Dareen Brown: Which do we belive?e

[PaulB]

He's claiming that his prediction was based on the idea that taking an average of a large number of people's guess of the weight of a cow will bring an accurate result.

Trouble is, when you guess the weight of a cow you've got the thing in front of you and have something to work from.

The lotto numbers are completely random, so the average of a number of guesses will be equally as random. And the same system doesn't work.

It was camera trickery, a very poor trick, and he's done himself a dis-service.

It must be ace to know everything like you do.

 

[srw]

The lotto numbers are completely random, so the average of a number of guesses will be equally as random.

Errmmm. No. If you ask people to guess a random number between 1 and 49 and take the average of the guesses, it will cluster around 25 or so. If you tell people you want lottery numbers and ask them for any random number, most will dutifully give you a number between 1 and 49 - and the average of the guesses will still be somewhere close to 25.

 

[PBancroft]

Errmmm. No. If you ask people to guess a random number between 1 and 49 and take the average of the guesses, it will cluster around 25 or so.

No it won't.

If you tell people you want lottery numbers and ask them for any random number, most will dutifully give you a number between 1 and 49 - and the average of the guesses will still be somewhere close to 25.

Seriously, no, it won't. Go do a blind study if you don't believe me.

 

[Over The Hill]

Clearly it is all just trickery and diversion.

It is interesting to see a lot of posts on here talking about him predicting the lottery. He did not do that.
He did not reveal the numbers before the call and then spent about a minute wasting time writing them down on a big board.
The trick was really how did he get the recently announced numbers onto the row of balls?
Now you could think up many ways to do that the most obvious is to put a little LCD screen into each ball programed to change to the right number or simply project the numbers up from below with the same kit you use for a powerpoint display.
I saw the "trick" but could not be bothered to hang around for the answer that he was supposed to reveal on Friday.

 

[Over The Hill]

And if you could guess the number then save all the hocus pocus and just get out the winning ticket from your pocket.

 

[srw]

Seriously, no, it won't. Go do a blind study if you don't believe me.

Yes it will. I just did. It's a piece of cake on Excel. =CEILING(rand(),1) is the magic incantation.

It's also bleeding obvious if you know the stats. You're generating numbers from a Uniform(0,49) distribution and rounding them up to the next whole number above. So what you're doing is finding 0.5 more than the mean of a Uniform(0,49) distribution. Which is 25.

If you allow people to pick any number they can think of, but tell them it's for lottery numbers, most will pick a number between 0 and 49 and your average is still likely to be around 25.

If you allow people to pick any number they can think of without restriction, the average of the picks is still likely to be a relatively small positive number. That's because most people don't naturally think of negative numbers or large numbers unless you suggest it to them.

One of the sets of "lottery numbers" Brown got his gullible victims to pick was very clustered around 25 - I believe it was the first one, which was the only one where he actually allowed them to see the calculations.

 

[srw]

Yeah ok. The point is though that each average has as much chance of being correct as any one person's guess. The odds of it being correct don't improve at all.

Oh, of course. I take that for granted. It's impossible to predict lottery numbers in a way that makes it more likely that you'll win.

I can trivially predict that the lottery numbers will be six different numbers between 1 and 49, of course - but that's not a very useful prediction.

 

[PBancroft]

Yes it will. I just did. It's a piece of cake on Excel. =CEILING(rand(),1) is the magic incantation.

With all due respect to Microsoft, Excel is not people.

People pick numbers for the lottery based on all sorts of whimsical things, from birthdays to ages to house numbers to phone numbers.

Birthday dates are going to run from 1 to 31. Months from 0 to 12. Years from anywhere between 1 (if they pick family) and 109 (but for the purposes of this argument 49).

Also, people find it more difficult to think of really big numbers. I'm not talking 49 here, I'm talking five or six digit numbers. But the same thing has an impact - people will simplify BIG numbers which mean something to them. If you ask people to pick a truly random number you might be correct about your averages.

But picking lottery numbers isn't an average. We are drawn to certain figures. Like I said, go out and do a proper blind test - I dare you to prove yourself right and me wrong.

 

[srw]

If you ask people to pick a truly random number you might be correct about your averages.

But picking lottery numbers isn't an average. We are drawn to certain figures. Like I said, go out and do a proper blind test - I dare you to prove yourself right and me wrong.

I am right about my averages if I'm asking people to pick a random number. And that's what Brown was asking (and that's the way I framed my scenario).

I know that real lottery numbers tend to be "favourite" numbers. But even allowing for this the average of a bunch of guesses will still be very predictable. It might be less than 25 because people slightly favour the bottom end of the distribution, but it won't be 49 and it won't be 1. It's very unlikely to be over 40, and it's very unlikely to be under 10.

 

[Piemaster]

Isn't there something in this about people not wanting random numbers to actually be random? The example I've seen quoted is a shuffle on an mp3 player. True random would give every track in the memory an equal chance of playing next, whereas what we actually want is all the tracks to be played in a random order. Not actually the same thing.

 

[PBancroft]

I am right about my averages if I'm asking people to pick a random number. And that's what Brown was asking (and that's the way I framed my scenario).

I know that real lottery numbers tend to be "favourite" numbers. But even allowing for this the average of a bunch of guesses will still be very predictable. It might be less than 25 because people slightly favour the bottom end of the distribution, but it won't be 49 and it won't be 1. It's very unlikely to be over 40, and it's very unlikely to be under 10.

But that's the point. The participants knew it was related to the lottery, therefore they had a lottery mindset. The average of a bunch of numbers is very guessable, but it wouldn't be 25 as you suggest. That's far too simplistic.

 

[gavintc]

Simplistic - just like the lottery. The government has successfully created a 'stupidity tax'. I am amazed at the level of interest in a television scam.

 

[02GF74]

With the lottery, every number on every ball is a complete guess. So even if you ask 1 million people then each one of them can only pluck a number out of the air. An average of these guesses will therefore be equally as random.

no, people playing the lotto do not use random guess but certain numbers that have menaing e.g. birthdays, age, house number you name it will pop more oftern. Also numbers associated with e.g. 7 and 13 will be more popular.

Aslo certain sequences e.g. 1,2,3,4,5,6,7 and 7,14,21,28,35 ... appear more than would if totally random.

 

[CopperBrompton]

It has been said that if 1,2,3,4,5,6 ever wins, each 'jackpot' winner will find themselves with about £400 each as so many people choose those numbers

 

[rich p]

Every lottery number chosen is a guess.

A collection of lottery guesses will be random, with account taken for some of the traits described.

Which would have a modicum of relevance if the actual numbers chosen were picked by a person.

As it is, the 6 numbers chosen are completely random by a machine. There is no way that using the 'guess the weight of the cow' idea will give you any more chance of winning than picking your own numbers out of a hat.

That's the point but it's taking a long time getting there!