Greedy Goblin

Monday, January 9, 2017

League of Legends update

Remember the winrate vs team score graph from last week?

Things are going differently recently:
As you can see, I'm winning in every teamscore bracket. I climbed back to Silver 3 from Silver 5 where I fell during experimenting with a rudimentary form of exploit mode. What is exploit mode? At first I evaluate the team. The current teamscore formula is A/2-B/2-C+D. Of course I keep you in suspense about what A, B, C and D type of players are. If the teamscore is zero or bigger, I play in normal mode. I focus on my farm, dragons, gank when convenient. As you can see, the various types give different win chance:
If the teamscore is below zero, I play in exploit mode. I - again - keep you in suspense about what it is, but I guarantee that it's a completely retarded playstyle in a truly random game. It starts to make sense if you assume that the game is rigged against you in a special way, the way Riot rigs the games. Then you get this:
The interesting thing is the dominance of Type C players. Normal play is not dominated by Type D, despite the formula gives +1 to D and -1 to C, so the two graphs should be mirrors. Please also note that more Type B gets to positive teamscore teams, despite they give -0.5 to the score. Something is up here.

I think I can climb to gold next week to prove how well the exploit works.


Anonymous said...

Looking forward to seeing you in challenger.

Gevlon said...

I don't think I'll ever get to Challenger, because
- Warwick is weak (some 0.01% pick rate in Platinum and above) and I need him to exclude "skill" from the research
- Because I don't really care about LoL besides proving it's rigged. Why should I play competitively a game I know to be rigged.

Anonymous said...

If you manage to get to G2 or higher using this method I will definitely look into it.
Best of luck.

Anonymous said...

This means you are looking for a rig that gives you something like +100 MMR to another equally skilled player, but stops working at that point? (like queue dodging, just more malicious)

As in let a bot play 10000 games using method a - avrg mmr x, let a copy of the bot play 10000 games normally leading to an on arvg lower mmr.

I assume this is done solely for entertainment purposes otherwise i am really curious how you can do this type of study non-blind and how you exclude increase in skill (and the most important lol skills are non-champ specific - beeing at the right place at the right time is one of the hardest skills to master (which also explains why challenger level players can pick up a new champ and not perform like a silver player)).

Slawomir Chmielewski said...

Let me guess:
You give a rank to each teammate.
AB are the middle ones, C is the apparent feeder, D is the apparent good guy.
Retarded tactic may be following the retard to make sure he won't feed, set wards around him and such; or following the good guy to help him gear up better on their own feeder (if they have one).
Looks like C can dominate the game, as in DotA games the weakest link allows opponents to beef up and dominate with better equipment.

Gevlon said...

@Anon: yes. It doesn't "stop working", it just cannot give infinite +MMR. I mean if my skill allows me to be in Silver 3, I can probably get to Gold with this. If someone is already Gold, he can probably get to Platinum with this. I think the rigging fades out at Platinum as there won't be any more Type D players.

@Slawomir: far from the truth. Not even close. I don't blame you though, this is where I started and didn't work and when did data analysis did I find who are these groups. Please note that your classification doesn't imply rigging. You claim that good players win, bad players lose. Which is normal.

99smite said...

Will get some popcorn before the next GG post...
Will be kept in suspense until then...

maxim said...

Just letting you know that i really don't appreciate being manipulated in this way.
Makes me preemptively biased against anything you might have to say in the future.

The fact that you really seem to have presupposed your own conclusion of rigging is not helping, either.

Gevlon said...

@Maxim: until I'm finished, I can't publish final results. The post is named "update" for a reason.

LoL is a free-to-play game, so rigging is just as obvious as "perpeetum mobile isn't working". Finding out HOW is interesting.

Anonymous said...

The win/loss ratio from the graphs still looks close to 50%. How likely your results are random chance? Do you have a control account that just plays normally not matter what?

Gevlon said...

The first graph is from a test when I played normally, no matter what

Anonymous said...

There is naturally happening

a) knowledge increase of the game
b) skill increase, yes, even with ww you can use q,w,e,r differently, also you may ward differently
c) variance due to luck etc. network disconnections etc.
d) mood variance, some people are more affected by this than others, but I don't think anyone is completely immune
e) etc.

If you get from ~B1-S1 to G5 it proves absolutely nothing of your system or that the game would be rigged. You must realize it yourself.

As an example in my first lol season ww was my most played champion with ~70% playrate, the low point of the season was s5, I ended up in s1. No system, nothing special, I didn't pay anything to Riot, just natural learning of the game & knowledge increase.

Gevlon said...

Me getting anywhere proves nothing.

Statistics prove everything to intelligent people.

"This is how you can exploit the system" guide proves it to mediocre people. The thing is that I don't play "generally better". I evaluate the team on before game and depending on the outcome I play very differently in one game than another. The real point is that "exploit mode" playing is completely stupid and guaranteed defeat in a normal play scenario, works only if the team is set up for defeat in the special way Riot sets it up.

Xmas said...

Based on the weighting I'm going to guess D is...

Players who have lost some number of games in a row or have a generally negative tilt in their last n (10?) games.

A, B and C are going to be player roles that aren't the one you are playing and some criteria for their weighting. Maybe one of those is players playing some core role when their last N games were support roles.

You're card counting?

I'm going to take a wild guess that your change in play style is going to be some sort of aggressive ganking or just stealing the lane from the the 'C' players. I know that LoL is a bit rigid in the "proper" roles for different characters.