Premium bonds winnings and odd pattern

HedgehogRulez

Clearly another instance of the patriarchy in action. Rise up sisters

PixelPound

42 is very small sample size for statistical odds. Come back in the year 2525 and see how it's looking then.

On PB most people get less than the advertised rate.

greekpig

Thanks for all your replies. I'll just keep tracking it and see if the disparity continues over a load more years.

Give it some more years for more data to accrue and before long we may have the next Govt scandal if Ernie keeps it up! 

solidpro

"I flipped a coin 5 times and it came up heads 4 times, so it's sure to keep coming up heads more often than tails".

No, it isn't.

fourmarks

akm2018 said:

greekpig said:

Thanks both for the replles.

I started tracking when we noticed the pattern, and I had enough information historically to go back to April 22.

Without giving much away, if you have a large holding you win something in most months.

Bearing in mind neither of us have outstripped the advertised return I'd suggest ernie doesn't like my wife!

Let's do some maths.

Say you have 10 bonds, that is, a £10 holding

In a particular month, each bond has a 1 in 22000 chance of winning something. Which is to say, a 21999/22000 chance of NOT winning something. So with your 10 bonds, your chances of NOT winning something, ten times are:

(21999/22000) ^ 10 = 0.9995455475

So almost every month, you won't win something. But you might!

At the other end of the scale, with the max holding of 50000 bonds, your chances of NOT winning something, with every bond, are:

(21999/22000) ^ 50000 = 0.10302548159

Note that this IS NOT ZERO. Yes, 90% of months you will win SOMETHING but 10% of months you won't.

In the middle, someone with a £10000 holding has a ~ 63% chance of getting nothing, in a given month. They have a 0.4% chance of getting nothing in a given YEAR. They'll probably think there's something wrong, or that ernie doesn't like them, if that happens! But now consider a group of 250 10k-holders. This "something wrong" is quite probably going to happen to one of them, every year...

Ah, now I understand.

wadsRtight

Came here to say exactly the same as OP. Wife and I each have the maximum allowance in Premium Bonds and have held them for a year. After the first few months wife commented how she was receiving zero or small monthly prizes vs my consistent amounts. By the 6 month mark mine was indicating the expected 3 percent ROI. While wife's looked more like 1 percent. At the 12 month mark we had a count-up. I'm at a bit over 3% ROI, she's at about 1.5%.

I've studied statistics up to a point and know about random numbers, sampling sizes and normal curves. Statistically this feels anomalous. With such a strong pattern established we're not expecting the situation to change and the wife is going to move her holdings elsewhere to get a more assured return.

 As said above, pointless querying this with the monolith that is NS&I. Maybe a real statistician can weigh in here. 

HedgehogRulez

A little knowledge is a dangerous thing….

twiglet98

I've held £14 - yes! - since my birth in 1956.  I won £250 in 1965 and £50 in 1975, and nothing since.

It's comforting to occasionally check the list of winning bonds and see that equally tiny holdings do well, sometimes!

Enzo_L

twiglet98 said:

I've held £14 - yes! - since my birth in 1956.  I won £250 in 1965 and £50 in 1975, and nothing since.

It's comforting to occasionally check the list of winning bonds and see that equally tiny holdings do well, sometimes!

According to the BofE inflation calculator £250 in 1965 would be £4270 now.

I bet you were a very happy nine year old!  

eskbanker

wadsRtight said:

Came here to say exactly the same as OP. Wife and I each have the maximum allowance in Premium Bonds and have held them for a year. After the first few months wife commented how she was receiving zero or small monthly prizes vs my consistent amounts. By the 6 month mark mine was indicating the expected 3 percent ROI. While wife's looked more like 1 percent. At the 12 month mark we had a count-up. I'm at a bit over 3% ROI, she's at about 1.5%.

I've studied statistics up to a point and know about random numbers, sampling sizes and normal curves. Statistically this feels anomalous. With such a strong pattern established we're not expecting the situation to change and the wife is going to move her holdings elsewhere to get a more assured return.

 As said above, pointless querying this with the monolith that is NS&I. Maybe a real statistician can weigh in here. 

If you've studied statistics then you should know that 12 draws is nothing like statistically significant, in the context of a random draw with very small expected returns, so (at this stage) your findings really aren't a strong pattern or anomalous at all, as per the comment further up the page by @PixelPound