|
United States47024 Posts
On June 12 2012 05:24 N3rV[Green] wrote: You missed the most important thing in my eyes, which is the ability to prevent the carry from reaching their next item by killing them and taking their gold away.
If you kill the AD carry in LoL, ya that's great, but he still has the same amount of money, and is barely behind in when he will get that next big item. Where in DotA you can bring a carry saving for radiance (3800 takes a while to get) down a LOT of gold by taking them out repeatedly preventing them from snowballing even if they got some advantage early. There's actually 2 separate issues here:
1) The gold differential that arises from kills. I don't consider the fact that you're *losing* gold to be significant (and in the grand scheme of things, you only lose 30xLevel gold per death in DotA--the creeps lost while you're dead are much more significant than the actual gold loss). What matters is the gold differential that is created between the killer and the victim--which translates into laning advantage when the victim comes back to lane. It's ironic that while LoL advertises the fact that you don't lose gold on death to be less punishing, the relative gold differential created by a kill in LoL is actually MORE than in DotA, particularly when measured in number of creeps.
2) The large timing window where a carry is simply banking gold towards a large item. As you mentioned this phase can take on the order of 3k-5k gold to save. This also ties into slot-efficiency, as oftentimes, a carry will fill out their slots on middle-tier teamfighting items, and then use their midgame teamfighting efficiency from those items to try and carry them over the large gap where they're saving for a big item.
Such timings are much harder to exploit in LoL, though they still do exist. The main thing is that comparably large items like NLR and BF are often completed during laning phase.
On June 12 2012 05:44 Shikyo wrote: In my opinion the best way to deal with this would be to introduce many more Doran-like items that are really really strong for their cost but are slot-inefficient.
It might not be enough but it'd be a step in the right direction I agree.
Such small cost-effective, slot-ineffective items would also help support itemization a lot, particularly if design space using item actives is explored.
|
Netherlands45349 Posts
Yango I lu
Don't forget that Dota has grossly ''overpowerd'' abilities on very long cooldowns, something that LoL doesn't have which allows teams to take fights even though they are behind on items or when aquiring a single item(Blink Dagger on Sand King/Earthshaker) but that is a core design idea I believe, so its not going to change.
|
United States47024 Posts
On June 12 2012 06:11 Kipsate wrote: Yango I lu
Don't forget that Dota has grossly ''overpowerd'' abilities on very long cooldowns, something that LoL doesn't have which allows teams to take fights even though they are behind on items or when aquiring a single item(Blink Dagger on Sand King/Earthshaker) but that is a core design idea I believe, so its not going to change. I have a long PM that I sent to someone actually explaining all these points. Wasn't going to copy-paste it here because I figured it wouldn't be interesting to everyone, but it seems everyone brought up all the individual points separately anyway. -_-
And I still think Medallion would be the easiest item ever to port from DotA. Great utility, right pricing range, easy to rescale armor/MR shred numbers to whats appropriate for LoL, useful for lots of different roles. It's like the perfect item active for them to port, but they still haven't done it.
|
Say that blue picked up a double kill with top and jungler each getting kills invading the first red buff of the red team after both jungler's started at blue buff. At 12 minutes the 10% gold advantage is in Blue's favor 16.5k - 13.5k, and the current gold values are Blue AD 3.5k, Blue mid 3.5k, blue support 2k, blue jungle 3.5k, blue top 4k. Red team is the same gold for mid, ad, and support, so that leaves a total of 4.5k split between jungle and top lane. So before looking at the experience differences in the 4 champions, how do you expect red top/jungle to control top lane with 4.5k gold between them, when blue top/jungle has 7.5k gold between them? You're going to have to bring the mid lane up to top and hope for a gank that leads to a 3k gold swing.
Not only is there a big gold difference in that senario, but their is probably a level difference as well. Leading to timings where the lane that is already behind has to either surrender gold and experience further, or they're just going to die and lose more gold and experience.
