Journalism? Statistics? Nonsense...
- Thread starter The Halk
- Start date
I can see what you're saying, in 240 years time it is almost certain that we will see this happen again - just with another MP. Definitely see what you're saying, and yes, you're right that the journalists probably are actually under the impression and are trying to put across that the probability of any MP being picked twice really is 58k to one, however they have worded it just enough to actually make sense. Only just though.
Indeed, the statistic isn't wrong for the article, the statistic for the group of pairs may be more representative but it's ultimately only presentational
It's a conditional probability, what is the chance he wins this year, given he won last year.
Where P denotes the probability,
P(he wins this year | he won last year) = P(he wins two years in a row) / P(he won last year) = (1/240 * 1/240) / (1/240) = 1/240.
So simply 1/240, which is obvious in itself anyway.
Someone would only be receiving 58k for a pound if they bet this would happen before he won it last year theoretically. Maybe that's what the article is 'getting at'.
Where P denotes the probability,
P(he wins this year | he won last year) = P(he wins two years in a row) / P(he won last year) = (1/240 * 1/240) / (1/240) = 1/240.
So simply 1/240, which is obvious in itself anyway.
Someone would only be receiving 58k for a pound if they bet this would happen before he won it last year theoretically. Maybe that's what the article is 'getting at'.
Sorry for the upcoming pedantry (and I'll probably get something wrong because its the middle of the night!).
Something like this, with a known average rate and independent events, will occur according to a Poisson distribution, which means there's a really large variance in how often it arises, meaning that its far from certain that we'll see another case within the next 240 years. Conversely, we could see several.
Within the next 240 years, the 95% confidence intervals for the number of times it is likely to have happened is 0.025 - 5.57. So it could be a lot longer, or a lot less time before it happens again. Which I suspect you knew, but for some reason I felt the need to pick up on your wording
Last edited:
Well that's good, as that's exactly what happened.
Wait I see what you're getting at, I had to re-calibrate my pedantometer to maximum to work it out though.
The probability of an individual MP getting picked twice in a row is 58k/1. The probability of that event occurring is 240/1. Statlol -mind your P's and C's!
Of all the examples you could have picked of journalistic mis-use of stats this is probably one of the least worst there is.
Last edited:
Soldato
- Joined
- 2 Mar 2004
- Posts
- 11,919
- Location
- SE England
We don't write da news, we create da news 
73% of statistics are made-up anyway.
73% of statistics are made-up anyway.
![[TW]Fox](/data/man-of-honor/2009.png){kind=avatar}
I agree.
Soldato
- Joined
- 16 Nov 2010
- Posts
- 16,513
- Location
- Swimming in a lake
We don't write da news, we create da news
73% of statistics are made-up anyway.
It is 90% fool
There are lies, damn lies, and statistics.
kd
Why dont you apply this logic to the MP ballot in your OP?
There are ~58k possible outcomes, and 1 of those is John MacDonell being selected twice.
1 in 58k?
edit: are you confusing the chances of any MP being selected twice with the chances of John MacDonell being selected twice?
Last edited:
Soldato
- Joined
- 8 Mar 2007
- Posts
- 10,938
nvm,
Last edited:
I don't believe he's confusing them, just that he thinks the odds of any MP winning would be a more representative probability to use.
Soldato
- Joined
- 8 Mar 2007
- Posts
- 10,938
You've not applied the same logic though. He is saying that out of the 36 possible out comes with the dice, 6 of them are matching numbers (so two 1's OR two 2's OR two 3's and so on).
Why?
The article is about John MacDonell. The odds for him of being selected twice are exactly the odds they use (as they are for each individual from their own point of view). Seems pretty straight forward to me.
I'm a proper dunceton when it comes to mathematics - always have been and always will be
Please could someone explain the differences between chance, probability and odds?
odds are for betting
probability is how likely something is to happen
chance, as a noun, means there is a non-zero probability of something happening. "there is a chance he could win". I don't think chance really means the same as probability, but colloquially it is interchanged.
I'm open to corrections
The problem with the '58000-1' odds headline is that it is true for all outcomes.
If person 1 wins first, then person 150 wins second, will you expect to see a headline,
"Person 150 wins after Person 1 in 58000-1 shocker"
?
I hope not, and that is why halk is right, and everyone who disagrees just doesn't get it.
If person 1 wins first, then person 150 wins second, will you expect to see a headline,
"Person 150 wins after Person 1 in 58000-1 shocker"
?
I hope not, and that is why halk is right, and everyone who disagrees just doesn't get it.
Soldato
- Joined
- 8 Mar 2007
- Posts
- 10,938
They're essentially different words for the same thing.
The biggest confusion most people have is "in" versus "to". 1 in 10 isn't the same as 1 to 10 for example (the first is more probable)