I would like to know this too, personally I have had those little streaks too where I would win 2-4 games in 1-2 days.
Comment has been collapsed.
Add a poll please, I also feel it is non-random/streaky or at least not pure arbitrary.
Comment has been collapsed.
I would, but when I try I don't think it'll let me go back and add a poll. Am I missing something?
Comment has been collapsed.
Streaks happen in random occurrences. I might win a few games one week, then none the next week, then one the week after.
Definitely haven't noticed any rhyme or reason to it.
Consider that sometimes you might enter a lot of higher level GAs all at once, or a train, etc, and other times you may not enter as much. Obviously the days you enter more, you're going to have more chance to win. Some days I might enter 5 GAs all day, other days I might enter 50+ if there's something I want.
Comment has been collapsed.
+1 for the point that sometimes you may be entering more giveaways or giveaways which end at a certain time, therefore higher chances of winning them at the time. As for the rest, it's pretty much all pure coincidence.
Try visiting this website. The chances of getting the same number twice are somewhat the chances of you winning twice.
Comment has been collapsed.
I was just reading this LOL. The Dilbert cartoon is right on the money.
https://www.random.org/analysis/
EDIT:
What Dilbert is told is correct: It is impossible to prove definitively whether a given sequence of numbers (and the generator that produced it) is random. It could happen that the creature in the comic strip has been generating perfectly random numbers for many years and that Dilbert simply happens to walk in at the moment when there's six nines in a row. It's not very likely, but if the creature sits there for long enough (and Dilbert visits enough times), then it will eventually happen.
Comment has been collapsed.
i agree its streaky w,e that means lol.... 2 games recently in the same 10 minute period, and there are many times ive won 2 games from multiple people where i shouldnt have cuz of the odd's
Comment has been collapsed.
Thanks - that's kindof what I'm wondering, because of the two I mentioned ending at almost the same time it makes me wonder if it was a single selection process for that period of time (edit: like maybe a random seed that ranks user id's into a table of priority for a given period of time) I've also noted something similar in other group giveaways, but with smaller #s the odds aren't as astronomical. thanks for the info about 3 at same time.
Comment has been collapsed.
Pretty sure you can only win a GA with you are friends with cg or wearing a tin hat.
BTW
The odds of someone winning the two GA with 394 & 438 entires is alot higher than 0.0000058. It is that times the number of people that entered both GAs. Each entrant has a chance to win both GAs. So could be as high as 1/1000 chance. Given the number of GAs on the site it probably happens most days.
Comment has been collapsed.
Someone described it here, but (1/394)*(1/438) are the odds for that particular person and not anyone who happened to enter both giveaways. It's a similar conundrum to the birthday paradox in that you don't need to get 365 people together to have any two people share a birthday.
For a really clear example:
Giveaway 1 and giveaway 2 happen to have the same 500 people entered in each.
Someone will win giveaway 1. The odds of that same person winning giveaway 2 are 1/500.
The odds of anyone winning both is 1/500 (500 people x (1/500) x (1/500)).
Comment has been collapsed.
Yeah. The birthday paradox is infamous for screwing with people's heads.
Let's say 1-6 on the dice represent a person and each roll represents a giveaway. A person winning both would be rolling any pair (1-1, 2-2, 3-3, 4-4, 5-5, 6-6). The odds of person 6 coming up twice would be 1/6*1/6 as you said, but the odds of any of the 6 numbers/people coming up twice in a row would be 6/36 as there are 36 possible outcomes and 6 of those are pairs.
When you say 499 unique people in giveaway 2 that changes a lot. To put it in your dice analogy:
Giveaway 1 has 500 people including you.
Giveaway 2 has 500 people but only you entered giveaway 1.
This means there are 500+499 unique people/numbers.
Let's assign you the number "500"
Giveaway 1 is a die with the numbers 1-500
Giveaway 2 is a die with the numbers 500-999
The only possible pair is rolling 500/500.