Another way to get to this magic 10% gold advantage, is every lane is getting out farmed in a 0 kill game. 45 cs to every 55cs in every lane, or two lanes and jungle are the same, but one lane is like 25 cs - 75 cs. If you have a 3-1 cs advantage that's absolute lane domination, and in a solo lane, it can easily allow you to carry your team to victory. I actually don't find this surprising at all actually. If I get a 3:1 cs advantage in a solo lane, I expect to be able to carry that game 90% of the time. If we're all out farming the enemy by 10%, I expect to be able to win that 90%. If the enemy jungler is 3-0-2, and the score is 0-5 in 12 minutes, I think we're going to have to play a 40 maybe even 60+ minute game if we're going hope to come back.
|
United States47024 Posts
On June 12 2012 06:14 Sabin010 wrote: Say that blue picked up a double kill with top and jungler each getting kills invading the first red buff of the red team after both jungler's started at blue buff. At 12 minutes the 10% gold advantage is in Blue's favor 16.5k - 13.5k, and the current gold values are Blue AD 3.5k, Blue mid 3.5k, blue support 2k, blue jungle 3.5k, blue top 4k. Red team is the same gold for mid, ad, and support, so that leaves a total of 4.5k split between jungle and top lane. So before looking at the experience differences in the 4 champions, how do you expect red top/jungle to control top lane with 4.5k gold between them, when blue top/jungle has 7.5k gold between them? You're going to have to bring the mid lane up to top and hope for a gank that leads to a 3k gold swing.
Not only is there a big gold difference in that senario, but their is probably a level difference as well. Leading to timings where the lane that is already behind has to either surrender gold and experience further, or they're just going to die and lose more gold and experience.
Another way to get to this magic 10% gold advantage, is every lane is getting out farmed in a 0 kill game. 45 cs to every 55cs in every lane, or two lanes and jungle are the same, but one lane is like 25 cs - 75 cs. If you have a 3-1 cs advantage that's absolute lane domination, and in a solo lane, it can easily allow you to carry your team to victory. I actually don't find this surprising at all actually. If I get a 3:1 cs advantage in a solo lane, I expect to be able to carry that game 90% of the time. If we're all out farming the enemy by 10%, I expect to be able to win that 90%. If the enemy jungler is 3-0-2, and the score is 0-5 in 12 minutes, I think we're going to have to play a 40 maybe even 60+ minute game if we're going hope to come back. The problem is that this leaves no room for error for teams that are inherently oriented toward one phase of the game or another. If lategame-oriented teams that are intentionally picked with weaker early-games are not sustainable with only a 10% gold disadvantage, then that has potential ramifications in terms of champion draft diversity.
|
On June 12 2012 06:05 red_ wrote:Show nested quote +On June 12 2012 05:45 Klive5ive wrote: Wow; for TL I would've expected a better understanding of stats.
"The team that takes as little as a 10% gold lead by 12 minutes wins over 90% of the time"
The key word is in bold. What that means is the stats take into account HUGE advantages at 12 minutes too. That means you cannot infer that if a team is up 10% they will win 90% of the time!!! You would have to take stats for teams up exactly 10% and no more to find the expected win percentage. No, you're putting words in the OP's mouth(as did others). How can I be putting words in his mouth when I'm quoting him? I'll just leave this here...... "Imagine watching game 7 tonight and seeing the Celtics go up 45-40 late in the second quarter and being able to say with 90% certainty that they are going to the finals."
If you read the "evidence" it gets even worse. Right at the top: Fnatic v TSM.evo GM2---------TSM.evo---------TSM.evo-----------lead taken at 15:40 So he decided to take the reading at 15 minutes 40 instead? He does this again in the notes later on. Moving the lead to another time.
I know I'm being pedantic but this is a pretty clear example of bad statistics. Stats used badly to justify a conclusion that has already been decided.
The title should say "The leader in gold at 12min wins over 90% of the time (as long as you ignore close games)"
|
so obvious.
12 min is end of laning phase. Winner of laningphase has more gold, and thus with a decent team makeup and won lanes the upper hand.
|
What i would like to know is, dose this also happen in DOTA? if not this will be a big drawback for LoL as a e-sport vs DOTA. Let see the average LoL game is ~30min, this would mean that after around a 3rd of the game you already know who will most likely win. In SC its rare in the pro scene that after a 3rd of total game time you can tell who will win.
|
Works the same in Starcraft. If you are behind, make smart moves and force mistakes from your opponent.
