I personally do NOT think it works well. That is a personal opinion and one others I know share. Chess leagues and tournaments all have different entry points and none of the have been rated "best" to use yet.



Quoted for humble to re-read. There is a reason Zirze often (not always) carries and Humble does not. Note that the more people we queue with the more likely we are to lose even with zirze. I believe this is because we end up matched with more premades who can deal with losing one lane by winning the other two. We're often on vent but we don't do much to control buffs, ward or adjust to the enemy.

Oddly enough duo-queues seem to be worse for me than solo queuing. I believe this is because they artificially inflate your elo when you duo queue to compensate. Pipi and I have queued a bunch lately with very mixed success.


/my elo thoughts

Actually, the couple of good players I found for DUO we blow the games up if I can manage to get with them. It's the completely solo I have mixed results in. It's half wins and half loses. The loses are usually because by the time I reach level 6 the other two lanes and or jungle on my team has failed BADLY assuming that I have a full team even playing (no afk or leaver). Usually when this happens I'm level 6 and the rest of the other team is anywhere between levels 6 and 10 (my counterpart typically matches my level unless I'm just dominating them). It's not because I was forced out of my lane or zoned out or anything. It's because my team mates fed like crazy. There is no carrying back from that.

 

[HumblePie]

As somebody who plays Leona, maybe you can answer this question for me. I was playing on a team with leona, me on Udyr. Leona was solo top and getting her ass handed to her by xin and katarina (yeah probably the two WORST characters in the game to feed). So I of course go up to her to try and gank. She kinda pulled them to the tower, and used the gap closer on them when they got close, and I charged in from the bushes. Xin is level 3 or 4, and has 800 HP at this point. As soon as she initiated I went in and attacked xin as that was who she was on and I swear he died in 2 hits. His HP literally MELTED. When they first started attacking, leona lost a good bit of HP early, and I was afraid they were going to kill her first, but the second we hit Xin he was dead from full. WTH gave us that much burst that quickly. I'm used to udry taking down people quick, but this was absurd.

Her passive.

Leona should NOT have a solo lane. Her passive leaves a "blow up" debuff on the target. Every time she hits a person with any of her abilities they get a debuff on them that lasts for a few seconds. The debuff she can not activate and does nothing with out a partner. The moment YOU as udyr hit her target, it pops her debuff. The debuff just does massive damage. At low levels, a couple of pops is enough to kill anyone. I mean ANYONE. That is where her damage comes from really. Most people don't pick up on this. Which is why I said that Leona REQUIRES a team mate that is on the ball with her.

Basically as Leona you need to initate and dive hard. Whomever you pick, you need to have one of your team mates ready to beat on that person with her immediately to make that enemy melt.

 

[xSkyDrAx]

Just got my first penta as Ashe yesterday. Finally though I thought I would have gotten it with vlad a long time ago but I usually ended up with a bunch of quadras.

 

[zebano]

Her passive.

Leona should NOT have a solo lane. Her passive leaves a "blow up" debuff on the target. Every time she hits a person with any of her abilities they get a debuff on them that lasts for a few seconds. The debuff she can not activate and does nothing with out a partner. The moment YOU as udyr hit her target, it pops her debuff. The debuff just does massive damage. At low levels, a couple of pops is enough to kill anyone. I mean ANYONE. That is where her damage comes from really. Most people don't pick up on this. Which is why I said that Leona REQUIRES a team mate that is on the ball with her.

Basically as Leona you need to initate and dive hard. Whomever you pick, you need to have one of your team mates ready to beat on that person with her immediately to make that enemy melt.

This is absolutely true, there is no reason for Leona to solo lane at all as her passive is designed to encourage teamwork. Combining it with Lux or Jarvan's passives have all yielded massive damage and almost certainly a kill.

