Forum
Collected data of a cursed account
|
fiero600 wrote
at 4:04 AM, Saturday December 20, 2008 EST
In the last 2 months, I noticed extremely bad luck while playing kdice. In november I shrugged it off, but as it got worse in December i started to alter my play style to win points. I started relying more and more heavily on early flags, early truces, and board manipulation to get what i could.
The better i got at this method, the more my points started to stabalize, and my statistics, (while not what they have been in the past) were decent. It was at this point that i noticed an even worse decline in my rolls, and began to track my rolls. Whereas i found it very tedious to track every roll, It was much easier to track the 8 stacks as they were clustered more and contained more data....So anyways, the longer i "cntrl c/v"'d my 8 stack rolls...the more intrigued i was. After tracking game after game, i ended up with data containing 100 dice rolls (800 dice!) Anyways...here's the summary of my dice [skrum if you want to review this, i have the data recorded (and some of it screen shotted when i got too lazy to cntl c)] 8v8's 58 rolls - 7 wins 51 losses 8v7's 9 rolls - 2 wins 7 losses 8v6's 15 rolls - 7 wins 8 losses (almost 50%) 8v5's 6 rolls- 1 win 5 losses (sigh) 8v4 3 rolls- 2 wins 1 loss 8v3 1 roll - 1 win 8v2 3 rolls- 3 wins 0 losses 8v1 5 rolls -5 wins 0 losses Worst 8v8 streak - 32 losses in a row (with Ghostface/ Kdicer X/wozoro watching) Shortest distance between 2 8v8 wins - 7 rolls Best suicide on me - Kdicer X (KDicer X's turn fiero600 defeated 8v8, 25 to 20, (2,4,6,5,2,2,2,2 to 1,6,2,1,2,4,1,3) fiero600 defeated 7v8, 27 to 22, (4,6,4,4,1,6,2 to 1,1,5,2,2,4,1,6) fiero600 defeated 6v8, 23 to 19, (2,6,5,3,4,3 to 6,2,1,1,2,1,4,2) fiero600 defeated 5v8, 23 to 20, (4,4,5,6,4 to 6,4,2,1,2,2,2,1) fiero600 defeated 4v8, 16 to 15, (6,6,3,1 to 1,1,1,2,6,1,1,2) +11 dice) BestT 8 stack roll = 35 Worst 8 stack roll= 11 (ive rolled an 8 though..just not in this set of 100) Overall Total of 800 dice = 1438 Overall Dice average of 800 dice = 1.8 (1.7975) [note : unless im wrong, average dice roll should be 3.5] Im not sure what this data means.... Is it possible for the average of 800 dice to be 1.8 when everyone is on an even scale with the "random luck generator" ? maybe? maybe not? I never used to believe that Ryan could/would "curse" an account... I remember at a time last month several people were complaining about "cursed accounts"... and i shrugged it off. I dont know what exactly Ryan (or who knows maybe a hacker ) did to my account... but i know 1 thing... my account is cursed. For anyone who doesn't believe this.. come watch me play a couple games... ill gladly build up (8 stacks) for you and let you watch as around 90% of them fail 8v8.. consistently. I know there's gonna be a lot of people who simply think this is funny... as many people obsessively hate me (kehoe, vermont, monte).. but i dont believe im the only account that's been cursed... and if Ryan/whoever can curse an account like this... i bet damn well they can make the average roll be 4.5 instead of 3.5... imagine how much that would affect an account. P.S. - Ryan, i dont expect an explanation or even a response because it's not your style to take credit for anything or ever admit to anything.. this was simply for community review and for those others who believe they have been cursed. If anyone wants screen shots / evidence of this data... feel free to message me and ill send it to you. Signed - Fiero600 |
|
|
fiero600 wrote
at 4:06 AM, Saturday December 20, 2008 EST Note : The 32 losses in a row extended 23 rolls past the original 100 but i still included the streak in the post.
|
|
|
fiero600 wrote
at 4:20 AM, Saturday December 20, 2008 EST Eek= ok 2 corrections after someone looked over my data sheet
Total dice should actually be 1938 not 1438 So average dice roll is actually 2.4225 so.... instead of 1.8... its actually 2.4225 Still considerable.. and all other stats check out |
|
moondust wrote
at 5:08 AM, Saturday December 20, 2008 EST Very interesting stuff you posted here.
