[Essay] Snowballing - Why DotA had it right and LoL is still learning

First Riot Post
Comment below rating threshold, click here to show it.

Miles Long

Senior Member

02-16-2010

I like your post Charmi. It neatly sums up everything I had become painfully aware of a while ago. If I could +10 your post you can be sure I would.

Quote:
Originally Posted by Asfastasican View Post
http://www.leagueoflegends.com/board...580#post619580

Some of these snowballing problems tie into why 5 man roaming is bad and why it exists...

Zileas' response is VERY surprising to me, considering that he just said that they want to encourage individual snowballing and discourage team snowballing, but everything they have done in the pre-development and in beta suggests that they did the exact opposite...

Pardon my ellipses, but don't most of you remember when beta first opened up and went into its 24/7 phase? Most of the devs, including players such as Larias yada yada yada would say "You can't push a base on your own, you will always need to move into the enemy base as a team and put in a group effort to finish the round" to the newcomers? Uh, oops? Seems like the design focus has shifted now in order to repair the meta-game, yet Summoner's Rift and other game mechanics haven't been tweaked to allow the devs to shift their focus.

Let's face it. Everybody has recall in this game and great designed items like scroll of town portal were s****ped (<-- s-c-r-a-p-p-e-d) early on for some odd reason. Also, it doesn't help that League is basically a short version of EM mode where the default map is a map that's just too small. You have a game where you have to deal with a number of design issues. It's hard to focus on real balancing dillemmas when you're still focusing on your foundation, such as your UI or game mechanics. Sure, a Ryze remake is great and all, but seriously flawed champions like Warwick are in much more desperate need of a remake. Meanwhile, new champions are being introduced to help fund the game via a microtransaction system and the playstyle of professional teams still strongly discourage players to play champions like Warwick. Having some variety in your team and trying out new strategies gets you majorly pwnt!

I can see why draft mode is taking forever to come out and the 6v6 hasn't been released yet. This MOBA game is an excellent game for casuals and its very fun. But still, its core game is just flawed when it comes to developing a challenging and interesting meta game. Especially with design philosophies such as "We don't want to make champions skills that could be used to troll or greive other teammates" still remain as one of the game's tenants. It's hard to expand on the meta-game when you're afraid of taking risks during development stages.

How can you introduce draft mode when most of your hardcore players are still concerned over a single champion and drafted 5v5's will most likely be the same 5-6 champions, let alone the same top 10?

I love this game, but I don't see it being more than a game that I play with my mediocre friends. It's fun playing with them, but I don't want to spend the time or effort to train myself and train them to "go pro." Especially when the current hardcore game is focused on an invisible score that doesn't go up and down according to your individual performance. We have Dota for that and also other genres as well.

I like this game a lot, especially since its a game where most of my online friends and their friends can all agree on playing it in a casual, time-killing sense. Still, League is far from being at the stage were you can finally start working on perfecting it.
I feel like it probably should've been /thread with this post. Madra's post on page six is also worth mentioning, so props to you as well.

Bolded for emphasis, I would also like to know what Riot's true intentions are and just how long the community has to wait for any sort of real change.

Word is that more diverse maps are in the works. So what, we can add this to the growing list of important features that will come out "soon?" At this point, I don't even want to play on summoner's rift anymore and yet DotA has gotten along for years with one map.

Something is rotten in the state of Denmark.


Comment below rating threshold, click here to show it.

Charmicarmicarmicat

Senior Member

02-16-2010

Quote:
Originally Posted by Miles Long View Post
Something is rotten in the state of Denmark.
Den mark... which mark?


Comment below rating threshold, click here to show it.

Tinithor

Senior Member

02-16-2010

Quote:
Originally Posted by TrickyCat View Post
I feel that Dota snowballs much worse than LoL for the simple reason that when you die in HoN you lose gold.

Just about all your points are jsut as valid for dota too. If a melee carry gets a good first blood/double kill early game, it can spell certain doom.

Not only does killing an Opponent in HoN give you gold, you're taking away the opponent's ability to keep up at all with items.

So say you kill someone in Dota, you gain 280g, they lose 200g. The net worth of that exchange is essentially 580 gold for one kill. And now you gain about a minute and a half of farm, and he just lost a minute of farm.
I think the big difference here is how the OP mentioned that farming lanes is worth so little in LoL, in HoN like he said there are alot of smaller size battles where you can easily kill that one carry who got the double kill and knock him back just as far as he knocked you back (therefor evening it up)

Also it just seems to me that carries get alot stronger ALOT faster in LoL than in HoN or dota.