My only gripe with Leona is how slow her skillshot is, you are almost required to be in the bushes or at point blank range. The only problem I ever have is in teamfights I will often initiate with shield (to apply passive to most of their team) then ult their carry/mage if I think my teammates can get to them and I will then use the stun + dash to peel 2 people off of my carry and I get cussed at for focusing the wrong person when I'm simply trying to protect my damage dealers.

Just got my first penta as Ashe yesterday. Finally though I thought I would have gotten it with vlad a long time ago but I usually ended up with a bunch of quadras.

Congrats! I have tons of quadras and no pentakills though I should have had one a couple weeks ago as Fiddle but our trist got the final hit on the Soraka.

 

[zebano]

Because even when they create smurf accounts, they are NOT starting at the bottom or alone. They duo with other good players, which means they know 2 out of the 5 are good, and usually duo with a higher ranked player. Which means their average elo for the "duo" is higher than the new smurf account. So a new smurf account starting at 1200 duoing with someone that has 2400 elo are going to have games with people in the 1800 elo range. Meaning they are going to be playing with people who are usually decent and are TRYING to win a ranked game. Not people playing a ranked game for the hell of it before dinner only to go AFK during the middle of the game. Or people just out to troll. Or even people just typically starting ranked for the first time.

Part of the problem i have with people arguing against elo is that they say trolls ruin it... Trolls unfortunately ruin any rating system not just elo ergo you are basically advocating to be randomly matched with people instead of randomly matched with people who have similar results to you.

So no, that example doesn't float because I've watched the streams of those players from M1lkfat and on. They ALL do that. I have yet to see a single "high" level elo player start from scratch and completely SOLO reach higher elo ranks in the solo queue. Give me one example and I'll bow down to your assertion. But as of yet I haven't found a single case of this happening.

I'm not going to hunt them down for you but there are lots of people just searching the reddit subreddit (very active, moderately better than the official forums) who went from 900-1600 or 1800. Some of them do it by carrying some by playing support. I have also seen from there high elo players who win taking over an existing ranked players account (i.e. they don't get the early high-k factor).

FACT: I KNOW they don't start out that high. But the fact is chess is SINGLE PLAYER and not multiplayer. I was trying to take into account the problems that a multi player ELO setup has over a single player.

I've read and understand this. http://en.wikipedia.org/wiki/Elo_rating_system

But one of the things with even a chess elo system is players at the higher ELO levels guard their elo as best they can. They choose when and where to play which skews the elo system there. It's a common concern. For LOL, the biggest problem is getting grouped with people who's skill level you have no control over. With the exception of 5 man pre-mades. Which I've tried to stay out of for now because I have yet to find 5 people I feel comfortable with to be on for a regularly scheduled ranked session of games. Not that I don't know good players, but getting them all together and working together with practice before hand is something different. Because team cohesiveness makes a huge impact in team based games regardless of skill. I know this from many, many, MANY years of team based sports. The underdogs with no stars on their team can beat the big dogs with all the stars for any sport so long as the underdogs play better as a team. Just a fact. There is a reason why team based sports do NOT used the ELO system. They use a regular win loss system. Then again, team based sports use the assumption that teams remain together for the most part through a season.

To use a quote from the Wiki link though.


Basically NEW chess players don't go to major tournies until have a high enough provisional rating. Even players with very HIGH elo recognize that there are players with lower scores who are probably better than they are, but get stuck with small gains due to the fact that no one above them in elo is willing to play them. The provisional ratings for new players start pretty high, and much higher than 1200. Here is another quote from the article talking about it.

Until 2 years ago I was a USCF tournament director, they simply don't start higher than 1200 though they can fluctuate much faster than LoL ratings do. The explanation I gave above is how the United States Chess Federation ranking system works. The problem of getting good people to play you only exists for people of Grand Master strength (2700+ is super-grand master and extremely extremely rare).

