Archive | Analysis

# We Might Be Doing Teamwork the Wrong Way

I remember that someone once told me (and I can’t find a paper on it, though) that the psychological trick behind the World War II holocaust was a very clever division of responsibility. As we all know, the SS (Schutzstaffel) used to be the paramilitary group in charge of finding the people. They searched for the targets and took them in. That’s it. They weren’t supposed to bring death, at first; they just brought them in.

Next on the relay race were the other part of the SS. These soldiers took in whatever the other part of the SS brought to their doors. Their job was clear: if they send them to you, they will eventually die.  Even though this sounds extremely harsh, they had no option. They were just at the end of the race and had their orders.

If you can spot it, then congrats. If not, here it goes: The ones who searched were in a way responsible for choosing who were to die. They were the grim reapers. But they didn’t go all the way. After determining the targets, their job was over. The killers at the concentration camps then took over. However, they weren’t completely guilty in this killing. They received wagons with people. So in a way, they simply treated them as animals. There was no difference: their job was to end the job.

This doesn’t excuse the sickness of mind behind the Final Solution. Either way, we have to admit that the execution of the plan was incredible.

What does this have to do with League of Legends? A lot. While the Schutzstaffel worked in sync, LoL teams don’t seem to care about teamwork. It’s almost sad to see how little the so-called teammates work. Just by the simple idea of “KS” (kill steal) we have a bad start. We know that one works his own kill score during a game, but as long as the support doesn’t get all the lane kills that the ADC should’ve gotten, then why bother? As I’ve shown in another post, what makes a team win is not the individual gold champs have earned rather than the total gold advantage.

Inside the team there should be cooperation. Let’s see…

## co·op·er·a·tion

noun \(ˌ)kō-ˌä-pə-ˈrā-shən\

: a situation in which people work together to do something

: the actions of someone who is being helpful by doing what is wanted or asked for

So how hard can that be.
Situation: match. People: players. Something: destroy enemy nexus.

I think that many players see LoL teamwork as less important than mechanics. Wrong. Mechanics is a prerequisite. The priority at the moment of truth is organization. By minute 5, the team should get a general idea of what the others players can do. Pressing a hesitant tank to make a decisive engage in mid lane isn’t a great idea. Either help him decide with a ping or find another way to engage.

Most of us try to blame another player for a defeat. Many times it is a correct statement: he really drove the team to defeat. True or not, by minute 20 the game’s outcome was most probably already determined. So it didn’t go wrong during that fight. It wasn’t that dragon. Those problems are part of the snowball effect of the match. Those mistakes are evidence that something was going wrong and are part of the events that allow a match to end.

So there’s that. When the timer hits 10:00, you can feel which team is doing better. Turrets, CS, KS are public numbers both teams can see. Was the outcome not predictable by minute 20, but by minute 5 or before? Maybe we are just under a deterministic rule. Riot must know this. There is no large company without someone or “someones” in charge of data mining. It’s up to us to turn the rule around then.

Work.

As.

A.

Team.

Come back and tell us if identifying your teammates’ abilities works out. There’s nothing else to lose, since you thought from the beginning you were going to lose. Team OP.

Best wishes,

The no-no Java

# Dracovic5 Placement Matches Summary

Game 1 : Defeat

Role : ADC           Champion : Lucian          KDA : 4/4/14            CS : 193         Build :         Total gold : 13.2 k

Enemy ADC       Champion : Caytlin      KDA : 9/7/7       CS : 180     Build :     Total gold : 13.4 k

Game 2 : Victory

Role : Jungler      Champion : Vi          KDA : 3/1/12            CS : 79         Build          Total gold : 8.3 k

Enemy Jungler       Champion : Wukong      KDA : 1/6/3       CS : 75     Build :       Total gold : 5.8 k

Game 3 : Defeat

Role : Jungler       Champion : Vi          KDA : 4/7/13            CS : 54         Build :         Total gold : 8.6 k

Enemy Jungler      Champion : Master Yi      KDA : 11/7/6       CS : 180     Build :       Total gold : 12.9 k

Game 4 : Defeat

Role :            Champion :           KDA : -/-/-            CS : –         Build :           Total gold : – k