Comment below rating threshold, click here to show it.

Comforts

Member

02-16-2010

Quote:
Originally Posted by Hearteater View Post
Wouldn't increasing gold for getting kills on powerful enemies (both higher level and kill streaks) help reverse the snowball effect? If they've got a 10 kill streak carry (who obviously is higher level) and we manage a kill, 1000 gold probably isn't going to swing things enough, but 2.5k might. Optionally, make getting killed from a kill streak not reset the streak in terms of gold awarded, so a back-to-back kill of a 10-streak target might award 1000 gold for the first kill, and maybe 750 for the second (instead of 300).
I think this is the simplest and most elegant solution. May also be applied to xp gain.


Comment below rating threshold, click here to show it.

xwrogue3

Junior Member

02-17-2010

Larias: Your assumption produces a much larger difference than you suggest (500g)

Quote:
READ THIS FIRST: I use the additive, where DR is added directly to armor/MR resistances, instead of the correct formula, which is it's added after. For example, 200 * .5 * .85 is different than 200 * .35 (85 vs 70). This means that my value for Leviathan is higher than it should be. However, it is not 2000g over. At most, it's around a 500g difference.)
At 100AR:
additive: 50% + 15% = 65% DR
multiplicative (right way according to you): 1 - (50% * 85%) = 57.5%

Both your calculations are spot on for the amount of armor needed to attain this differential (I calculated it independently of you)

Additive: 100AR needs ~86AR to achieve 65% DR
Multiplicative: 100AR needs ~35AR to achieve 57.5% DR.
at 16g/AR, the difference is 816.

Assuming equal values for MR and a cost of 16.6 (MR is more expensive), the difference is 846.6 for a total difference of 1662g between the two methods using a base of 100 AR/MR.

TLDR; It matters which way the DR is combined w/ AR/MR. If the proper way to combine the 15% DR w/ AR/MR is multiplicative then the added value is very low as Constrictor pointed out.


Comment below rating threshold, click here to show it.

Viro Melchior

Senior Member

02-17-2010

Quote:
Originally Posted by Larias View Post
So I was playing around with math, and you're a little off. Note: I include aegis aura because I don't know any group that doesn't have one.

READ THIS FIRST: I use the additive, where DR is added directly to armor/MR resistances, instead of the correct formula, which is it's added after. For example, 200 * .5 * .85 is different than 200 * .35 (85 vs 70). This means that my value for Leviathan is higher than it should be. However, it is not 2000g over. At most, it's around a 500g difference.)

Starting MR of 30 (amumu), chalice, mercury treads, aura from aegis of the legion = 119, or 54.3% reduction. To get to 69.3% reduction, you would need around 226 magic resist. This is a total of 107 magic resist. Assuming 16g per point of mag resist (roughly what negatron cloak is - nullmagic cloak is actually 16.6g per point), this is 1712g.

Starting armor of 51.3 (level 12 amumu), just 1 armor item (sunfire cape, standard amumu build), and chalice aura: 120.3 armor, or 54.6% reduction. To get to 69.6%, you would need 229 armor. This is a gain of 108.7 armor. Assuming armor is valued at 16g (again, split between cloth armor price/chain vest price), this is worth 1739.2g

Health gained from Leviathan fully stacked: 640. At 2.6g per health, this is 2132.

Total gold: 5583. Not 3500.

Don't try to say "complicated math" to back up your facts.

Edit: apparently http://www.leagueoflegends.com/items..._s_soulstealer is out of date. Updated mejai's gold worth.
Oh, random note. AP is worth ~21.6 gold. Mejais fully stacked gives 220. This is 4752 gold. Cooldown reduction is cheap, roughly 21g per %(this is assuming the steep cost of Fiendish Codex. For ridiculously lower prices, check out Brutalizer), or 315g for mejai's. Total worth? 5067.
Additive: 50% armor, 15% reduction = -65% damage.
Multiplicative: 50% armor, 15% reduction = -57.5% damage.

This grossly changes your numbers, reducing the armor/MR bonus attributed to Leviathan by just over 54%.
(1712 + 1739) * .46 = 1587 gold, added to the 2132 from health value...
Is a grand total of 3719 gold.

You slam someone for saying "complicated math", and then knowingly use "bad math" instead...

Mejai's gives a far better bonus in the 1-19 stack range than either other snowball item.
Leviathan's gives the biggest bonus from the 20th stack, but not enough to eclipse the prior 19 stacks.
Sword of the Occult is ****, and only ever useful if you have high crit chance and can keep at least 8 stacks as well - and then it pales in comparison to the other two.

