Bbranning2004 (With Statistics!) - FARMER! [warned] BG

[Tenebrus]

Accused:

bbranning2004



The accused are suspected of:

Other: Farming



Game number(s):

Statistical Analysis of all his 3 player Halloween Hollows games


Comments:

I realise that I only just posted up a report about this, and you guys were good enough to (i) explain the kind of games that NRs can play, and (ii) note it. Sometimes though, I just get curious and do a bit of data mining because I realise that the feeling that someone is cheating is usually just randomness and it's a good way to tell yourself you're just being crazy. But every now and again you come up with a pattern that demonstrates a statistical certainty of cheating, which I think Bbranning2004's games do.

I originally reported him because he seemed to have a lot of NRs in his games. Fair enough people said, NRs join a lot of three player games because they can't join games with more than four players. And also, he can't invite people because he set the games up. All good points, but there has to be an explanation for this:

Number of Bbranning2004 games: 219
Number including a NR: 121
Percentage of NR games: 55%

Number of all 3 player HH games (24 hr/public): 621
Number including a NR: 181
Percentage of NR games: 29%

That doesn't appear to be that much of a discrepancy, but the right comparison is with the 3 player HH games when you strip Bbranning2004 out, then you get:

Number of all 3 player HH games: 402
Number including a NR: 60
Percentage of NR games: 15%


When you are talking about a couple hundred games over an extended period, those kinds of anomalies are a statistical impossibility. To be extra sure, I looked at a random sample of 3 player games generally. Looking at the last 3 complete pages of games on game finder for the Africa map - 16% of games include an NR, for Alexander's Empire it dropped to 1%, so there's map variation. I tried a random sample of all 3 player (24 hr/public) games, and went back quite a way and the consistent percentage is 10%.

It's also worth noting that his games don't start out this way, they change. If you slice it so that you disregard his early games here, the percentage rises as high as 75%. If you include games with players who played less than ten games but just happened to not show up as a ? any more on Game Finder, it's just about every game.

I'm not an IT guy, I don't know how this could be done, but generally data doesn't lie. There is some cheating going on here, fo sho. That's even without getting into a more detailed analysis of how those NRs appear to be helping Bbranning2004 win, there are some particularly damning games of the NR not showing up (as they always deadbeat in his games), until he needs them to or of them always hitting the third player.

Thanks for your time!

[D00MandD3A7H]

Cmon stop this, it is called farming and ranching. No one enjoys 3 player games, only a new recruit would join this.

[Tenebrus]

That's the whole point of my post, dude. To work out what % of NRs you can realistically expect to join an open 3 player. It's about 10-15%. With Bbranning it's 55%-75%. That's a big enough statistical anomaly that it needs an explanation. And the only explanation is some form of action on Bbranning's part which makes it (at best) farming/ranching.

[jltile1]

Accused:

bbranning2004



The accused are suspected of:

Other: Farming



Game number(s):

Statistical Analysis of all his 3 player Halloween Hollows games


Comments:

I realise that I only just posted up a report about this, and you guys were good enough to (i) explain the kind of games that NRs can play, and (ii) note it. Sometimes though, I just get curious and do a bit of data mining because I realise that the feeling that someone is cheating is usually just randomness and it's a good way to tell yourself you're just being crazy. But every now and again you come up with a pattern that demonstrates a statistical certainty of cheating, which I think Bbranning2004's games do.

I originally reported him because he seemed to have a lot of NRs in his games. Fair enough people said, NRs join a lot of three player games because they can't join games with more than four players. And also, he can't invite people because he set the games up. All good points, but there has to be an explanation for this:

Number of Bbranning2004 games: 219
Number including a NR: 121
Percentage of NR games: 55%

Number of all 3 player HH games (24 hr/public): 621
Number including a NR: 181
Percentage of NR games: 29%

That doesn't appear to be that much of a discrepancy, but the right comparison is with the 3 player HH games when you strip Bbranning2004 out, then you get:

Number of all 3 player HH games: 402
Number including a NR: 60
Percentage of NR games: 15%


When you are talking about a couple hundred games over an extended period, those kinds of anomalies are a statistical impossibility. To be extra sure, I looked at a random sample of 3 player games generally. Looking at the last 3 complete pages of games on game finder for the Africa map - 16% of games include an NR, for Alexander's Empire it dropped to 1%, so there's map variation. I tried a random sample of all 3 player (24 hr/public) games, and went back quite a way and the consistent percentage is 10%.

It's also worth noting that his games don't start out this way, they change. If you slice it so that you disregard his early games here, the percentage rises as high as 75%. If you include games with players who played less than ten games but just happened to not show up as a ? any more on Game Finder, it's just about every game.

I'm not an IT guy, I don't know how this could be done, but generally data doesn't lie. There is some cheating going on here, fo sho. That's even without getting into a more detailed analysis of how those NRs appear to be helping Bbranning2004 win, there are some particularly damning games of the NR not showing up (as they always deadbeat in his games), until he needs them to or of them always hitting the third player.

Thanks for your time!

Very good stats. However it will be the same shit and they do nothing as always. It pretty obvious to me, but the people in charge to busy cleaning there finger nails to hit the keyboard. It kinda funny if you read his game logs too. You see how he only attacks the ?.He will attack the other player to get his bonus in most games then it's all on the player who is actually playing? That's just a accident.


Looking at the logs something is going on.

[TheForgivenOne]

That's the whole point of my post, dude. To work out what % of NRs you can realistically expect to join an open 3 player. It's about 10-15%. With Bbranning it's 55%-75%. That's a big enough statistical anomaly that it needs an explanation. And the only explanation is some form of action on Bbranning's part which makes it (at best) farming/ranching.

Tenebrus, you also have to take into account that ranks aren't archived, so just because you don't see NR's in the older games without Bbranning, doesn't mean they weren't at the time. So I'm guessing your percentage is a lot lower then it actually was at the time of the games.

[WILLIAMS5232]

That's the whole point of my post, dude. To work out what % of NRs you can realistically expect to join an open 3 player. It's about 10-15%. With Bbranning it's 55%-75%. That's a big enough statistical anomaly that it needs an explanation. And the only explanation is some form of action on Bbranning's part which makes it (at best) farming/ranching.

Tenebrus, you also have to take into account that ranks aren't archived, so just because you don't see NR's in the older games without Bbranning, doesn't mean they weren't at the time. So I'm guessing your percentage is a lot lower then it actually was at the time of the games.

well that also works into Bbranning's games. some of the players that he was playing earlier may no longer be new recruits now. but they were when he played them.

[Tenebrus]

That's exactly right, Williams5232. These games are not a brief snapshot of a recent week, they are a statistical analysis going back a whole year (and the NR % stays consistent). My cross-references are also done on a similar timescale (showing the NR % for other maps/all maps/HH), so TFO's explanation doesn't explain the discrepancy. Happy to reconsider if anyone has an explanation that does do so, but I don't believe there is one other than some action on Bbranning2004's part.

[BGtheBrain]

*****

[Extreme Ways]

bbranning2004 has been WARNED for farming new recruits.

I do appreciate the amount of work that Tenebrus put in to his report, and the information he provided was very helpful in the case.

Yes, I am fully aware that this contradicts my previous ruling on bbranning2004 where I cleared him of farming back in February or March.