First of all: I really do hope that people (and Ryan or Skrum in particular) will give you some feedback here and I hope that people don't start again to talk about July 2008 [because THAT gets really tedious]. Anyway... I had a look at your stats and wow... it looks like something is really wrong there. Personally, I am surprised that you haven't quit yet, because with luck that low, normal game play is not possible anymore as it gets frustrating after some time. As I already said I hope you will get some constructive replies from people how have something to say, because I am also interested if this is still within the frame of "Statistical Fluctuation" as the call it. |
|
|
moondust wrote
at 5:09 AM, Saturday December 20, 2008 EST edit: who*
|
|
Vermont wrote
at 7:54 AM, Saturday December 20, 2008 EST "Obsessively hate" anyone over a game of kdice? Get over yourself. People can disagree without it instantly being "hate." Making everything black and white does no good to further any sort of analysis.
(That being said, Ryan probably just saw what you've been doing late at night with rnd and leekstep.) You're certainly not the only one to think their account has been de-lucked, though. It would be quite interesting if Ryan took the time to add code to do this, but given my observations of a few different accounts it doesn't seem like there is a realistic alternative explanation. Since it should be fairly easy for you to compute as you have the data - what was the average roll for your opponents? I'd be curious to see if that falls near the expected mean. |
|
nuflis wrote
at 11:19 AM, Saturday December 20, 2008 EST I remember last July when..... oh wait.
Man, that's bad luck. |
|
skrumgaer wrote
at 3:12 PM, Saturday December 20, 2008 EST fiero,
The number of successful rolls for a given matchup should follow the binomial distribution. The expected number of wins for n rolls is np and the variance is np(1-p) where p is the probability of winning. For an 8v8, the value of p is 0.471. For 58 rolls your expected number of wins would be 27.3 and the variance would be 14.5. For your 7 wins, the difference between your 7 and the expected 27.3 divided by the standard deviation, the square root of 14.5, would be (7.0 - 27.3)/3.81 or z = -5.33. This is a highly significant result. The hypothesis that this would occur by chance is rejected. Now what would be the cause? It could be bias in the rolls, or bias in the sampling. If the sample of 100 rolls is itself randomly selected (that is, you say "I will measure the next 100 rolls" instead of going back and selecting 100 consecutive rolls from your roll history), you have eliminated sampling bias. The suicide stats are a biased sample since they are selected ex post facto. |
|
|
skrumgaer wrote
at 3:42 PM, Saturday December 20, 2008 EST I have looked at the luck stats of the current top 100 and there are two negative outliers: jcp4p and fiero600, that lie in the 46% range while most eveyone else is 48-50%. There is one positive outlier, Bahaduras, but he has played only one game. What do jcp4p and fiero600 have in common that they do not have in common with the other top 98? They both have high percentages of first and second place, but so do some of the others.
One think I do not know about the luck stat is whether the monthly stat is the average of the game stat or is the roll-weighted average of the game stat. One thing to consider about the behavior of first- and second-place players is do they tend to sit and wait while others fight it out for third, fourth, etc. If they tend to sit, they will have fewer rolls per game and their luck stats may not average out as quickly. |
|
dasfury wrote
at 4:37 PM, Saturday December 20, 2008 EST Did you track your defends too? Otherwise, this is only half the story.
|
|
|
fiero600 wrote
at 5:17 PM, Saturday December 20, 2008 EST skrum, thank you for analyzing and looking at the data... it doesn't surprise me at all that im one of the 2 negative outliers... and if JPC has the same curse...yeah sorry bro it does suck.
as for the defends... they were a little bit better than the attacks.. maybe 20% defend on 8v8? i didn't save the defending data but it wasnt very good either... im just curious as to why ryan did this... were me and JPC winning too much and needed to be given a handicap or what? |