So players can start higher, depending upon the circumstances, and would have bigger swings in ELO based on performance for their first set of games. LoL took the standard minimal approach to starting new players with ELO and I personally do NOT think it works well. That is a personal opinion and one others I know share. Chess leagues and tournaments all have different entry points and none of the have been rated "best" to use yet.

I could understand this, but if trolling is a problem, consider what would happen if you dumped new people in as 2000 rated? They wouldn't be playing with the jiji and hotshotggs. All that would happen is the ratings would shift and those people would be rated 3000 and the current 1200s would be 2000s. It's an arbitrary point. One thing that might be effective would be to have an increased K factor early (more so than it is now) but that will still have the same problem that you're dependent on your teammates. You're right that it's a problem but statistics shows us that it works over a very large sample size. K/D is an interesting try but it fails for some obvious reasons:


If my teammates feed early and I'm Ashe, I'll finish the game 1/2/x just because I play it safe, try to farm and am mostly impotent. If I'm trundle or another melee character I have no influence if I don't get into at least the edges of the fray meaning I will most certainly end the game 2/7/x. Is this really worse than the Ashe who sat back and volleyed but realized she couldn't do anything? Also what about supports who go 0/2/20 or a Master Yi who realizes what is happening split pushes down 5 towers gets caught once or twice and ends the game 3/5/x while taking out an enemy inhibitor?

My main problem is that you complain complain complain and I while I admit that you're better than the average player, you're far from the best players I've seen (you expect to be) and you haven't offered a single reasonable option that is better than what exists.

Quoted for humble to re-read. There is a reason Zirze often (not always) carries and Humble does not. Note that the more people we queue with the more likely we are to lose even with zirze. I believe this is because we end up matched with more premades who can deal with losing one lane by winning the other two. We're often on vent but we don't do much to control buffs, ward or adjust to the enemy.



Actually, the couple of good players I found for DUO we blow the games up if I can manage to get with them. It's the completely solo I have mixed results in. It's half wins and half loses. The loses are usually because by the time I reach level 6 the other two lanes and or jungle on my team has failed BADLY assuming that I have a full team even playing (no afk or leaver). Usually when this happens I'm level 6 and the rest of the other team is anywhere between levels 6 and 10 (my counterpart typically matches my level unless I'm just dominating them). It's not because I was forced out of my lane or zoned out or anything. It's because my team mates fed like crazy. There is no carrying back from that.

When duo-queues go right, they go really right, I've just found I'm more likely to have abad feeder in duo queue than I am in solo queue which is really odd. This is probably small sample size on my part or perhaps it's what we chose to play - pipi was also in the duo queue while I jungled meaning we left both solo lanes to chance. The one advantage we enjoyed from this was that when I went for dragon someone from my bottom lane always joined me.

 

[HumblePie]

Part of the problem i have with people arguing against elo is that they say trolls ruin it... Trolls unfortunately ruin any rating system not just elo ergo you are basically advocating to be randomly matched with people instead of randomly matched with people who have similar results to you.

Trolls ruin team based video gaming rating systems. You don't get trolls in real life or when money is on the line such as a chess or poker tournament. I said that trolls are part of the problem, not the whole problem, with the current implementation of the ELO system. I specifically stated that the current implementation of ELO for LoL is very basic and does not try to account for these very persistent problems. It is one thing if trolls were limited to a small number of games and it was something of less than 1% of games. That's a statistical outlier that can be ignored. The problem is, that 1% may be true for higher elo games 1800+, but is certainly NOT true for lower elo games. Me and most others all feel that the current system needs to be adjusted for them. Hence my suggestions above. But trolls are the only problem with the current system.

I'm not going to hunt them down for you but there are lots of people just searching the reddit subreddit (very active, moderately better than the official forums) who went from 900-1600 or 1800. Some of them do it by carrying some by playing support. I have also seen from there high elo players who win taking over an existing ranked players account (i.e. they don't get the early high-k factor).