|
On June 12 2012 07:03 Goozen wrote: What i would like to know is, dose this also happen in DOTA? if not this will be a big drawback for LoL as a e-sport vs DOTA. Let see the average LoL game is ~30min, this would mean that after around a 3rd of the game you already know who will most likely win. In SC its rare in the pro scene that after a 3rd of total game time you can tell who will win. It still happens in dota but not as much. Dota is more diverse in the meta/items so you can't really tell who's gonna win unless they have the lead and a better mid game/late game comp on top of that. There's mid game items and late game items as well. Take my anecdotal evidence for what you will, but snowballing happens a LOT more often in league of legends, and it's not always due to one team playing better all the way through.
|
On June 12 2012 07:03 Goozen wrote: What i would like to know is, dose this also happen in DOTA? if not this will be a big drawback for LoL as a e-sport vs DOTA. Let see the average LoL game is ~30min, this would mean that after around a 3rd of the game you already know who will most likely win. In SC its rare in the pro scene that after a 3rd of total game time you can tell who will win.
People have outlined the differences of Dota to LoL in this thread so I suggest just paging through. I think one thing no one has mentioned is the pick/ban phase. I think LoL would benefit from switching to a dota style system with the 2 bans after first 3 picks. Feels like currently people can't really deviate from the norm since they don't know what the opponent is doing when they banning so can't actually do anything other than guess. We might see some more interesting things than protect the Kog with this I think.
EDIT:
On June 12 2012 07:29 Itsmedudeman wrote:Show nested quote +On June 12 2012 07:03 Goozen wrote: What i would like to know is, dose this also happen in DOTA? if not this will be a big drawback for LoL as a e-sport vs DOTA. Let see the average LoL game is ~30min, this would mean that after around a 3rd of the game you already know who will most likely win. In SC its rare in the pro scene that after a 3rd of total game time you can tell who will win. It still happens in dota but not as much. Dota is more diverse in the meta/items so you can't really tell who's gonna win unless they have the lead and a better mid game/late game comp on top of that. There's mid game items and late game items as well. Take my anecdotal evidence for what you will, but snowballing happens a LOT more often in league of legends, and it's not always due to one team playing better all the way through.
I think the problem is also that in Dota you can get teams that peak at different stages of the game so while one team might be winning earlier they also peak earlier so if they can't gain a lead that wins them the game they other team still has a chance when they themselves peak. Not sure if what I mentioned above might allow LoL to gain a similar system.
One thing I noted in this weekend is that even when teams are so far ahead games still drag on for every long even though the outcome never seemed to change unless you Dignitas going for baron. I'm not sure if this is an issue of the players not knowing exactly how to finish off a game or the way LoL works. Feels like without items("Aura actives") that extend to creeps games go on for far longer than they should.
|
Thanks for putting all this info up. While I've (and many others I'm sure) have semi-assumed this. It's nice to have some data.
And it is sad to be honest. Especially with the statistics being as highly skewed as they are. Like someone said. Pretty much 2/3rds of the game is being played knowing no matter how hard you try, you pretty much can't win and are just playing out of sheer desperation. Which is why when a team is up 1-0 and it gets to a bad point. You almost always see a 25-30 minute surrender. Because they know the game just isn't worth trying since they aren't in desperation to win.
Riot did make a step in the right direction allowing for more exp to be gotten from higher level champions. But it doesn't seem enough. I constantly find myself turning games off because you can clearly see the outcome of the game. And it's boring as hell to watch a team who is already ahead just win teamfights with ease and secure every global objective and then teamfight some more and win easily.
I think this really needs to be worked on to allow teams to not be as punished for small mistakes. But that's just my opinion.