----------------------------------------

As for the other points in this thread...

I think the easiest ways to discourage the 5v5 and 5 man gank squads are:
  1. Make towers matter more late-game, with buffs based on nearby enemy champions. When it becomes possible to hold a tower 3v5, it will encourage players to split up some. When the defense is able to split between two lanes, it will force the offense to do the same.
  2. Remove boots from the game. As is, they are an auto-buy for every character. Replace with items that provide the smaller bonuses granted by upgrading from boots 1 to the better versions (ie decrease all advanced boot costs by 350 gold, and reduce the benefit of their speed by the bonus of boots 1). Slower overall movement will mean positioning is more important, and 5 players in the bottom lane will take much longer to reach top.
  3. Allow inner turrets to be repaired AT COST. Currently, a single ace can often decide a match 20 minutes later simply because it leaves an exposed inhibitor (no defensive turret) for the remainder of the game. Or simply make the inner turret respawn half as fast as the inhibitors.
  4. Decrease the net reward for a gank based on the participating players. 1 player is 100%. 2 nets you 90%. 3 is 80%. 4 is 70%. And 5 is 60%. Before splitting the reward.
  5. Similarly, decrease death timers as well. 100/96/90/82/72. If all 5 players get in on your kill, you respawn in 3/4 the time.

I also like the idea of more consumables as well as more efficient weak items. Perhaps include them with a +1/10 gold production as well. Worse for the long-game than the 5/10 gold items, and worse per item slot than saving up for the big items...
But better for the short term, and their money "cost" decreases the longer you hold onto them (but at just 1/10 they should never pay for themselves - a cost of 500 gold would take 2500 seconds [40+ minutes] to break even). At that point, just making them efficient upgrades to the sub-500 gold items makes them perfect. Effects like "-5% damage taken", "+10% vision range", and so forth without significantly changing the base stat of the item seem to work well to me.


Comment below rating threshold, click here to show it.

Charmicarmicarmicat

Senior Member

02-18-2010

Zileas, I played around 6 games yesterday. All of them were 10 Flash games. It's getting quiet rediculous. Shaco with flash. Kassadin with flash. Corki and tristana with flash.

There needs to be something done about this, other than another 20-30 seconds cooldown increase, it's core mechanics needs to be altered.

The metagame at higher levels is REALLY suffering from this, it's hurting and bleeding. Just about anyone worth listening to agrees that flash is by far the best summoner spell in the game. It skews the balance and makes people play passive when their flash is down. It also makes ganking a lane difficult.

There have been a number of suggestions, and the best one (I think) is making it work like blink dagger in DotA - If you get hit by a hero, 2 seconds of unusable flash. Long enough so that you cant escape during focus fire, but if you juke through brush you might make it.

Give people an incentive to pick ghost, heal or cleanse over flash - currently, there is none except on a very very limited amount of champions.


Comment below rating threshold, click here to show it.

TrickyCat

Senior Member

02-18-2010

Quote:
Originally Posted by Tinithor View Post
I think the big difference here is how the OP mentioned that farming lanes is worth so little in LoL, in HoN like he said there are alot of smaller size battles where you can easily kill that one carry who got the double kill and knock him back just as far as he knocked you back (therefor evening it up)

Also it just seems to me that carries get alot stronger ALOT faster in LoL than in HoN or dota.

nah, carries may curve in strength faster, but carries in HoN are just stupid powerful when they have farm.


Comment below rating threshold, click here to show it.

Stinyo

This user has referred a friend to League of Legends, click for more information

Senior Member

02-18-2010

I agree 100 percent. Too many of my games are decided by a Shaco who knows a quick way to grab first blood or an Annie that feeds off of my teammates. Or a Karthus who harasses until level six, essentially stealing first blood.

I was watching my buddy play HoN and I think there are a lot of things that Riot can learn from S2. I love the idea of denying. The map is bigger and the turrets target champs. Although, I hate HoN because of all the flaws (granted it's still in Beta).


Comment below rating threshold, click here to show it.

ta11geese3

Senior Member

02-18-2010

Quote:
Originally Posted by TrickyCat View Post
nah, carries may curve in strength faster, but carries in HoN are just stupid powerful when they have farm.
The point is though you can actually shut them down, before they get that much farm. And in most cases the harder they carry the weaker their early game, so they don't generally get kills early on either.