Welcome to the Forum Archive!

Years of conversation fill a ton of digital pages, and we've kept all of it accessible to browse or copy over. Whether you're looking for reveal articles for older champions, or the first time that Rammus rolled into an "OK" thread, or anything in between, you can find it here. When you're finished, check out the boards to join in the latest League of Legends discussions.

GO TO BOARDS


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

Comment below rating threshold, click here to show it.

Charmicarmicarmicat

Senior Member

02-16-2010

Quote:
EasymodeX:
THIS. You kind of slipped past this point, but this is major. In DotA, cheap items are way more efficient than expensive items. Even if you are behind by 1000 gold versus your opponent, if you spend on cheap items (Bracer stack) instead of saving up 3200 for a Reaver, you can still keep up with their character strength (or even exceed it in the short-term).

One of the #1 problems with LoL's item system and growth of champions is that the low-tier and mid-tier items are not highly efficient compared to the 3k+ items. Haunting Guise, Brutalizer -- all these items should be vastly more efficient than an IE.

Big items do not have a premium cost from fewer inventory slots in LoL. This, combined with the lack of a gold drain on death, means a losing team cannot strategically gank an enemy carry. The losing team cannot invest in mid-tier items or low-tier items to gain an early/mid advantage.

Fed players should gain an advantage from raw power totals, not from efficient stats. Sure, a fed <insert carry> can have their Trinity Force or IE. However, a player who has less gold should be able to buy a Brutalizer + Exec Calling and be almost as effective. In a similar way where, between a Reaver and 5x Bracers ... the Bracers are far and away superior.


I think that you make a good point Easymode. My guess is that Riot has chosen dileberately not to make it so that smaller items are more cost effective than big ones.


Comment below rating threshold, click here to show it.

Charmicarmicarmicat

Senior Member

02-16-2010

Quote:
Heero Maki:
flash shouldn't be a free summoner ability period it's to strong of an ability. you should be forced to spend about 3k in game to buy such a strong ability.


Actually, it's worth it on every hero except maybe kassadin and shaco, but even that is a stretch


Comment below rating threshold, click here to show it.

Constrictor

Recruiter

02-16-2010

Quote:
Larias:
TiberiusAudley, we were not discussing 19 stacks. We were discussing 20 stacks. The player in question said Leviathon, due to "complex math," was worth 3500g at max stacks, and that Mejai's was worth "over 5000."

If you want to discuss level 19 stacks, feel free


FINE! You're going to make me re-do all my calculations again, aren't you?

Here we go:

First, we need to calculate the gold value of each stat point for the various stats involved. To get this, we start with the most basic items.

To do this, I simply added the total cost of all such base items and divided by the total stat points of all such base items.

EXAMPLE: Attack Damage value = (Long sword 415 + Pickaxe 975 + BF Sword 1850) / (Long sword 10 + Pickaxe 25 + BF Sword 1850) = 38.12

So we see attack damage has an expected gold cost of 38.12.

We do this for all base values. I won't list them out here, but I have them all saved on a spreadsheet.

Now we have to derive some values because we don't have base items with just that stat on them, example: Movement speed %, Critical Damage %, etc.

To get these values, we use Quintessence expected values. First we calculate the actual quintessence value for those items which we found in step one.

Example:

  • Attack damage is 38.12 gold per point.
  • Attack damage qunit gives 2.25 points
  • Attack damage quint gold value = 38.12*2.25 = 85.77
We do this for all such values (10 total). From here, I added up all of the quint values and divided by the total (10). This gives me an expected average quint value (82.44 if you care)

From here, I can use the expected end value and the quint bonus to calculate the gold per point cost of things such as cooldown reduction.

The values we care about for the snowball items are:

  • Cooldown reduction -> 50.27 gold per point
  • Critical damage -> 18.48 gold per point
  • Movement speed -> 54.96 gold per point
Percentage damage reduction is still not calculable using this method, so we need to use another method, which I will explain now.

Percentage damage reduction can be achieved in three ways:

  • Adding armor and magic resist
  • Taking the % damage reduction mastery
  • Getting full 20 stacks on Leviathan
As reducing the % of damage taken gets more costly the more you currently reduce, the value of the % reduction stat goes up as your base magic resist and armor goes up.

