NS&I still worth it?

AnotherJoe

polymaff wrote: »

That web-site is a joke,

"post-doctoral cosmology statistician"

indeed! Try it out and you'll find that it makes up exactly the same return on radically different stakes, and even a lesser return on a higher stake.


MSE conned - again!

Pleas give some examples.

polymaff

AnotherJoe wrote: »

Pleas give some examples.

https://imgur.com/klvaemn

Now, who is going to justify

"The MSE premium bond calculator does it more thoroughly"

AnotherJoe

polymaff wrote: »

https://imgur.com/klvaemn

Now, who is going to justify

"The MSE premium bond calculator does it more thoroughly"

You're right.
With 39k you win £500
With £50k you win £500
Yet with £11k (difference between 50 and 39) you win £100.
So if £39 wins £500 then £39k + £11k (=£50k) should win £600.
And 1.2% of £50k is £600, so thats what it should be.

polymaff

AnotherJoe wrote: »

You're right.
With 39k you win £500
With £50k you win £500
Yet with £11k (difference between 50 and 39) you win £100.
So if £39 wins £500 then £39k + £11k (=£50k) should win £600.
And 1.2% of £50k is £600, so thats what it should be.

No - that website is just wrong. As is your analysis, I'm afraid, not least in that it assumes linearity so that the first £11k would receive the same treatment as the last 11k in your investment. It doen't work that way - hence my previously stating that small investors are losers.

I'm a realist, so not at all surprised at the gullibility of even seasoned MSE posters -

hence my challenge to LondonInvestor to justlfy his dotty "The MSE premium bond calculator does it more thoroughly" statement

and also the even barmier "yes you could of course get 0 return but you could equally win something substantial" from 18cc.

or even the super-innumerate "They are meant to earn around 1% interest. With £50K in there you should be winning/earning £25 every other month" from AndyPK.

Here is a simple analogy that you don't need to be a "post-doctoral cosmology statistician" to understand You just need to shop at Sainsbury's.

When you shop in Sainsbury's, your stake is the money you are going to spend and your prize is your Nectar Points. Those Nectar points represent 0.25% - i.e. you get one point per £1 spent and one point has a value of 0.25p. Now just to maintain the analogy let's say you shop on the first of each month - and, remember, just like premium bond your stake is preserved - its just that you take in a stake and come out with a steak. Sorry, to continue.

Your typical stake is £1: half the time you get no nectar points, the other half you get 1 nectar point. 0.5 points per visit !

Your typical stake is £10: well it is either 9 or 10 points per visit.

Your typical stake is £100: well it is either 99 or 100 points per visit.

Now the £1 guy is aready yelling "why are you guys getting proportionally more than me?" Huh - he ain't seen nuthin yet !


Sainsbury thinks that this is such a crowd-puller that they decide to scale the whole system by 100. So 25p Nectar points but only on each £100 spent.

The £1 guy is totally wiped out - so is the £10 guy. The £100 guy is now down to 0.5 point and only a £1,000 or more guys are still doing well

Sainsbury's goes nuts and scales by 100 again. £25 points/prizes and points earned on £10,000.

Well, you can see how things are going. We are now in the land of Premium Bonds where the minimum prize is also £25 and the only thing small holders are going to get is stuffed.


Now, the rate of return. The prize fund varies month-to-month. The only thing that doesn't change very often is the way NS&I divi out the prize money. Currently that is:

Higher Value: 5% of the prize fund (£5,000 and up)
Medium Value: 5% of the prize fund (£500 and £1,000)
Lower Value: 90% of the prize fund (below £500)

As an investor your attitude to the the Higher and Medium Value prizes is "You die before you win" so, you, pragmatically, accept that you only have access to 90% of the prize fund - which means that your interest rate is 90% of the prize fund rate. The present fund rate is 1.4% so the expected rate is 1.26%

That, of course is for the £10,000-plus Sainsbury's spenders. All the small holders get is older.


Now nothing about the above is precise - but the art of not being taken for a ride is to try to find something that is simple enough to understand and also something that is precise enough for your needs.