And every single one of them did it as a DUO with another higher ranked player. I've seen videos on Gaurdsman Bob, and Saint Viscious, and even m1lkfat, who are "paid" to play someone else's account to get them higher in elo. What was their method besides obviously playing good? They DUO'd with a higher elo team mate. Every single time. It improves your chances of winning dramatically. By REMOVING the ability to DUO in a supposedly solo queue takes that out of the equation for people trying to rise through the ranks.

Until 2 years ago I was a USCF tournament director, they simply don't start higher than 1200 though they can fluctuate much faster than LoL ratings do. The explanation I gave above is how the United States Chess Federation ranking system works. The problem of getting good people to play you only exists for people of Grand Master strength (2700+ is super-grand master and extremely extremely rare).

congrats, and while I wasn't a USCF tournament director, the wiki I posted stated otherwise. The norm is 1200, but even then, that is starting in the middle at Class C. No one starts at the bottom in an ELO system. There are reasons for starting people at least in the middle. I was trying to give reasons to start people even higher with wider fluctuations, because of this being a team based game. I was not dictating what should be done in chess, but what I feel will fix the current ELO system. I am not alone in this opinon either as I didn't come up with it out of the blue. I actually read some statistics on the official forums of people talking about this which is what swayed my opinion to this.

I could understand this, but if trolling is a problem, consider what would happen if you dumped new people in as 2000 rated? They wouldn't be playing with the jiji and hotshotggs. All that would happen is the ratings would shift and those people would be rated 3000 and the current 1200s would be 2000s. It's an arbitrary point. One thing that might be effective would be to have an increased K factor early (more so than it is now) but that will still have the same problem that you're dependent on your teammates. You're right that it's a problem but statistics shows us that it works over a very large sample size.

As I said just right above, I was not the person to think up of this suggestion and there were others with their own math formulas as to why a higher starting ELO should be used with a more varied depression/inflation.

K/D is an interesting try but it fails for some obvious reasons:


If my teammates feed early and I'm Ashe, I'll finish the game 1/2/x just because I play it safe, try to farm and am mostly impotent. If I'm trundle or another melee character I have no influence if I don't get into at least the edges of the fray meaning I will most certainly end the game 2/7/x. Is this really worse than the Ashe who sat back and volleyed but realized she couldn't do anything? Also what about supports who go 0/2/20 or a Master Yi who realizes what is happening split pushes down 5 towers gets caught once or twice and ends the game 3/5/x while taking out an enemy inhibitor?

Never said it was perfect. But then again, if your team is feeding and you end up 1/2/x with an Ashe that is going to be better than your team mates scores even if you lose. So that should be reflected as your performance for ELO changes. Also, I never stated the elo changes should only be a reflection of your K/D score, but a reflection of what your score was compared to team mates as well as the competition. For example...

You score as an ashe....

4/2/0 at the end of the game. Looks pretty good right? 2:1 KD ratio.

But I scored as the tank.... 2/8/14. Same 2:1 ratio. What's the difference? You could tell I more of a factor for the team in a team based game because I was there for the FIGHTS. so even if we lose I would have the lesser ELO penalty than an Ashe that huge back and didn't contribute. I never said to make it a static algorithm based on K/D only. Still, nothing is perfect, but things CAN be better than what they are.

Also, K/D wasn't the only measurement I was saying to take into account either. There are other things they keep track of such as tower and inhibs taken down, minion kills, gold earned, damage dealt and received. A certain amount of weighting for ALL those categories n RELATION to the rest of your team needs to be considered in the current ELO formula.

My main problem is that you complain complain complain and I while I admit that you're better than the average player, you're far from the best players I've seen (you expect to be) and you haven't offered a single reasonable option that is better than what exists.

I've pointed out problems and fallacies with the system. I made suggestions of what I feel would improve them. That's not "complain, complain, complain" as you put it. The current ELO system is off. In chess the ELO players have a +/- 100 in "performance" from their current ELO. You don't see people dropping drastically once they reached 30+ games. They may increase as their skills increase, but they don't really drop. The same can not be said for LoL. I've dropped and raised 500 point swings through no real changes to my play. Many others have. Actually it's more common than less so. That speaks of an inherent major flaw in the current system.