|
On June 12 2012 07:35 Numy wrote:Show nested quote +On June 12 2012 07:03 Goozen wrote: What i would like to know is, dose this also happen in DOTA? if not this will be a big drawback for LoL as a e-sport vs DOTA. Let see the average LoL game is ~30min, this would mean that after around a 3rd of the game you already know who will most likely win. In SC its rare in the pro scene that after a 3rd of total game time you can tell who will win. People have outlined the differences of Dota to LoL in this thread so I suggest just paging through. I think one thing no one has mentioned is the pick/ban phase. I think LoL would benefit from switching to a dota style system with the 2 bans after first 3 picks. Feels like currently people can't really deviate from the norm since they don't know what the opponent is doing when they banning so can't actually do anything other than guess. We might see some more interesting things than protect the Kog with this I think. EDIT: Show nested quote +On June 12 2012 07:29 Itsmedudeman wrote:On June 12 2012 07:03 Goozen wrote: What i would like to know is, dose this also happen in DOTA? if not this will be a big drawback for LoL as a e-sport vs DOTA. Let see the average LoL game is ~30min, this would mean that after around a 3rd of the game you already know who will most likely win. In SC its rare in the pro scene that after a 3rd of total game time you can tell who will win. It still happens in dota but not as much. Dota is more diverse in the meta/items so you can't really tell who's gonna win unless they have the lead and a better mid game/late game comp on top of that. There's mid game items and late game items as well. Take my anecdotal evidence for what you will, but snowballing happens a LOT more often in league of legends, and it's not always due to one team playing better all the way through. I think the problem is also that in Dota you can get teams that peak at different stages of the game so while one team might be winning earlier they also peak earlier so if they can't gain a lead that wins them the game they other team still has a chance when they themselves peak. Not sure if what I mentioned above might allow LoL to gain a similar system. I'd also like to point out baron. Baron is the easiest way to make a comeback, but also the easiest way for a leading team to close it out. If a team has an edge and they grab baron the game is pretty much over. I'll give a suggestion but I know riot will never implement this into the game. When's the last time they made a drastic change to try and alter the game? What they COULD do is have baron follow the team out if they engage. Obviously I don't really know what would happen if that were changed. Could it make the game worse? Yes. It could make it so team fights don't happen for a very long time, but we won't know how the game would change after 4-5 months, but for now, entire games are decided at baron.
|
In terms of basic design I feel that Riot purposely evens out the scaling all over the place as long as I can remember. The argument being that strong late- and strong earlygame champs are "unfun" means everyone and their mom scales the same basically.
While champions still scale differently the differences between them (think Cait vs Vayne) are too small to make them actually shine in their respective strong phases. I feel that strengths during certain time frames add more variety, allow for more complex strategies and therefor make the game more fun in the long run.
The question is whether balancing around what top teams can do with this is okay, even if it makes the random solo q dude cry on the forums about "imbalance" because he doesn't understand the underlying concepts.
|
On June 12 2012 06:26 Klive5ive wrote:Show nested quote +On June 12 2012 06:05 red_ wrote:On June 12 2012 05:45 Klive5ive wrote: Wow; for TL I would've expected a better understanding of stats.
"The team that takes as little as a 10% gold lead by 12 minutes wins over 90% of the time"
The key word is in bold. What that means is the stats take into account HUGE advantages at 12 minutes too. That means you cannot infer that if a team is up 10% they will win 90% of the time!!! You would have to take stats for teams up exactly 10% and no more to find the expected win percentage. No, you're putting words in the OP's mouth(as did others). How can I be putting words in his mouth when I'm quoting him? I'll just leave this here...... "Imagine watching game 7 tonight and seeing the Celtics go up 45-40 late in the second quarter and being able to say with 90% certainty that they are going to the finals." If you read the "evidence" it gets even worse. Right at the top: Fnatic v TSM.evo GM2---------TSM.evo---------TSM.evo-----------lead taken at 15:40 So he decided to take the reading at 15 minutes 40 instead? He does this again in the notes later on. Moving the lead to another time. I know I'm being pedantic but this is a pretty clear example of bad statistics. Stats used badly to justify a conclusion that has already been decided. The title should say "The leader in gold at 12min wins over 90% of the time (as long as you ignore close games)"
Only smart guy here lol. Throw out the games where team has 20%+ lead. If you find that too arbitrary, find the times a team has come back from gold deficit.
|
On June 12 2012 12:21 r.Evo wrote: In terms of basic design I feel that Riot purposely evens out the scaling all over the place as long as I can remember. The argument being that strong late- and strong earlygame champs are "unfun" means everyone and their mom scales the same basically.
While champions still scale differently the differences between them (think Cait vs Vayne) are too small to make them actually shine in their respective strong phases. I feel that strengths during certain time frames add more variety, allow for more complex strategies and therefor make the game more fun in the long run.
The question is whether balancing around what top teams can do with this is okay, even if it makes the random solo q dude cry on the forums about "imbalance" because he doesn't understand the underlying concepts. Well, you CAN abuse strong earlygame champs in LoL, you just have to tailor your whole team towards it. But you can't really abuse strong lategame champs with weak earlygame. Those champs just aren't strong enough lategame to offset their earlygame disadvantages. Let's see, who's weak early these days? Nasus: he's also a joke during lategame if he didn't get to keep up in farm with everyone. Poppy: At least until now it doesn't seem like people are solid enough with her to make her consistently strong early. But she definitely is strong lategame... So yeah Poppy might actually work. Anivia would be one of these strong lategame champs, but she happens to be strong all game long.
