Browse
Category
DiscussionsUncategorized

What could a statistical analysis of someones entries tell us ?

Closed
Bilateralrope

Take a list of all the closed giveaways that someone has entered. Do some statistical analysis on them. What pieces of interesting information could be extracted ?

Would displaying that information to the person in question be useful ?

The obvious one would be to calculate the average win chance of the giveaways someone entered, so it can then output "You should expect to win a game every x entries". Which would cut down on some of the "I entered y giveaways, why haven't I won" whining. But not all, some people will just delay their whining till they hit x entries because they don't understand averages.

1 decade ago*

Comment has been collapsed.

14 Comments
aapinator

"You should expect to win a game every x entries" is a very wrong statement to make. Keep in mind that every giveaway has different variables to take into account; such as duration, desirability of the game in question, day of the week on which it's been given away and most obviously the amount of entries. And even if these variables were the same with every giveaway, then still the 'you can expect' part would be wrong.

A "You have a 0.XXX% chance of winning" statistic for every giveaway, that would update as people enter a giveaway could tell people "something" although I doubt this statistic would be encouraging or fun to know for anyone. If you desperately want to know the chance of winning per giveaway in percentage, you can simply divide 100 by the amount of entries.

Edit: the stats page already displays the odds of winning per entry, per day.

1 decade ago
Permalink

Comment has been collapsed.
Bilateralrope

You should expect to win a game every x entries" is a very wrong statement to make.
Not really. The chance of winning is a fraction of 1/x, meaning that a person will win, on average, every x entries.

Keep in mind that every giveaway has different variables to take into account; such as duration, desirability of the game in question, day of the week on which it's been given away and most obviously the amount of entries.
After a giveaway has closed, all of those variables are reflected in the number of people that enter the giveaway.

Edit: the stats page already displays the odds of winning per entry, per day.
And those statistics are only useful for a single person if that person enters every single giveaway here. Since nobody can enter every giveaway, their personal statistics are going to be different from the overall statistics. So that page isn't so useful.

1 decade ago
Permalink

Comment has been collapsed.
Jaina

You should expect to win a game every x entries" is a very wrong statement to make.
Not really. The chance of winning is a fraction of 1/x, meaning that a person will win, on average, every x entries.

"A person will win, on average, every x entries" and "A person should expect to win every x entries" are VERY different sentences.

If a large number of people enter x contests each, where the chance of winning each is 1/x, then it's true that the average number of wins per person will be x. Keyword being 'average'. That's because some people will win more, while others will win less.

If one person enters x*y contests, where the chance of winning each is 1/x, then it's tempting to say that they will probably win y of those. That wording is very misleading, since the odds of them winning that number exactly are actually pretty low. You're better off saying that they have good odds of winning y of those draws, plus or minus a certain amount (which is where standard deviations come into things, for those who remember that term).

Now, if this means that someone has entered 50,000 draws, with a 1/100 chance of winning each, then they're probably not going to care much whether they're below the 'expected' 500 wins or above it, so long as it's not by much. Most people don't get nearly so upset when you're dealing with large whole numbers, since even unlucky results usually still mean they've at least won something.

If someone's only entered 100 giveaways, though, each with a 1/100 chance of winning, then the thing to keep in mind is that they're still not guaranteed to win 1 of them. They're only statistically likely to win some number of them very near to 1. That could mean they've won 2, or possibly even more. Or it could mean 0.

In fact, if you do the math in the above example, the odds are actually OVER 36% that they won't have won anything. That's over a third. Scaling the value of x used doesn't change this value very much.

If you tell a group of people that they can "expect to win a game every x entries", and over a third of them don't, they'll usually complain and claim they were misled. Hence it is considered a "very wrong statement".

1 decade ago
Permalink

Comment has been collapsed.
satoru

http://www.steamgifts.com/stats

We already have this and no one complaining about not winning ever looks at it, nor do they have any remote concept of how statistics and probability works. Thus any data given will not stave off the inevitable threads of "Why I did not win" because such people do not care to even inform themselves.

1 decade ago
Permalink

Comment has been collapsed.

Deleted

This comment was deleted 4 years ago.

1 decade ago
Permalink

Comment has been collapsed.
Martin

you can use me as a testing bunny if you want , 323 entries , 2 gifsts won . my secret weapon is being chill and believing in the gifting god up in the sky.

1 decade ago
Permalink

Comment has been collapsed.
DanKnee

Eleven, The answer is eleven.

1 decade ago
Permalink

Comment has been collapsed.
anonynamja

you could do some kinda multivariate probit

1 decade ago
Permalink

Comment has been collapsed.
neocow

No, you could never say "expect to win every X entries" with statistic, unless you were the only entry.

1 decade ago
Permalink

Comment has been collapsed.
xarabas
(Junior Moderator)

I hereby apply as a test subject

1 decade ago
Permalink

Comment has been collapsed.
CITIZ3N

i dont understand why people would whine about not winning ....its not cost them anything to enter...i can not think of another auction that gives you free credit to bid on the free games for FREE.... if i bid on 10,000 games and only won 1 time its still a free bonus and i would be happy..... The only effort involved is logging into here each day and using your FREE credits and start clicking on the games u hope to win.....

But lets not forget none of this would be possible if it was not for the generous people who freely give thier unwanted/unused games away for FREE.....

the only true statistic is that we all have the same chance in the auction as the next person has to win a game...

1 decade ago
Permalink

Comment has been collapsed.
helraiser31

I completely agree with this. You are getting this for free!!!

1 decade ago
Permalink

Comment has been collapsed.
Fiddlestyx

I have only know about this website for ~four days. After 41 entries I won a game. I may enter another 10,000 contests and not get anything.

Random is random, and nobody should complain about free things.

1 decade ago
Permalink

Comment has been collapsed.
satoru

You should read the constant bitching and whining on the Steam forums about the Winter contest and how 'free' somehow entitles them to be whining 10 year olds on Christmas.

1 decade ago
Permalink

Comment has been collapsed.

Closed 1 decade ago by Bilateralrope.