you already have gang or want to make your own? i want to train my own gang for whole year to next GST, and already have 5 loyal friends, gonna start base next monthQuote:
Originally Posted by Don Alan
Printable View
you already have gang or want to make your own? i want to train my own gang for whole year to next GST, and already have 5 loyal friends, gonna start base next monthQuote:
Originally Posted by Don Alan
Actually hate every villian on era lol
Yeah, I seen that.Quote:
Originally Posted by SashaGryg
To clear the air, let me break down how the Gang Spar ranking was determined:
Wins - Losses = Score.
The scores have gone through several filters to weed out possible boosted wins.
First off, when a gang beats the same gang consecutively, it will only count the first three rounds.
Meaning, if they won 100 rounds back to back, against the same gang, only the first three counted.
Any win from round 4 to round 100 is eliminated; because, well, find a new opponent.
After those are weeded out, I go through the data to identify outliers.
An outlier is data that skews results, and in this case, potentially points towards boosting.
Outliers are easily identifiable when I break down the LOSSES a gang incurred by percentage.
It asks the question, “of the total losses that GangA(the loser) took, what percentage came from GangB(the winner)?”
To do this, I go through each and every gang’s winning percentage, in this fashion:
http://puu.sh/gT5ya/16372a614c.png
In that example, I see a possible outlier. “C R I P”.
(C R I P) sticks out to me, because, as far as the rankings go, they’re ranked 445 out of 453 total Gangs participated.
That ranking tells me that they've not really contributed to the competition in a positive way.
I've been watching gang spar and going through data all week.
(C R I P) hasn't appeared on any other meritable gang's top wins list, so it gets me thinking: "What the hell are these guys doing here?"
I investigate by breaking down the Losses of (C R I P), by percentage:
http://puu.sh/gT60Y/41d3a9e9d0.png
You’ll see that (C R I P) lost a total of 52 times. 44 of those losses went directly to Real Savages.
All 44 wins were recorded within a time frame of 1 hour.
http://puu.sh/gT7bZ/4c5d001ee4.png
This is one of two things:
1) Boosting, or
2) There was no other competition for Real Savages.
Either way, I cannot consider that data as relevant, so I eliminate that data from the equation.
Those fights weren’t counted at all.
On the matter of Villains, They actually did have the most outliers.
Villains GST had a total of 257 wins that were outliers.
All of those fights were completely eliminated from the data- they received absolutely no points for them.
I don’t recognize data that was potentially fraudulent.
All outliers for every gang were eliminated from the results.
However, even with those outliers eliminated, Villains GST still beat Cons by 70 points.
http://puu.sh/gT7y3/b20fba0991.png
No need for this thread, the point has already been made by layerd.
Who is Cons?
one of the best spar gangs.Quote:
Originally Posted by Cinos
Meph did good and long work, i think. so many filters and so many different results... i am from phone, so i did not see the pictures BUT he found out many outliers. cant wait for GST results XD
Heheheh! Glad you happy! :)Quote:
Originally Posted by SashaGryg
Then you should start worrying. If a gang got so many matches without winning, and all the losses come from the same gang is most probably boosting. Even if wins did not counted, you should check them outQuote:
Originally Posted by Meph
- - - Updated - - -
EDIT: You told us if a gang get more than the 30% of wins on the same gang, it counts at boosting.
even that those 257 point didn't count Villains beater Cons by 70 points [emoji50][emoji13]
Yes but if they were boosting (looking at mephs results it looks like they were) then they should be disqualified no matter if the legit wins got them through. Anyway im not takin sides but that seems suspicious to me.