I’m not convinced there’s a lot of science behind the justification of premia, if you enquire they don’t seem to be able to come up with a sensible explanation.

I once enquired about van insurance as I was thinking about a small Citroen. When they found out that I would be carrying a competition motorcycle they got a bit chairy saying that I could have a professional rider in the passenger seat which might involve large claims for loss of earnings in the event of an accident.

Leaving aside that any driver could have a high earning passenger in any vehicle, I pointed out that the only professional rider in Britain (at that time) in that particular branch of the sport was not an acquaintance and lived hundreds of miles away so the likelihood of their scenario was nil they came out with the “ah but” stuff.

They realised they had painted themselves in to a corner when I pointed out that they had happily insured me for twenty years to tow a three bike trailer and thus, using their logic, I could have had two professional riders in the car.

Shouldn’t the margin of error depend on the sample size?

I think, but could be wrong (because although I paid attention in my stats class, I wasn’t top of the form) that the margin of error is usually the 2.5% at either end of a normally distributed population, i.e. a total of 5%, the outliers that cannot be accounted for.

Sample size is important, but for any given ‘population’ (and the key word here is population) a sample size of 30 is enough to give a reliable prediction of future events. The problem is that your ‘population’ has to homogenous. Something that hoomans are not very good at.

I’m not convinced there’s a lot of science behind the justification of premia, if you enquire they don’t seem to be able to come up with a sensible explanation.

Probably because you’re talking to a keyboard monkey. The guy/girl who wrote the pricing algorithm is sitting in a comfortable house in the home counties at the long end of 7 years minimum studying Dead Hard Maths and a few years writing pricing algorithms.

And where the hell did we get “premia” from? 7 years in insurance and this is the first I’ve heard of it.

Cougar, I ageree with you, on an intellectual level,

That’s good, I’m not overly sure I do. I’m just trying to spark some discussion.

It’s very easy to go “yeah, well, it’s the fairest way of calculating insurance is to do a risk assessment. We know that those young people are all hooligans, that if you live in Liverpool you’re going to get your alloys stolen, and that women never go above 40mph so when they hit things they don’t do as much damage; but of course, you can’t take national descent into account because that’s racist.” But when you stop and think about it, however you look at it, it’s a massively hypocritical situation and we take it as normal (and argue about it on the Internet).

I’m not actually suggesting that the insurance procedure changes, particularly. I’m not sure as I’ve reached a firm conclusion either way, and I don’t really want to be paying through the nose so that Johnny Halfwit can pass his test at 17 and get a Subaru Impreza. (Off the top of my head maybe a fuel tax would be fairer, sporty cars and higher mileage drivers are both higher risk and would pay more accordingly? I don’t know, I’ve not given it a lot of thought.)

Point is, I just can’t help wonder why this is the only industry that’s allowed to do it, and why some demographics are acceptable grounds for analysis and some cause all our knees to jerk.

No one is saying you are excluded from buying this service, they are just saying, the price to customer A is different to the price to customer B because they aren’t buying the same service, i.e. they are buying cover against a risk, but that risk is different with each person, ergo, the product is different, ergo the price is different.

Ok. How would people feel if the London Tube security were routinely checking that young males with tanned complexions, beards and rucksacks weren’t carrying large quantities of explosives (assuming they have this power legally)? After all, the statistical evidence here for terrorism demographics (post-IRA days anyway) must be close to 100%. That’s just using risk assessment to better deploy your manpower.

Finally, you use of ‘statistically’ needs to be sharpened up. The whole ‘scousers more likely to rob cars thing’, you’ll likely find that there is no correlation or if there is it’s pretty weak.

Well no, your example recognition needs to be sharpened up. I never said (or intended to imply) that there was a correlation, I said “If I had evidence that…” and even clarified that it was hypothetical in a follow-up post.

But if there’s some underlying factor which causes men to present a greater risk than women it’s fairer to use that factor.

I suspect that you’re assuming that a causal factor will be more tightly correlated to whatever you’re trying to predict than any proxy, and what you actually want, if you really get down to it, is not causality but very strong correlation. That’s fine, it’s undeniable that (higher correlation)=(better prediction)=(fairer premium), no argument there. The thing about cause is that:

1. To untangle all the causes of a particular behaviour/phoneomenon, particularly if it’s human, can be fiendishly difficult or impossible. A proxy (like the red car example) can be better, and therefore fairer, than an incomplete model of cause-effect.

2. Even if you somehow manage to untangle all the causes of something, they may not be measurable or practical to measure. So an easily collected proxy may be your best bet at fairness, even if it’s not perfect.

Like I said before, don’t let the perfect be the enemy of the good.

And where the hell did we get “premia” from?

Premium: Noun (plural premiums or premia)

Cougar – there is a HUGE amount of data nehind premium rating, it’s just that as you only speak to a customer service drone nowadays rather than an actual underwriter, they won’t have a clue!

How would people feel if the London Tube security were routinely checking that young males with tanned complexions, beards and rucksacks weren’t carrying large quantities of explosives

I would think that al qaeda will certainly be stepping up recruitment amongst 70 year old chinese women.

Just found out that it’s the Belgians that are to blame for all of this. Pesky Belgians, what good has their existance ever done?!

I would think that al qaeda will certainly be stepping up recruitment amongst 70 year old chinese women.

(-:

‘s not the same. The whole point is to instil terror by the act of someone so dedicated to their cause that they’re willing to die for it. You’re supposed to feel hopeless in the face of their fanaticism, or something. An “outsider” bomber wouldn’t be the same, you might as well just hide a bomb.

Pesky Belgians, what good has their existance ever done?!

Apart from the beer, and the chocolate, and the waffles? And, curry ketchup.

it’s a massively hypocritical situation and we take it as normal

Hear hear.
That’s why there shouldn’t be any variables which are off limits.

Touche Cougar, actualy have a bottle of Curry Gewurz Ketchup in the cupboard! The g/f is half Belgian too, but that doesn’t count 😉

Which half?

I was kinda going for the “Romans, what have they ever done for us” thing, but I don’t think it worked.

The whole point is to instil terror by the act of someone so dedicated to their cause that they’re willing to die for it

Is it?!
My impression is that al qaeda et al are waging a war using any and all means available to them, and wouldn’t be too picky about the method of delivery. It’s just that suicide bombers can be efficient without the need for great technological sophistication.

In fact I’m not sure if I’d be more terrorised by those who blow themselves up for religion than those who hide bombs for money. The second one is scarier.

But maybe you’re right, because it’s true that people dying for their beliefs does hold a certain special message.

Assuming you ARE right, and if the tube decided to start searching people, tanned beardies with backpacks do seem like an effective place to start. IF the goal is to find terrorists. If the goal, on the other hand, is to make you feel all warm and fuzzy inside, random will do.

Which half?

left iirc

Yup, went over my head

for instance, if someone is called ‘steve’ they get a higher premium by someone called ‘shirly’

So does Stevie Nicks pay more for insurance than Big Daddy?

…and how much does Steve Shirley pay for insurance?