 

[zebano]

congrats, and while I wasn't a USCF tournament director, the wiki I posted stated otherwise. The norm is 1200, but even then, that is starting in the middle at Class C. No one starts at the bottom in an ELO system. There are reasons for starting people at least in the middle. I was trying to give reasons to start people even higher with wider fluctuations, because of this being a team based game. I was not dictating what should be done in chess, but what I feel will fix the current ELO system. I am not alone in this opinon either as I didn't come up with it out of the blue. I actually read some statistics on the official forums of people talking about this which is what swayed my opinion to this.

This is not the case. Elo is not hard and fast 2k is good, 1200 middle (Class D in chess btw, not C, c is 1400-1599). It is a relative system meaning that if you reset the ladder and made 2000 the starting point, elos would range from 1 to 3xxxx instead of 1-2500. If you made the change without a reset you would seriously piss off some good players by making them play with fresh level 30s who still don't really know how to play. Assuming a reset, perhaps the distribution would spread more, but then either matchmaking gets slower or you get matched with a wider range of elos in a given game. A similar approach would be to keep the current implementation and tell the mm system that instead of a range of 200 elo in a given game, the max is 100 (I'm not sure what it really is but I've seen 300 when I cared to look everyone up, but that was a duo queue).

Honestly I think the best solution is twofold:

1. remove the ability to duo-queue for solo-queue ranked.
2. balance champions for TT allowing small groups of friends to easily queue together. You would have to make the skills scale differently on TT than on SR so you could keep the balancing separate.

Then again, I don't see it as pressing of a problem as you do. When top dies twice, they are usually willing to lane switch and if not, I as a jungler will give up jungle and lane with them (which sucks, but alleviates many problems).

But I scored as the tank.... 2/8/14. Same 2:1 ratio. What's the difference? You could tell I more of a factor for the team in a team based game because I was there for the FIGHTS. so even if we lose I would have the lesser ELO penalty than an Ashe that huge back and didn't contribute. I never said to make it a static algorithm based on K/D only. Still, nothing is perfect, but things CAN be better than what they are.

Also, K/D wasn't the only measurement I was saying to take into account either. There are other things they keep track of such as tower and inhibs taken down, minion kills, gold earned, damage dealt and received. A certain amount of weighting for ALL those categories n RELATION to the rest of your team needs to be considered in the current ELO formula.

Interesting idea, but once again you penalize supports that willfully give up cs. Their CS looks bad, they won't get kills, they may even die for their team and they certainly won't get towers. Best case they go 0/0/20. How do you weight tower then? If I'm alistar with a 4v5 mid and I have a Yi/TF/Sivir/etc split pushing and we trade one or two deaths for a tower, I'm ok with that (if we keep our tower). They lose a lot of map control by losing that tower. If instead we've got one dead and we're 3v5 mid and TF continues to push instead of teleporting mid, that's bad but in both cases they can choose to take that tower and I don't know how you differentiate between the two cases using statistics. Another poor player that may get rewarded is the mid that constantly ganks but loses their tower to a Caitlyn that pushes non-stop. The mid can end up with a nice K but there is no way to assign that tower loss to them.

I appreciate that you're throwing real ideas out here, I just don't find any of them to be quite feasible due to the complexity of the situation. If you really think duo-queing is the answer, hit me up when I get back from vacation, I've never trolled a ranked game.

 

[hans030390]

Those are two very different characters. There is no "best" ranged either.