Enemy        Champion :       KDA : -/-/-       CS : –     Build :       Total gold : – k

Game 5 : Defeat

Role :            Champion :           KDA : -/-/-            CS : –         Build :           Total gold : – k

Enemy        Champion :       KDA : -/-/-       CS : –     Build :       Total gold : – k

Game 6 : Defeat

Role : ADC           Champion : Lucian          KDA : 4/4/14            CS : 193         Build :         Total gold : 13.2 k

Enemy ADC       Champion : Caytlin      KDA : 9/7/7       CS : 180     Build :       Total gold : 13.4 k

Game 7 : Defeat

Role :            Champion :           KDA : -/-/-            CS : –         Build :           Total gold : – k

Enemy        Champion :       KDA : -/-/-       CS : –     Build :       Total gold : – k

Game 8 : Defeat

Role : ADC           Champion : Lucian          KDA : 4/4/14            CS : 193         Build :         Total gold : 13.2 k

Enemy ADC       Champion : Caytlin      KDA : 9/7/7       CS : 180     Build :      Total gold : 13.4 k

Game 9 : Defeat

Role :            Champion :           KDA : -/-/-            CS : –         Build :           Total gold : – k

Enemy        Champion :       KDA : -/-/-       CS : –     Build :       Total gold : – k

Game 10 : Defeat

Role :            Champion :           KDA : -/-/-            CS : –         Build :           Total gold : – k

Enemy        Champion :       KDA : -/-/-       CS : –     Build :       Total gold : – k

# Approximating Win Minute?

I’ve posted before that gold seems to be the dependent variable for gold, $gold = f(ability,\: team\: composition,\: luck,\: ping,\: time)$. We might not have a direct way of getting the value of gold at any given moment, but we know that gold does express a combination of (probably) every factor in the game summed up. If gold expresses all of this variables, then depending on the gold performance of each team, we might be able to find a trend that relates gold ratio to winning minute. Shall we?

Considering the fact that I’m not that into server programming to collect massive amounts of data, I try to take representative sample i.e. some beginner matches, a few normal games, and a couple of ranked matches of different tiers. The question here is, seen in the graph, which gold advantage : minute is the match probable to finish?

I think that gold advantage might be helpful. It is just $\Delta gold = gold_{1} - gold_{2}$. If we now differentiate with respect to time, we get $\frac{\Delta gold}{dt}$ which is basically how the gold difference/advantage is changing over time. Leaving aside the assumption of a linear tendency of god earnings, which seems to fail during the last ~20% minutes of the game, all we want is to get our best gold advantage advantage, so to speak.

We just want to maximize $\frac{\Delta gold}{dt}$. We then do $\frac{gold_{1}}{d^2 t} = \frac{gold_{2}}{d^2 t}$ to find our best point. This tells us one this almost instantaneously: if minions were the only gold source and both teams farmed efficiently, then the game would just go on till the aces started to appear. To make this more interesting, adding gold to champion kills makes the gold difference snowball, I do believe.

The point is that at every moment in the map there is a fixed amount of potential gold for each team. If there were no gold rewards for kills, both teams would have the same potential gold on the map. So as the game progresses, the potential gold of the team with a level lead actually decreases. Still in most cases, the team with level advantage keeps rolling over its opponents and regardless of the extra gold available to the apparently losing team. Once again, CS is taken for granted. Until the point where the kills make a team fall behind in CS.

In a way, there is a gold cap per state on the map. Our first derivative for gold earnings tells us how efficiently each team is farming and seizing opportunities for extra gold, namely kills, dragon, jungle, and Baron. Yet, what we might want to really look at is the second derivative. This one tells us how well a team is adapting to the situations. if $\frac{gold_{1}}{d^2 t} > \frac{gold_{2}}{d^2 t}$ then team 1 is responding to the situations in a more favorable way than team 2.

Finally, approximating the win-time?

Well, as a rule of thumb, if the gold advantage by minute 20 is less than 6k, the game will last longer than 30 minutes. If the advantage is greater than 6k, the game will last less than 30 minutes. As for extreme cases, an advantage of less than 2k usually results in game of over 35 minutes and an advantage of less than 1k might mean that the game will extend until full late-game. I took quite a few sample from Match History and plan on doing more later on, but this is how the dispersion looks like.

