4 ms·

> There are many properties that are expected to burn more frequently than that. There are 200,000 homes insured by CalFAIR plan, meaning they are in a fire pr

> There are many properties that are expected to burn more frequently than that. There are 200,000 homes insured by CalFAIR plan, meaning they are in a fire prone region, “expected to burn” with some frequency. There are 876 structures damaged or destroyed in 2022 fire season in California. So, the real frequency looks like once in 228 years, while I am paying - as you’ve calculated- for once in 175. CalFIRE plan looks solidly in black.

SideQuark 4y ago
One sample for an incomplete year is not a strong predictor of frequency, especially going forward where the trend is more fires.
- MikePlacid 4y ago
  Why do you call the whole data a sample? Why do you call a year incomplete - the fire season is practically over, we had the first week of rain already, the grass is green again. Anyway, your prediction that I am paying too little for fire insurance seems wrong so far, isn’t it.
  - SideQuark 4y ago
    >Why do you call the whole data a sample? One year is a sample considering that different years have different results and you you pay yearly premiums for pricing. Another way to think: is there more data for houses burning that are not from 2022? For example, you claimed 876 structures so far in 2022, but in 2020 there were 11,116 structures damaged or destroyed [1]. In 2021 there were 3629 [2]. See why making claims on one data point is not statistically valid? A few more years from the same data source: 2012 => 876 2013 => 456 2014 => 471 2015 => 3159 2016 => 1274 2017 => 10280 2018 => 24226 2019 => 732 2020 => 11116 2021 => 3629 Notice any possible trend? >Why do you call a year incomplete because >the fire season is practically over you answered it: the year is not over. And all examples of past data I just presented had relevant fires well past this date. To expect no more this year ignores demonstrable history. >your prediction I made no such prediction. The mean over the above data (and ignoring the expected and demonstrated upward trend) is 5190, far above your single (incomplete) data point of 876. So you computed 1 in 228, but the avg above puts the last 10 years (using your math) at 1 in 38. This also ignores that previous years often had less houses in CalFAIR, which makes it even less in your favor. >I am paying too little for fire insurance seems wrong so far, isn’t it Go ahead and update your calculation using your method above with this better data and see what you obtain. [1] https://www.fire.ca.gov/incidents/2020/ https://www.fire.ca.gov/incidents/2020/ [2] https://www.fire.ca.gov/incidents/2021/ https://www.fire.ca.gov/incidents/2021/

SideQuark 4y ago

One sample for an incomplete year is not a strong predictor of frequency, especially going forward where the trend is more fires.

MikePlacid 4y ago
Why do you call the whole data a sample? Why do you call a year incomplete - the fire season is practically over, we had the first week of rain already, the grass is green again. Anyway, your prediction that I am paying too little for fire insurance seems wrong so far, isn’t it.
- SideQuark 4y ago
  >Why do you call the whole data a sample? One year is a sample considering that different years have different results and you you pay yearly premiums for pricing. Another way to think: is there more data for houses burning that are not from 2022? For example, you claimed 876 structures so far in 2022, but in 2020 there were 11,116 structures damaged or destroyed [1]. In 2021 there were 3629 [2]. See why making claims on one data point is not statistically valid? A few more years from the same data source: 2012 => 876 2013 => 456 2014 => 471 2015 => 3159 2016 => 1274 2017 => 10280 2018 => 24226 2019 => 732 2020 => 11116 2021 => 3629 Notice any possible trend? >Why do you call a year incomplete because >the fire season is practically over you answered it: the year is not over. And all examples of past data I just presented had relevant fires well past this date. To expect no more this year ignores demonstrable history. >your prediction I made no such prediction. The mean over the above data (and ignoring the expected and demonstrated upward trend) is 5190, far above your single (incomplete) data point of 876. So you computed 1 in 228, but the avg above puts the last 10 years (using your math) at 1 in 38. This also ignores that previous years often had less houses in CalFAIR, which makes it even less in your favor. >I am paying too little for fire insurance seems wrong so far, isn’t it Go ahead and update your calculation using your method above with this better data and see what you obtain. [1] https://www.fire.ca.gov/incidents/2020/ https://www.fire.ca.gov/incidents/2020/ [2] https://www.fire.ca.gov/incidents/2021/ https://www.fire.ca.gov/incidents/2021/