That means, the higher your base levels of magic resist and armor, the more % damage reduction is doing for you in comparison to just getting armor and magic resist.

Example:

At 0 armor, with full stack Leviathan, you would take 85 damage from a 100 point hit. In order to get the same reduction from armor/magic resist, you would need 18 points of defense.

100 * (100/118) = 84.75 damage taken.

So, as you gain armor/magic resist, the armor/magic resist value of % damage reduction increases, according to the following schedule:

  • Base (0) = 15% reduction is worth 18/18 armor/mr
  • Base (50) = 15% reduction is worth 27/27 armor/mr
  • Base (100) = 15% reduction is worth 35/35 armor/mr
  • Base (150) = 15% reduction is worth 45/45 armor/mr
  • Base (200) = 15% reduction is worth 53/53 armor/mr
  • Base (250) = 15% reduction is worth 63/63 armor/mr
  • Base (300) = 15% reduction is worth 70/70 armor/mr
In a typical game, a tank will probably have ~150 - 200 armor/magic resist, which is then reduced by enemy player masteries/runes/and items. I figure a base value of 150 for the purposes of this review is fair, though the numbers don't change that significantly at other values near the norm.

With all that out of the way, we can now compare the values of all three snowball items at each charge level:

Blade of the Occult:

Charges - Value:

0 381.1764706
1 494.3784706
2 607.5804706
3 720.7824706
4 833.9844706
5 947.1864706
6 1060.388471
7 1173.590471
8 1286.792471
9 1399.994471
10 1513.196471
11 1626.398471
12 1739.600471
13 1852.802471
14 1966.004471
15 2079.206471
16 2192.408471
17 2305.610471
18 2418.812471
19 2532.014471
20 3469.616471


Leviathan:

Charges - Value

0 479.5081967
1 564.7528
2 649.9976
3 735.2424
4 820.4872
5 905.732
6 990.9768
7 1076.2216
8 1161.4664
9 1246.7112
10 1331.956
11 1417.2008
12 1502.4456
13 1587.6904
14 1672.9352
15 1758.18
16 1843.4248
17 1928.6696
18 2013.9144
19 2099.1592
20 3611.19

*Note - this is using our expected base armor/mr of 150. The end value of Leviathan changes based on your base level of armor/magic resist. At 100 base, the end value becomes 3294.13 and at 200 base, the end value becomes 3864.84.


Mejai's Soulstealer:

Charges - Value

0 431.6666667
1 647.496
2 863.326
3 1079.156
4 1294.986
5 1510.816
6 1726.646
7 1942.476
8 2158.306
9 2374.136
10 2589.966
11 2805.796
12 3021.626
13 3237.456
14 3453.286
15 3669.116
16 3884.946
17 4100.776
18 4316.606
19 4532.436
20 5502.29

So, there you have it. ALL the data.

Make of it what you will, but what the numbers tell me is that Mejai's gives the most gold value per charge (breaking even at only 4 charges, while the other items break even at around 9 charges). Mejai's also has the largest total gold value when fully charged. Approximately 5500 vs 3500, which is nearly 60% better.

For the sake of balance, they need to either nerf Mejai's or buff the other two snowball items.


Comment below rating threshold, click here to show it.

SlyGoat

Member

02-16-2010

Quote:
EasymodeX:
THIS. You kind of slipped past this point, but this is major. In DotA, cheap items are way more efficient than expensive items. Even if you are behind by 1000 gold versus your opponent, if you spend on cheap items (Bracer stack) instead of saving up 3200 for a Reaver, you can still keep up with their character strength (or even exceed it in the short-term).

One of the #1 problems with LoL's item system and growth of champions is that the low-tier and mid-tier items are not highly efficient compared to the 3k+ items. Haunting Guise, Brutalizer -- all these items should be vastly more efficient than an IE.

Big items do not have a premium cost from fewer inventory slots in LoL. This, combined with the lack of a gold drain on death, means a losing team cannot strategically gank an enemy carry. The losing team cannot invest in mid-tier items or low-tier items to gain an early/mid advantage.