It’s not like the gold at 20 completely defines the outcome of the game, since a failed Baron or wrong positioning may turn the game around, especially late mid-game. It doesn’t matter if you can’t calculate the difference between the second derivatives of the gold earnings of the teams. What matters is that you should keep in mind that the team that feels like the hare from the story of “The Turtle and the Hare” might just let the other team become item-wise invincible. Get those dragons down.

Don’t troll,

The no-no Java

# Doom Bots… OF DOOM!!!

After those Solo queue matches I had to unwind a little and normal games didn’t do it for me… I made mistakes, got frusterated, got trolled etc… the usual.

Then finally came Doom Bots of DOOM! It was tough at first and truthfully I just wanted the icon (I’m a free-icon whore)… but after three games I got the hang of it an had a good enough team to beat them!!! (lvl 1 of course)

Laning is sorta hard, but I’ve found mid 1 v 1 to be easiest, I took Viktor, ‘cuz he has good wave clear and decided to take it slow.

Spoiler: I LOVED Doom bots of DOOM!! (oh and I won)

IMO Doom bots helps you learn how objective control is infinitely more important than kills. Normally people play against a tough team and as they start losing they also start to commit more and more mistakes. Usually they fail to recognize that that small bit of gold or that kill makes a difference when your up against some one better. And people FAIL to learn from those mistakes too.

ENTER DOOM BOTS OF DOOM

By making the bots so ridiculously over-powered (and until you realize that dying to them a couple of times) you decide that trying to assasinate them or duel them is in fact ridiculous. You discard that as a posibility from the get go focus on something you CAN do.

Lemme stop here to declare I’m talking about low ELO and low lvl players.

So without making it a choice made not by planning or strategizing too hard, people may be able to understand that the way you CAN win against a more “skilled” or “OP’d” team is by making smarter choices than too 1v1 them or duel them.

I have to say I feel my laning skills with Viktor have been polished somewhat. if you see the match history you can see I didn’t carry (obvsly) if anything they carried me again, but I could lane very solidly, and my combo stuck a lot, I could get kills, and be useful to the team.

Doom Bots of DOOM have been a learning experience for me. I sure hope to get to play with good players and my friends again. Maybe clear the next lvl too!

Zexycommander, signing off…

EDIT: League of Legends did a Q&A about Doom Bots of DOOM They also mention URF….!!

# Is It Possible To Predict a Game’s Outcome?

With the new match history that Riot now publishes I confirmed something I believed: games are predictable outcome-wise. As I published in an earlier post, gold seems to rule the match, but it seems to be much more than what it seems. As a simplification of different total gold graphs published in Match History beta, the gold earnings seems to like like this almost every time:

It’ll take me some more time and analysis to explain this linear tendency, but if this is a rule good enough, just by taking two reference total gold points with enough separation, an approximation of the winner side could be pretty accurate.

A partial explanation to why this works may be seen in the following graph:

To justify this, we simply write the equations for the total gold earnings:

$y = m_1x+475$

$y = m_2x+475$

With some research, I concluded that end-game builds end cost somewhere in the interval $[11000,17000]$. All we have to do is multiply by 5, to get the interval $[55000,85000]$. If we substitute in the equations, and being optimistic with the CS (which I’ll try to quantify someday), $m_1$ and $m_2$ are between $[1200,2700]$. Our time in minutes for end-game build, in average, of $34.59 \: min$. So depending on which team has the better $\frac{gold}{minute}$, the better the ratio compared to the other one:

By this point, it is up to you: if you are on the team with the apparently lower productivity and trust the statistics, then either push your income to get on par with the total gold by any means, or stall until gold is useless i.e. the final builds are ready and it is up to cooperation and individual skills.

While you are in-game (I believe) you can’t see how much gold the other team has gathered, but if you get someone to do this approximations for you, regardless of the 3 minute delay, you might get an idea of what the Math thinks about the game. If you are up to trying this out, then please do and later on come and comment here to say if this is complete trash or has something true in it (hopefully). I’ll try to bring some more algebraic and Calculus analysis to the game later on. Could Statistics be the true ruler behind this game?

