3 ms·

Mathematically, if you sell insurance at break even, you're guaranteed to go bankrupt - on an infinite time scale, the "spike" of a random walk martingale (this

by tomp 3y ago

Mathematically, if you sell insurance at break even, you're guaranteed to go bankrupt - on an infinite time scale, the "spike" of a random walk martingale (this last word means, it doesn't make a profit) will exceed every level, i.e. it will wipe out any amount of collateral / capital / equity the company might have. https://en.wikipedia.org/wiki/Law_of_the_iterated_logarithm https://en.wikipedia.org/wiki/Law_of_the_iterated_logarithm

cortesoft 3y ago
This is why you have re-insurance https://en.wikipedia.org/wiki/Reinsurance https://en.wikipedia.org/wiki/Reinsurance If the final insurer is the government, you don't have the risk of ruin because you have control of the money printer.
- TylerE 3y ago
  If money printer goes brr… you’re losing, not winning.
- MajimasEyepatch 3y ago
  Reinsurance isn’t magic. This helps with one-off losses, but if you’re fundamentally not able to make a profit, they’re not going to cover you, because all you’ve done is shift the negative expected value to them.
  - TylerE 2y ago
    It also probably increases
  - TylerE 2y ago
    It also probably increases the odds of total ruin... think a Katrina or an Andrew but a bit worse. On a smaller scale, it's never gonna be just one car in a town with hail damage.
bdjsiqoocwk 3y ago
What a bunch of nonsense. If you believe that in an infinite time scale the spike of a "random walk martingale" will exceed every level, then you also believe that you'll go bankrupt even if you don't sell insurance at break even. Maybe mathematically incorrect, but entirely irrelevant in the real world. IN ADDITION, the money that insurers make isn't just the underwriting profit but also the investment profit. You you're talking twice as much shit as the average HN commenter.
- tomp 3y ago
  I don’t ”believe” in math, I can prove it. If you don’t sell at break even, it’s not a martingale, so the Law doesn’t apply.
  - bdjsiqoocwk 2y ago
    You can prove mathematical propositions, you obviously can't make truthful conclusions about insurance. And that's really the crucial parts. Anyone can make prove mathematical statements. Reminds of the guy who lost his keys in the darkness and was looking for the keys under the lamp because that's where he can see. Likewise, you're using your tools and hoping that the tools have some connection to real life. You sound like one of those "that's all good in practice but it would never work in theory" type of people.
    - tomp 2y ago
      The real world is much more complicated that mathematical models. But if your business goes bankrupt with probability 100% with even a simplified mathematical model, I wouldn't want to invest in it.

by tomp 3y ago

cortesoft 3y ago

This is why you have re-insurance https://en.wikipedia.org/wiki/Reinsurance https://en.wikipedia.org/wiki/Reinsurance If the final insurer is the government, you don't have the risk of ruin because you have control of the money printer.

TylerE 3y ago
If money printer goes brr… you’re losing, not winning.
MajimasEyepatch 3y ago
Reinsurance isn’t magic. This helps with one-off losses, but if you’re fundamentally not able to make a profit, they’re not going to cover you, because all you’ve done is shift the negative expected value to them.
- TylerE 2y ago
  It also probably increases
- TylerE 2y ago
  It also probably increases the odds of total ruin... think a Katrina or an Andrew but a bit worse. On a smaller scale, it's never gonna be just one car in a town with hail damage.

bdjsiqoocwk 3y ago

What a bunch of nonsense. If you believe that in an infinite time scale the spike of a "random walk martingale" will exceed every level, then you also believe that you'll go bankrupt even if you don't sell insurance at break even. Maybe mathematically incorrect, but entirely irrelevant in the real world. IN ADDITION, the money that insurers make isn't just the underwriting profit but also the investment profit. You you're talking twice as much shit as the average HN commenter.

tomp 3y ago
I don’t ”believe” in math, I can prove it. If you don’t sell at break even, it’s not a martingale, so the Law doesn’t apply.
- bdjsiqoocwk 2y ago
  You can prove mathematical propositions, you obviously can't make truthful conclusions about insurance. And that's really the crucial parts. Anyone can make prove mathematical statements. Reminds of the guy who lost his keys in the darkness and was looking for the keys under the lamp because that's where he can see. Likewise, you're using your tools and hoping that the tools have some connection to real life. You sound like one of those "that's all good in practice but it would never work in theory" type of people.
  - tomp 2y ago
    The real world is much more complicated that mathematical models. But if your business goes bankrupt with probability 100% with even a simplified mathematical model, I wouldn't want to invest in it.