AnotherJoe

I think once you get above £10k or so it will scale roughly linearly.
Can you explain why you think the first 11k doesn't work as much as the last?
With £100 or £200 or even £1k eg small numbers of course it works that way, but by the time you get to the tens of thousands averaging is working much more effectively. It certainly makes no sense to me that £39k has an expected payout same as £50k or that £50k doesn't have an expected payout of £600 with "average" luck. At the maximum value you can hold you should be in the rough area of 1.2% rather than 1% otherwise something seems wrong, wheres the 0.2% going?

I also find your analogy flawed, in that you posit all or nothing for each £100 spend at Sainsburys (eg one chance to win per £100) but of course you do get 100 chances to win with PB's minute though they may be.

polymaff

AnotherJoe wrote: »

I think once you get above £10k or so it will scale roughly linearly.

That qualitative detail I intentionally didn't deal with, so feel free to elaborate. As I hope that I made clear - mine is an attempt at an explanation which attempts that most-difficult of activities - a simplified model which doesn't actually mislead. Bit like my old simple, visual models

AnotherJoe wrote: »

Can you explain why you think the first 11k doesn't work as much

Just because of the non-linearity. See below.

AnotherJoe wrote: »

I also find your analogy flawed, in that you posit all or nothing for each £100 spend at Sainsburys (eg one chance to win per £100) but of course you do get 100 chances to win with PB's minute though they may be.

That's true, and every pound you have in Premium Bonds carries a micro-chancelet - but you have to have so many of those to get over each £25 threshold in a single draw that low stake-holders are inevitably disadvantaged. The little guy finds that far more of his stake - percentage-wise - never earns anything.

The mathematical basis of all of this is granularity. The granularity of most savings accounts is the penny. We usually lose a fraction of a penny every time interest is computed - but this is so nugatory for most that we all work on the basis that a graph of return per interest period versus sum invested is a straight line heading north-east. Actually, if you zoomed in enough you'd see something a bit odd going on at the origin - and greater granularity just makes that oddity more prevalent. The only things you can say with certainty about that Premium Bond graph is that it starts at the same origin, has an increasing and never negative gradient and that it converges on the straight line. If you accept only that then you confirm that low stakeholders lose out.

eskbanker

polymaff wrote: »

https://imgur.com/klvaemn

Now, who is going to justify

"The MSE premium bond calculator does it more thoroughly"

The answer to this lies in your reference to granularity, in that the underlying analysis of the odds is sound (as far as I can see) but the attempts to present them simplistically (condensed down to a single number) are flawed.

The detailed odds tabs within those results show that for £39K over a year, the odds of winning at least £500 are 51.6%, whereas for £50K it's 87%, clearly a significant difference but one that isn't reflected in the headline presentation of the results. Because the next of the arbitrarily-chosen data points is £750, with odds of only 24.7% (for £50K), they're presenting £500 as the median, when realistically it should be more like £625.

So, anyone choosing to assess comparative returns solely on the basis of the single figure presented on the main results page is likely to be misled, as it simply relays the last of the arbitrarily-chosen data points with a percentage higher than 50%, rather than making any attempt to choose appropriately granular data points....

RG2015

eskbanker wrote: »

The answer to this lies in your reference to granularity, in that the underlying analysis of the odds is sound (as far as I can see) but the attempts to present them simplistically (condensed down to a single number) are flawed.

The detailed odds tabs within those results show that for £39K over a year, the odds of winning at least £500 are 51.6%, whereas for £50K it's 87%, clearly a significant difference but one that isn't reflected in the headline presentation of the results. Because the next of the arbitrarily-chosen data points is £750, with odds of only 24.7% (for £50K), they're presenting £500 as the median, when realistically it should be more like £625.

So, anyone choosing to assess comparative returns solely on the basis of the single figure presented on the main results page is likely to be misled, as it simply relays the last of the arbitrarily-chosen data points with a percentage higher than 50%, rather than making any attempt to choose appropriately granular data points....

Bl**dy hell, mate!

Didn’t understand a word but I would definitely vote for you .

Topaz99

So in short, worth it or not? LOL

dales1

Not worth it for a non-taxpayer or a basic-rate taxpayer - there are higher rates available elsewhere.
But for higher-rate taxpayers, 1.2% or so (tax free and not taxable) is not a bad rate if you do need a 5 figure sum in an easy access account.
Dales.