Meh, champs really seem to streamlined these days, can hardly find any champs who have a distinctive lategame powercurve... Poppy definitely is the best example, and you can clearly see how strong she is if she doesn't get shut down completely. But apparently people hate her for being inherently powerful lategame.
Also, at best you can fit 1 of these weak early, strong late champs into your team. Any more and you will get crushed by an earlygame team. You could blame Dragon + Buffs + the way top lane works for that I guess.
|
1) Cause and effect:
The OP has presented data which shows that a 10% gold lead by 12 minutes leads in 90% of the cases to win for the leading team. People here interpret this in a way which indicates that the rest of the game is "meaningless". You could also interpret this in way that the stronger team will manage to pull ahead early and this shows.
2) What about the other games?
90% is an impressive number but how many games show such a clear advantage after 12 minutes? So what is the total number of games which show a clear advantage after 12 minutes and how many are about even? So how much percentage of the total games are decided by 12 minutes?
3) The better team looks better X)
In any sport that you are familiar with you have a good idea about the likely outcome after some time into the game. 12 minutes is a quarter of a "regular" lol game. In soccer I for example will also have a very clear idea, which team will most likely win after around 22 minutes. Doesn't mean I can't enjoy the the remaining 68 minutes.
4) Dota
In Dota the gold graph is only half the story because XP/Lvl are also very important. Furthermore buybacks/TPs/"Lost gold" are not taken into account. You could improve that by showing a graph which shows the total gold value of the team, which would be all items+saved gold. The gold graph in dota tells you how well a team is farming not how well they are doing overall.
Conclusion)
The gold graph is a meaningful tool to asses which team is in the lead. It also seems to indicate that comebacks in lol are difficult. I think everybody here knows that and hopefully they will try to change that (if they can).
The gold graph is a symptom of that, not the problem. Cause and Effect
|
Eh, throw out any data of shitty teams playing real pros. MRN will typically have a 10%+ gold deficit to TSM at 12 minutes of any game they play. I'm not just shitting on MRN, but they're a pretty good example of the types of matches that are predetermined before the matches are played.
In what world would Redact, or Wolf Pack ever beat TSM, M5, etc?
|
On June 12 2012 06:26 Klive5ive wrote:Show nested quote +On June 12 2012 06:05 red_ wrote:On June 12 2012 05:45 Klive5ive wrote: Wow; for TL I would've expected a better understanding of stats.
"The team that takes as little as a 10% gold lead by 12 minutes wins over 90% of the time"
The key word is in bold. What that means is the stats take into account HUGE advantages at 12 minutes too. That means you cannot infer that if a team is up 10% they will win 90% of the time!!! You would have to take stats for teams up exactly 10% and no more to find the expected win percentage. No, you're putting words in the OP's mouth(as did others). How can I be putting words in his mouth when I'm quoting him? I'll just leave this here...... "Imagine watching game 7 tonight and seeing the Celtics go up 45-40 late in the second quarter and being able to say with 90% certainty that they are going to the finals." If you read the "evidence" it gets even worse. Right at the top: Fnatic v TSM.evo GM2---------TSM.evo---------TSM.evo-----------lead taken at 15:40 So he decided to take the reading at 15 minutes 40 instead? He does this again in the notes later on. Moving the lead to another time. I know I'm being pedantic but this is a pretty clear example of bad statistics. Stats used badly to justify a conclusion that has already been decided. The title should say "The leader in gold at 12min wins over 90% of the time (as long as you ignore close games)"
Glad I'm not the only one who looked at the stats and decided he manipulated the hell out of them.
Hmm... this close game with two really good teams goes against my thesis... better mark it down as a tie. Dafuq...
|
Awesome post!
I've been getting more into playing and watching LoL myself, and I was also feeling the drag of the late game.
I have seen some cool back and fourth in tourney play based on Ninja baron kills, but many of the games do seem to be a slow grind.
I wonder, is the data available to do a similar study of Dominion? That game mode seems very fast paced and dynamic, I wonder why we don't see pro play there?
|
|
|
|