For the better ranged AD carries here is the list:
Ashe
Vayne
Caitlyn
Corki
Miss Fortune
Kog Maw
Twisted Fate

For the better ranged AP carries here is the list:
Annie
Malzahar
Morganna
Anivia
LeBlanc
Kennen
Brand
Karthus
Veigar
Tristana (she is devastating right now as AP build)
Twisted Fate
Vladamir


for the better support who CAN be almost a carry:
Lux
Orianna
Janna

I wouldn't call LeBlanc a carry. Sure, she can deal a ton of burst damage and is a great assassin/1v1 champ, but not a carry. Her ult is the main reason she has such great burst damage, and that has a ~30 second cooldown. When you're waiting for that, though...you aren't too useful.

She also is probably the worst pusher I've ever played. It's very important to keep her teleport move ready (~15 sec cooldown) in case you need to escape or murder someone, so she can't farm very well. Her auto-attack barely does anything, so she can't do much to turrets.

I think I would only label her as a carry if the entire enemy team was comprised of squishy characters with no magic resist. But once you put her awful pushing skills and cooldown times together, she's no carry. Hell, if the enemy team has even a bit of MR and health, she can be pretty useless.

 

[HumblePie]

This is not the case. Elo is not hard and fast 2k is good, 1200 middle (Class D in chess btw, not C, c is 1400-1599). It is a relative system meaning that if you reset the ladder and made 2000 the starting point, elos would range from 1 to 3xxxx instead of 1-2500. If you made the change without a reset you would seriously piss off some good players by making them play with fresh level 30s who still don't really know how to play. Assuming a reset, perhaps the distribution would spread more, but then either matchmaking gets slower or you get matched with a wider range of elos in a given game. A similar approach would be to keep the current implementation and tell the mm system that instead of a range of 200 elo in a given game, the max is 100 (I'm not sure what it really is but I've seen 300 when I cared to look everyone up, but that was a duo queue).

That's the other problem with the game. It had a wider ELO matchmaking basis to keep match making from taking too long. Something that doesn't happen in other ELO systems. The system should be more dynamic in that it has tighter ELO match makings when more are on and less when there are less. It should also group new players to ranked more often with other new players.

As for the 2K I threw out, that was not a hard and fast number either. It was an example. I stated instead of starting in the middle have them start higher on the scale. 2000 was the number I threw out because that is a 3/4ths level in the current ELO scale that LoL uses. Basically start new players at 3/4ths, and try to keep them grouped up as best as can be done with other new players to ranked. Have their rankings fluctuate a bit more based off scores.

Interesting idea, but once again you penalize supports that willfully give up cs. Their CS looks bad, they won't get kills, they may even die for their team and they certainly won't get towers. Best case they go 0/0/20. How do you weight tower then? If I'm alistar with a 4v5 mid and I have a Yi/TF/Sivir/etc split pushing and we trade one or two deaths for a tower, I'm ok with that (if we keep our tower). They lose a lot of map control by losing that tower. If instead we've got one dead and we're 3v5 mid and TF continues to push instead of teleporting mid, that's bad but in both cases they can choose to take that tower and I don't know how you differentiate between the two cases using statistics. Another poor player that may get rewarded is the mid that constantly ganks but loses their tower to a Caitlyn that pushes non-stop. The mid can end up with a nice K but there is no way to assign that tower loss to them.

I appreciate that you're throwing real ideas out here, I just don't find any of them to be quite feasible due to the complexity of the situation. If you really think duo-queing is the answer, hit me up when I get back from vacation, I've never trolled a ranked game.

Again, that's why it's statistics and not hard and fast examples. On AVERAGE you would use those scores. You can also change weights based on the type of character. Sona for example would have a much higher weight for some of her game play stats compared to other characters. As you've pointed out, this is something easily recognizable and thus can be accounted for. Sona isn't going to get as many minion kills as say Ashe, so her having less minion kills won't factor in as much for an elo penalty. Her not having as many assists as Ashe WOULD factor in hugely though for the type of character she is. The game already categorizes champs into certain types so the developers and players know the capabilities and roles generally of all champs.