Insurance madness.
Discussion
Online quotes for a Bandit 650 to have a bit of fun with and maybe commute.
Estimated annual mileage of 9000 is £2 cheaper for the year than estimated annual mileage of 5000 and the excess also drops from £250 to £175; same company and details throughout.
I just don't understand algorithms.
Gavia, care to explain...?
Estimated annual mileage of 9000 is £2 cheaper for the year than estimated annual mileage of 5000 and the excess also drops from £250 to £175; same company and details throughout.
I just don't understand algorithms.
Gavia, care to explain...?
I can’t explain it, but I can hazard a few guesses.
When you say same company, who is it? Quite often people get confused between v
Broker amd underwriter. For example Carole Nash might be quoting both of those prices, but they are a broker, so could be two completely different underwriters.
Other than that, does it matter, you get more miles for less money. Result
When you say same company, who is it? Quite often people get confused between v
Broker amd underwriter. For example Carole Nash might be quoting both of those prices, but they are a broker, so could be two completely different underwriters.
Other than that, does it matter, you get more miles for less money. Result
Could it be that some people who do really high mileage will lie and say under 5000 a year, whereas anyone who says up to 9000 is telling the truth.
So in the up to 9000, no one is exceeding that, but in the up to 5000 group, you have a small percentage of liars doing 50K miles a year. Hence up to 9000 has better claims results.
Just a theory.
So in the up to 9000, no one is exceeding that, but in the up to 5000 group, you have a small percentage of liars doing 50K miles a year. Hence up to 9000 has better claims results.
Just a theory.
TwigtheWonderkid said:
Could it be that some people who do really high mileage will lie and say under 5000 a year, whereas anyone who says up to 9000 is telling the truth.
So in the up to 9000, no one is exceeding that, but in the up to 5000 group, you have a small percentage of liars doing 50K miles a year. Hence up to 9000 has better claims results.
Just a theory.
Unlikely, bike insurance is a fairly old school way of doing things, as the money is small fry to insurers so they more or less let brokers do what they want. Quite often the brokers systems can’t cope with all the algorithms in play, so it spits out the odd discrepancy. So in the up to 9000, no one is exceeding that, but in the up to 5000 group, you have a small percentage of liars doing 50K miles a year. Hence up to 9000 has better claims results.
Just a theory.
My money is still on it being via one broker, but two different underwriters.
Gavia said:
TwigtheWonderkid said:
Could it be that some people who do really high mileage will lie and say under 5000 a year, whereas anyone who says up to 9000 is telling the truth.
So in the up to 9000, no one is exceeding that, but in the up to 5000 group, you have a small percentage of liars doing 50K miles a year. Hence up to 9000 has better claims results.
Just a theory.
Unlikely, bike insurance is a fairly old school way of doing things, as the money is small fry to insurers so they more or less let brokers do what they want. Quite often the brokers systems can’t cope with all the algorithms in play, so it spits out the odd discrepancy. So in the up to 9000, no one is exceeding that, but in the up to 5000 group, you have a small percentage of liars doing 50K miles a year. Hence up to 9000 has better claims results.
Just a theory.
My money is still on it being via one broker, but two different underwriters.
Dog Star said:
Reminds me of that weird stuff that went on with Bennett's quote engine last year - I was getting quotes for Mrs DS's R1 and the more goodies I added made the premium go down - for example lower excess, lower premium, higher mileage, NCB protection etc. 
Same this year.Renewal - £640
All mods declared with Bennetts - £350.
Quotes were from Bennets.
Started at 5000 miles then hopped up and down by 1000 miles each time.
No difference at all between the different prices (although lower mileage meant pennies cheaper but a much higher excess) until it hit the magic number of 9000.
Going beyond that just returned the premium to the same level as before.
Started at 5000 miles then hopped up and down by 1000 miles each time.
No difference at all between the different prices (although lower mileage meant pennies cheaper but a much higher excess) until it hit the magic number of 9000.
Going beyond that just returned the premium to the same level as before.
Dog Star said:
Reminds me of that weird stuff that went on with Bennett's quote engine last year - I was getting quotes for Mrs DS's R1 and the more goodies I added made the premium go down - for example lower excess, lower premium, higher mileage, NCB protection etc.
Aaaaand the Bennetts quote engine does it again.Sorting out Mrs DS's R1 policy and Bennetts quote (with me on as named rider and commuting) goes down £8 to £95 if I *ADD* protected NCB. I've tried all the other tricks of increasing mileage but can't get it lower than that.
Rubin215 said:
Online quotes for a Bandit 650 to have a bit of fun with and maybe commute.
Estimated annual mileage of 9000 is £2 cheaper for the year than estimated annual mileage of 5000 and the excess also drops from £250 to £175; same company and details throughout.