In my scenario the same 500 people entered giveaway 1 and 2.
This would be having two dice with the same numbers: 1-500,1-500.
There are 500 possible pairs you could roll with those two dice.
Comment has been collapsed.
I think you're starting to get it! 200 uniques would mean 438-200=238 duplicates, so you could make 238 different pairs.
The exact probability of "that guy" is still (1/394)x(1/438), but the odds of someone winning both depend on how many entered both, so 238 duplicates means 238 people each have a probability of (1/394)x(1/438), for a combined probability of 238x(1/394)x(1/438) :)
Comment has been collapsed.
Glad you got it!
That's the same confusion people have with the birthday paradox! Birthdays are 1/365 (in theory, not in practice), but you only need to get 70 people in a room for a 99.9% chance someone meets another person who shares their birthday :)
The birthday paradox also applies to random number generators.
Comment has been collapsed.
...times the number of pairs of giveaways that end within one minute of each other since the creation of the site makes it basically a 100% chance of this happening at some point.
Comment has been collapsed.
I've noticed that I tend to win more right after I do a giveaway myself. Can't say if there's any correlation but it's happened every time I've done a giveaway myself so far.
Comment has been collapsed.
dont think so, but at least: making giveaways can leed sometimes to get whitelisted and increases your level as well. both will help you to win giveaways^^
Comment has been collapsed.
"I just had the same person win two giveaways with 394 and 438 entries that ended 1 minute apart. The probability of that happening is 0.0000058 assuming randomness (or pseudo randomness w/ an algo and seed)."
That's not true.
The probability that that particular person won both of them is low, but the probability that somebody won both is much higher.
If all 394 of the people entering the first giveaway also entered the second, then the probability assuming randomness that the person who won the first giveaway also won the second is 1/438 = 0.002. Even if only half of them also entered the second giveaway, the probability that the same person won both would be 1/2(1/438) + 1/2(0) = 1/876 = 0.001.
Comment has been collapsed.
thanks for the detailed correction w/ math. appreciated.
Comment has been collapsed.
In addition, you'd have to know how many total giveaways the creator made and how many of those giveaways the same people entered.
As it stands with your equation, every subsequent GA entered by the same people (which can happen a lot when someone posts all the games in a new bundle, or a bunch of whitelist GAs at once, etc.) increases the odds of the same person winning multiple giveaways from the same creator. If that creator made 100 GAs all at once (which can sometimes happen in a train), there's a much, much better chance of the same person winning 2 or more games from that train.
Comment has been collapsed.
Both! :D
Winners are random. Randomness is often streaky (more often then most people would think). Often enough, a probability professor asks his students to go home and throw a coin 1000 times, and return the results to him. You can EASILY pick the people that faked the results instead of throwing the coin (or using a computer) because people do not fake the long streaks of Heads (or Tails) that SHOULD happen in such an experiment.
Besides, when you calculate probabilities, you have to be aware of the question you are asking:
-- The probability of THIS particular guy winning THOSE particular 2 giveaways is that number you mentioned, 1 in (394)(438), that is, 1 in 172572.
-- The probability of SOME guy winning THOSE particular 2 giveaways is a lot more; say, if 100 people were in both giveaways, it would be about 100/394/438, that is, about 1 in 1726.
-- Now, if you do about 100 giveaways with similar numbers, the probability of having SOME guy winning SOME CONSECUTIVE 2 giveaways will be about 1 in 18 or so (try 1-(1725/1726)^99). Still somewhat unlikely, but not enough to doubt the randomness of the site.
So again: streaks are often due to randomness, a lot more than most people think.
P.S.: In 20 Dec 1986, the German Lotto (winning chance = about 1 in 14 million) picked the numbers 15-25-27-30-42-48. Guess what numbers were picked in 21 Jun 1995? :)
P.P.S.: All of that said -- yeah, what an incredibly lucky person! :)
Comment has been collapsed.