Fed players should gain an advantage from raw power totals, not from efficient stats. Sure, a fed <insert carry> can have their Trinity Force or IE. However, a player who has less gold should be able to buy a Brutalizer + Exec Calling and be almost as effective. In a similar way where, between a Reaver and 5x Bracers ... the Bracers are far and away superior.


This is startlingly insightful. The only low tier items I build in LoL that don't build into something bigger are Chalice and Guise. Most of the low tiers just aren't as beneficial in the short term to make up for the long term power of high tier items.

I agree with this entirely. LoL needs more low tier combination items that are cost effective to build, because they come at the cost of more item slots, which means you'll have to sell them back later. The only problem with this, though, is the sell back for most items in LoL is 75% instead of 50%, so it might be too beneficial unless all items are made 50% sellback (which, honestly? doesn't seem like a huge deal to me. the items you end up selling back are most often gold over time items which are already 50%).


Comment below rating threshold, click here to show it.

Karandor

Adjudicator

02-16-2010

Quote:
EasymodeX:
THIS. You kind of slipped past this point, but this is major. In DotA, cheap items are way more efficient than expensive items. Even if you are behind by 1000 gold versus your opponent, if you spend on cheap items (Bracer stack) instead of saving up 3200 for a Reaver, you can still keep up with their character strength (or even exceed it in the short-term).

One of the #1 problems with LoL's item system and growth of champions is that the low-tier and mid-tier items are not highly efficient compared to the 3k+ items. Haunting Guise, Brutalizer -- all these items should be vastly more efficient than an IE.

Big items do not have a premium cost from fewer inventory slots in LoL. This, combined with the lack of a gold drain on death, means a losing team cannot strategically gank an enemy carry. The losing team cannot invest in mid-tier items or low-tier items to gain an early/mid advantage.

Fed players should gain an advantage from raw power totals, not from efficient stats. Sure, a fed <insert carry> can have their Trinity Force or IE. However, a player who has less gold should be able to buy a Brutalizer + Exec Calling and be almost as effective. In a similar way where, between a Reaver and 5x Bracers ... the Bracers are far and away superior.


This is something I never thought of before and I completely agree. Though it does exist in some form with Heart of Gold. High end items are probably all too efficient and it kills any chance of stacking a bunch of cheap items to compete.

The only problem I see is teams abusing this and just grabbing 5-6 low tier items and finishing a game extremely quickly.


Comment below rating threshold, click here to show it.

Charmicarmicarmicat

Senior Member

02-16-2010

Quote:
Constrictor:
...alot of number crunching


Wow i can't believe you did that x)


Comment below rating threshold, click here to show it.

AdmiralSnackbar

Recruiter

02-16-2010

Quote:
Zileas:
Our philosophy on this issue is that individual snowballing is good, and team snowballing is bad. Right now, team snowballing occurs too much -- one super carry twitch is a lot easier to handle than 5 guys who are 3 levels higher than each of you.

We are going to make some experience table tweaks to reduce team snowballing without hurting individual snowballing much. We are also going to make a new 3 lane map that is larger, has less cross-map connectivity (so that there is higher cost to 5 man roam, and less shutdown of the farming/laning phase via the first tower being down), etc.

We probably will nerf flash.

We wont be nerfing snowball items. Emotional highs from the game you super carried make this genre sticky. We removed it too much in the beta, and it cost us. Sorry, but being super carry annie or super carry corki once in a while gets a player through the 8 defeat in a row streak they hit the next day.

I'm on the fence about early gold for kills. It's been something we have been talking about, for the reason you cite. OF course, an early kill is also harder than a late kill.


Sweet, that'll all happen right after you release 20 more champions and skins, and balance/add the 3v3 map to the MM, right? So I can uninstall and wait till 2011?


Comment below rating threshold, click here to show it.

Constrictor

Recruiter

02-16-2010

Quote:
Charmicarmicarmicat:
Wow i can't believe you did that x)


I did the calculations before this thread popped up, for my own education/benefit.


Comment below rating threshold, click here to show it.

Charmicarmicarmicat

Senior Member

02-16-2010

I think you had a braingasm.


Comment below rating threshold, click here to show it.

L0CUST

The Council

02-16-2010

Zeal / Executiners / Brutalizer are very cost efficient >:O

Last Whisper / Void Staff / Catalyst are too..

and so are the elixirs..

-L0CUST