I just don't understand algorithms.
Gavia, care to explain...?
very few people understand algorithms.Estimated annual mileage of 9000 is £2 cheaper for the year than estimated annual mileage of 5000 and the excess also drops from £250 to £175; same company and details throughout.
I just don't understand algorithms.
Gavia, care to explain...?
I would venture and say actually noone understands them by now, probably created by some guy 10 years ago, no documentation, 100s of variables, dozens of scripts highly interconnected computing something with some sort of a regression full of adjustments that you can't find and learn about in any book because it was something that the developer "knew" from somewhere. The current operators are probably just happy when the scripts run without throwing off an error of some sort...
Ho Lee Kau said:
very few people understand algorithms.
I would venture and say actually noone understands them by now, probably created by some guy 10 years ago, no documentation, 100s of variables, dozens of scripts highly interconnected computing something with some sort of a regression full of adjustments that you can't find and learn about in any book because it was something that the developer "knew" from somewhere. The current operators are probably just happy when the scripts run without throwing off an error of some sort...
Usually the opposite - someone (or more likely a team of someones) is working every single day tweaking it to their advantage, and researching how to improve it.I would venture and say actually noone understands them by now, probably created by some guy 10 years ago, no documentation, 100s of variables, dozens of scripts highly interconnected computing something with some sort of a regression full of adjustments that you can't find and learn about in any book because it was something that the developer "knew" from somewhere. The current operators are probably just happy when the scripts run without throwing off an error of some sort...
Can confirm we have 100's of people constantly fiddling with the algo's. All data goes into a data lake, which is constantly analysed and reviewed, we also pull data in from third parties. Due to comparison websites we have about 70% of the population (each country) in our database. The data mining potential is huge and is continuously updated. The closer our algo's are perfected the closer we are to ensuring we are competitive, managing our risk, look after our customers, and make money.
On occasion you will get an erroneous response and that could be because we simply 1) we want to increase the numbers on our books 2) we do not want you on our books 3) It could be a package deal to which we might be pushing with kickbacks 4) algo error. For point two, this may be due to location, gender, age, risk, machine, etc. etc.
We insure based on the herd. e.g. we offset your potential loss against all the other people pooled in that group. This is why we limit the number of people and risk in that group - to many people in one location insured, it only takes a flooding and we take a pounding (3). New bikes mean parts are expensive, so if there is not enough people on the books or we can not cover the risk elsewhere then we don't insure.
On occasion you will get an erroneous response and that could be because we simply 1) we want to increase the numbers on our books 2) we do not want you on our books 3) It could be a package deal to which we might be pushing with kickbacks 4) algo error. For point two, this may be due to location, gender, age, risk, machine, etc. etc.
We insure based on the herd. e.g. we offset your potential loss against all the other people pooled in that group. This is why we limit the number of people and risk in that group - to many people in one location insured, it only takes a flooding and we take a pounding (3). New bikes mean parts are expensive, so if there is not enough people on the books or we can not cover the risk elsewhere then we don't insure.
BobSaunders said:
Can confirm we have 100's of people constantly fiddling with the algo's. All data goes into a data lake, which is constantly analysed and reviewed, we also pull data in from third parties. Due to comparison websites we have about 70% of the population (each country) in our database. The data mining potential is huge and is continuously updated. The closer our algo's are perfected the closer we are to ensuring we are competitive, managing our risk, look after our customers, and make money.
On occasion you will get an erroneous response and that could be because we simply 1) we want to increase the numbers on our books 2) we do not want you on our books 3) It could be a package deal to which we might be pushing with kickbacks 4) algo error. For point two, this may be due to location, gender, age, risk, machine, etc. etc.
We insure based on the herd. e.g. we offset your potential loss against all the other people pooled in that group. This is why we limit the number of people and risk in that group - to many people in one location insured, it only takes a flooding and we take a pounding (3). New bikes mean parts are expensive, so if there is not enough people on the books or we can not cover the risk elsewhere then we don't insure.
There have been, (and are still) a number of insurance types on piston heads over the years.On occasion you will get an erroneous response and that could be because we simply 1) we want to increase the numbers on our books 2) we do not want you on our books 3) It could be a package deal to which we might be pushing with kickbacks 4) algo error. For point two, this may be due to location, gender, age, risk, machine, etc. etc.
We insure based on the herd. e.g. we offset your potential loss against all the other people pooled in that group. This is why we limit the number of people and risk in that group - to many people in one location insured, it only takes a flooding and we take a pounding (3). New bikes mean parts are expensive, so if there is not enough people on the books or we can not cover the risk elsewhere then we don't insure.
But that is the best explanation of how insurance companies approach setting premiums i've read on here.
Thanks!
Gassing Station | Biker Banter | Top of Page | What's New | My Stuff