"There is no fate but what we make ourselves"... so, someone is cheating on SG?! :P
Comment has been collapsed.
"The probability of that happening is 0.0000058 assuming randomness ("
There's more than one giveaway! Figure the odds of that happening once for all the giveaways in the last month.
Comment has been collapsed.
to ensure that you are maximising your chances to win please make sure to provide offerings regularly to the RNG Gods. they can be fickle and petty, but provide great bounty to those who please them.
Comment has been collapsed.
i think it's 100% random, but i'll keep sending cg money each week just in case. :3
Comment has been collapsed.
i once won the same game from the same user for 2 diff GA he made. so what? doesn't mean anything, that was truly lucky
Comment has been collapsed.
Yeah, of my recent b-day giveaways, one user won two of them (~250 entries each).
This "phenomenon" shows itself often in the Simultaneous Giveaways group events, where hundreds of giveaways end at the exact same time. There will be a few users that end up winning a large number, and some that end up winning just a couple.
Comment has been collapsed.
I don't know what algorithm does SG uses to simulate random function, but I'm guessing it needs some rework.
Judging from your description, I would say that Date/Time plays some important part in that function as does it could be also linked to issued raffle tickets for each GA.
As programmer myself, I'm quite interested to see the inner workings of running algorithms.
Comment has been collapsed.
I would say streaky, but only because of bursts of giveaways. While indiegala always delivers, bundlestars often doesn't have a really good bundle for weeks, Humblebundle's main bundle barely have Steam codes right now - without Killer 6, there wouldn't really be good giveaways in high numbers. But considering how good (and unbundled previously) BS's bunde's games are, people entering much more. More giveaways --> more entries --> more chance to win more than one game.
Comment has been collapsed.
I voted streaky from anecdotal experience, but as other have mentioned already that may just be psychological bias at work. The only real way to properly assess this would be to take a wide statistical sampling of actual results and compare them to expected levels of variance.
I'm interested in where this poll ends up though, as if the anecdotal evidence is high, it may justify putting some time/resources into digging into the statistical data.
Comment has been collapsed.
It is random. It doesn't feel like one because human brain is a pattern detection machine. We do notice coincidences and make patterns out of them, at the same time ignoring everything else. Ie: how many times two different persons won giveaways ending at the same time? )
Comment has been collapsed.
I wonder does it really matter when you entered the giveaway (in its very beginning, in the middle or in the end)?
Comment has been collapsed.
A perfect example for the streak theory: http://www.steamgifts.com/discussion/y8XJz/thank-you-sg
I'm totally convinved it is based on streaks. I always win in clusters and then nothing for a while.
Comment has been collapsed.
The probability is 1/394+1/438 not 1/394*1/438. Not so impossible after all
Anyway I have win streaks too, after one I stop winning for a while XD
Comment has been collapsed.
that happend to me too......
a lot of time whitout win nothing then pummmmm 2-3 consecutives wins in a few days.
it's not 100% random, probably you are in a "qeue" or something similar when entering giveaways of same games etc..... that's how i think.
Comment has been collapsed.
Hi all, I just had a wow moment that made me ask this question.
I just had the same person win two giveaways with 394 and 438 entries that ended 1 minute apart. The probability of that happening is 0.0000058 assuming randomness (or pseudo randomness w/ an algo and seed).
I've also noticed that when I win, I also often win in bunches. That can happen with randomness for sure, but it feels a bit clustery sometimes.
I'm curious, do any of you get the feeling that maybe winning isn't random on a per giveaway basis, but maybe some days you're streaky? Just curious.
And I'm not insinuating anything improper is going on, I'm just trying to understand if something about the selection algorithm might make streakiness more probable? Is it a simple rand(n), where n is # of entrants, or is it perhaps somehow tied to particular user id's on certain days that makes winning more probable? If it's been explained somewhere that'd be cool to read. thanks
EDIT: Link to Poll as provided by rred (thanks)
http://strawpoll.me/6225603
Comment has been collapsed.