Premium bonds - interesting stats since the beginning of the year

eskbanker

DaveMM said:

eskbanker said:

DaveMM said:

for the bigger prizes it looks like holding the full allowance does provide more chance of a higher prize

That analysis is flawed and misleading, in that over any sensible timeframe it should be roughly linear - are you ignoring the fact that the number of people holding the maximum £50K will hugely outnumber each smaller holding (according to some analysis last year, 43% of all premium bonds are held by maximum holders)?  In other words, any given £50K holding should win about twice as many prizes as any given £25K one, rather than it being a distorted spike like that chart suggests....

it is not comparing anything - that spike exists in the data (4 months) given the value of holding and prizes won counted. Not ignoring any facts not provided - if the other data was available it would make more interesting reading

Yes, I wasn't suggesting that the number of holders by each size of holding is readily available data, but just highlighting that without that weighting the chart is objectively meaningless and certainly can't legitimately be used to support a conclusion about chances of winning prizes based on size of holding!  If (visually from that chart) £50K holders win about twenty times as many prizes as £49K holders, that probably signifies no more than that there are about twenty times as many £50K holders as there are £49K holders, but that says nothing about the chances of winning with either holding....

eskbanker

FatherTireseus said:

eskbanker said:

hangryconsumer said:

Why have you excluded Scotland but included Wales?

The geographic information appears to be truncated to fit the page, so it doesn't show all of the 'counties', just the top 25-30.

However, NS&I's approach to location mapping is highly inconsistent, so in the currently published big prize results they report Wales as one entity but Scotland is subdivided into Central, West, South, North East, and Highlands & Islands, while they show London, Inner London, Outer London, and (separately) Wandsworth, Ealing, and Bromley!  In other words, trying to analyse results based on location needs a more rigorous structured data mapping than that currently performed by NS&I....

How NS&I announce locations depends on how many bond holders live in the area.  In the case of Ealing, for example, based on NS&I's information,  there must be at least 100,000 bond holders living there.  But in other London boroughs there are presumably fewer which is why they may say 'outer London'.

https://www.nsandi.com/get-to-know-us/prize-winner-locations

I'd have thought that setting a limit of 100,000 bond holders in determining a location is a pretty high bar.  After all, no-one is going to know you hold premium bonds unless you go round telling everyone.   And if you do win the £1m prize,  the shiny new car in the drive (or whatever) may be more of a giveaway to the neighbours that you've come into money.

Thanks - I hadn't seen that page so now understand how they do it, but even though there is some logic to it, it still leaves inconsistencies.  It's unclear what would count as simply London, i.e. not in Inner London or Outer London, and the latter effectively means 'Outer London excluding boroughs with more than 100,000 PB holders, such as Bromley and Ealing'!

JohnnyB70

DaveMM said:

eskbanker said:

DaveMM said:

for the bigger prizes it looks like holding the full allowance does provide more chance of a higher prize

That analysis is flawed and misleading, in that over any sensible timeframe it should be roughly linear - are you ignoring the fact that the number of people holding the maximum £50K will hugely outnumber each smaller holding (according to some analysis last year, 43% of all premium bonds are held by maximum holders)?  In other words, any given £50K holding should win about twice as many prizes as any given £25K one, rather than it being a distorted spike like that chart suggests....

it is not comparing anything - that spike exists in the data (4 months) given the value of holding and prizes won counted. Not ignoring any facts not provided - if the other data was available it would make more interesting reading

It’s not showing what you are implying. Each extra pound invested brings the same expected return; the graph that you posted just shows that there are a lot of people with the maximum investment allowed.

talexuser

Premium bonds have a place, but unless you've maxed out your PSA, there are better options. 

The tax position is the primary consideration that needs minimising for us. I have no personal savings allowance and the OH adding interest going into 40% would be self defeating.

DaveMM

JohnnyB70 said:

DaveMM said:

eskbanker said:

DaveMM said:

for the bigger prizes it looks like holding the full allowance does provide more chance of a higher prize

That analysis is flawed and misleading, in that over any sensible timeframe it should be roughly linear - are you ignoring the fact that the number of people holding the maximum £50K will hugely outnumber each smaller holding (according to some analysis last year, 43% of all premium bonds are held by maximum holders)?  In other words, any given £50K holding should win about twice as many prizes as any given £25K one, rather than it being a distorted spike like that chart suggests....

it is not comparing anything - that spike exists in the data (4 months) given the value of holding and prizes won counted. Not ignoring any facts not provided - if the other data was available it would make more interesting reading

It’s not showing what you are implying. Each extra pound invested brings the same expected return; the graph that you posted just shows that there are a lot of people with the maximum investment allowed.

Not implying anything - that is the data - agreed every pound brings the same expected return - but the winners do spike at that point, and yes is most likely due to more people holding the maximum, and it does not allow knowing about the same people winning. The only thing I have done is used the nearest (up) £1000 to only give 50 points

the raw data - of 8608 prizes since 2022 -

548 held Less than £10000

733 held between £10000 - £20000

940 held between £20000 - £30000

1037 held between £30000 - £40000

1453 held between £40000 - £50000

3897 held £50000

Like I say - just thought it was interesting - obvs not

one person held £50 since 2007 and won £10k in 2022
one held £180 since 1998 and won £25k in 2022
those are great returns 

DaveMM

eskbanker said:

DaveMM said:

eskbanker said:

DaveMM said:

for the bigger prizes it looks like holding the full allowance does provide more chance of a higher prize

That analysis is flawed and misleading, in that over any sensible timeframe it should be roughly linear - are you ignoring the fact that the number of people holding the maximum £50K will hugely outnumber each smaller holding (according to some analysis last year, 43% of all premium bonds are held by maximum holders)?  In other words, any given £50K holding should win about twice as many prizes as any given £25K one, rather than it being a distorted spike like that chart suggests....

it is not comparing anything - that spike exists in the data (4 months) given the value of holding and prizes won counted. Not ignoring any facts not provided - if the other data was available it would make more interesting reading

Yes, I wasn't suggesting that the number of holders by each size of holding is readily available data, but just highlighting that without that weighting the chart is objectively meaningless and certainly can't legitimately be used to support a conclusion about chances of winning prizes based on size of holding!  If (visually from that chart) £50K holders win about twenty times as many prizes as £49K holders, that probably signifies no more than that there are about twenty times as many £50K holders as there are £49K holders, but that says nothing about the chances of winning with either holding....

Agreed - do not have the rest of the data that would make the weighting more relative 

plumb1_2

Bought mine end of March, 1 st draw this month and won 2x£25 

lozzy1965

plumb1_2 said:

Bought mine end of March, 1 st draw this month and won 2x£25 

Not wishing to criticise you but there is already a thread where people delight in telling everyone what they win each month and reading about what other posters' won each month.
https://forums.moneysavingexpert.com/discussion/81778/premium-bond-winner#latest

coyrls

lozzy1965 said:

plumb1_2 said:

Bought mine end of March, 1 st draw this month and won 2x£25 

Not wishing to criticise you but there is already a thread where people delight in telling everyone what they win each month and reading about what other posters' won each month.
https://forums.moneysavingexpert.com/discussion/81778/premium-bond-winner#latest

The OP in that thread was testing their theory that newer bonds win more prizes.  After 17 years of non-scientific data, I don't think that they have ever come back to share their conclusions.