The no-no Java

# The Shine of Victory

Arnor once told us that what really wins a game is gold. Evidently the in-game levels establish a stark difference between the players pre-6, but what would happen is a player were to get a 400 gold head start, considering that everyone begins at level 1?

For a start, an ADC would feel inclined to spend all that $875$ gold, right? What would one afford to pay with that?

Well, we have the Pickaxe as an option right at 875 gold. The attack damage of the ADC is in the interval $[49,51]$. Not he other hand, attack speeds are in the interval $[0.625,0.679]$.

If we are interested in DPS, damage per second, we get the following table:

# DPS

• Ashe: $32.3407\: dmg/s$
• Caitlyn: $31.25\: dmg/s$
• Corki: $32\: dmg/s$
• Draven: $33.95\: dmg/s$
• Ezreal: $31.3750 \: dmg/s$
• Graves: $33.812\: dmg/s$
• Jinx: $33.125\: dmg/s$
• Lucian: $31.262\: dmg/s$
• Twitch: $33.271\: dmg/s$
• Vayne: $35.0385\: dmg/s$

That’s the base AD that the champs get regardless of items. (Yes, Vayne, you hurt.) But if we add a Doran’s, assuming it is the standard, the DPS changes to:

 Champ DPS+Doran’s DPS+Pickaxe Ashe 36.9467 48.7907 Caitlyn 35.625 46.875 Corki 36.375 47.625 Draven 38.703 50.925 Ezreal 35.75 47 Graves 38.1875 49.4375 Jinx 37.5 48.75 Lucian 35.728 47.212 Twitch 38.024 50.246 Vayne 39.6445 51.4885

And after three minutes of laning, if no ADC has been forced to return to base, the total (flat AA) damage looks like this:

 Champ AA Doran’s AA Pickaxe Ashe 6650.406 8782.326 Caitlyn 6412.5 8437.5 Corki 6547.5 8572.5 Draven 6966.54 9166.5 Ezreal 6435 8460 Graves 6873.75 8898.75 Jinx 6750 8775 Lucian 6431.04 8498.16 Twitch 6844.32 9044.28 Vayne 7136.01 9267.93

If we do the ratios, the interval of increase would be $[29,32]$ percent. So that might not actually be worth it. Replacing a Doran’s for an early Pickaxe would be non-sense for a player who cannot freeze a lane, not lose a single CS point, and keep auto attack pressure on the enemy laner.

The interesting part would be that those 400 gold would be like getting two dragon’s for a single player. But dragon is usually done by three champs. That means that you get a double insta-dragon and your team’s total gold is boosted by $\frac{2}{3}$ of a dragon. What do you do with all that gold? I would just upright save those 400 gold and rush a Pickaxe after those good 3 minutes of farming.

Does gold win the game? It might, especially at higher levels, who would be able to squeeze out all of that auto attack damage. I’d say that gold wins the game if you know what the hell to do with it.

MfG

The no-no Java

# Caitlyn v. Vayne: My Bane

In theory, Caitlyn is meant to rule during “laning”. The range she has is insane; in fact she has the longest range among champs. Also, by the time you get acquainted with her play style, you realize that there is really no useless ability, as long as you are conscious of what you have at your disposal.  You basically feel omnipotent. Then you find out that your lane opponent is Vayne. Hell breaks lose. Why? Umm.

Considering Vayne’s squishiness and taking Cait’s range into account, I really begin to fall into a cycle where I clear the wave and push the lane, hoping to make the other Vayne lose farm. Then the jungler comes. I am forced to back up and the wave goes back to neutral push. After a couple of back-and-forth waves, I put pressure on Vayne. Those passive hits simply make you want to hit her. From this you might notice that my CS begins to fall behind hers.

Eventually mid game kicks in and Vayne with lifesteal becomes a huge problem. She might be squishy, but that lifesteal makes her seem immortal, especially against Caitlyn: The Bad Duelist. From here things begin to snowball and I hope that the other team’s Vayne is bad at positioning herself during team fights, while I abuse my range.

Vayne is supposed to be an easy champ as a matchup, but I do get carried away by my pre-6 advantage.

Thank you, Götze

The no-no Java