"Overdue for a hurricane..."

[Toadstool]

I notice a few news shows on TV and posters stating certain places are "overdue for a hurricane." (Such as NYC and Miami, etc.)

Statistically this doesn’t make sense to me, nor does it align with the field of statistics. If someone gives you a 6-sided die and asks you what are the chances you’ll roll a 4, the answer (if it’s a fair die) will be 1/6. If you are then asked what are the chances you will roll a 4 again on the second try, the answer is still 1/6. The probability does not change.

I know some people look at other phenomena like earthquakes and they hear LA is overdue for one, but that is based on a building of pressure in the tectonic plates over time so there is a reason behind it. I do not yet see any reason why a particular city is "overdue" for a hurricane.

It’s kind of like how people took hold of the Coriolis effect where tropical cyclones spin counter-clockwise in the Northern hemisphere vs. clockwise in the Southern hemisphere; and then extrapolated that their sinks and toilets must show the same behavior (which they don’t).

[RL3AO]

You are correct. Hurricanes are pretty much independent trails.

[terstorm1012]

I think "overdue" is based on statistical probability. I personally pay very little attention to it. I mean if we're going to talk overdue, the Midwest is "overdue" for a moderate Earthquake and where I live is "overdue" for a big flood. All based on very loose probability.

[Toadstool]

I think "overdue" is based on statistical probability. I personally pay very little attention to it. I mean if we're going to talk overdue, the Midwest is "overdue" for a moderate Earthquake and where I live is "overdue" for a big flood. All based on very loose probability.

Hi Terstorm,

I used to live in the midwest so I know what you mean about the New Madrid fault line being overdue, but again that's due to stresses building up over time. I still am not seeing a statistical link for knowing when a given city (or lat/long) is overdue for a hurricane. There very well may be a hidden pattern not yet discovered, but so far I have not seen one.

[Regit]

I think "overdue" is based on statistical probability. I personally pay very little attention to it. I mean if we're going to talk overdue, the Midwest is "overdue" for a moderate Earthquake and where I live is "overdue" for a big flood. All based on very loose probability.

Hi Terstorm,

I used to live in the midwest so I know what you mean about the New Madrid fault line being overdue, but again that's due to stresses building up over time. I still am not seeing a statistical link for knowing when a given city (or lat/long) is overdue for a hurricane. There very well may be a hidden pattern not yet discovered, but so far I have not seen one.

With events with no exhaustion of choices, there is no change in probability. For instance, if you have a child, the odds of it being a boy are approximately 1 in 2. If you've had 9 children and they're all girls, the odds of the next being a boy are still approximately 1 in 2. It's impossible to be "overdue" when the previous events have no baring on the next.

Now if you have a bag with 16 red balls and 4 green balls, and you draw 8 reds in a row, discarding each ball that is drawn, you are overdue for a green ball in that event since your odds have improved from 1 in 5 to 1 in 3. Also there is a guarantee you'll eventually draw a green ball.

In the child example, there's no guarantee you ever get a boy, and with weather there's no guarantee that any certain event will ever happen again, much less "on time."

[GeneratorPower]

Regit and others, a very good discussion of probability! I haven't heard such a great discussion in a long time. Very accurate and scientific.

[Toadstool]

I think "overdue" is based on statistical probability. I personally pay very little attention to it. I mean if we're going to talk overdue, the Midwest is "overdue" for a moderate Earthquake and where I live is "overdue" for a big flood. All based on very loose probability.

Hi Terstorm,

I used to live in the midwest so I know what you mean about the New Madrid fault line being overdue, but again that's due to stresses building up over time. I still am not seeing a statistical link for knowing when a given city (or lat/long) is overdue for a hurricane. There very well may be a hidden pattern not yet discovered, but so far I have not seen one.

With events with no exhaustion of choices, there is no change in probability. For instance, if you have a child, the odds of it being a boy are approximately 1 in 2. If you've had 9 children and they're all girls, the odds of the next being a boy are still approximately 1 in 2. It's impossible to be "overdue" when the previous events have no baring on the next.

Now if you have a bag with 16 red balls and 4 green balls, and you draw 8 reds in a row, discarding each ball that is drawn, you are overdue for a green ball in that event since your odds have improved from 1 in 5 to 1 in 3. Also there is a guarantee you'll eventually draw a green ball.

In the child example, there's no guarantee you ever get a boy, and with weather there's no guarantee that any certain event will ever happen again, much less "on time."

Hey Regit, I definitely agree, and a good way of stating it. In earthquakes, each year that goes by builds up stress, so that like reducing your choices are you stated. With hurricanes I agree there's no exhaustion of choices (that I have seen so far).

[Downdraft]

I think "overdue" is based on statistical probability. I personally pay very little attention to it. I mean if we're going to talk overdue, the Midwest is "overdue" for a moderate Earthquake and where I live is "overdue" for a big flood. All based on very loose probability.

Hi Terstorm,

I used to live in the midwest so I know what you mean about the New Madrid fault line being overdue, but again that's due to stresses building up over time. I still am not seeing a statistical link for knowing when a given city (or lat/long) is overdue for a hurricane. There very well may be a hidden pattern not yet discovered, but so far I have not seen one.

There is a major difference between predicting or calling a hurricane "overdue" and a seismic event. One is based upon a probability of occurrence the other is based upon a mathematical certitude of occurrence. Big difference! In the case of a hurricane striking a particular area such as New York or Tampa the probability is not based on a changing set of odds each year. Every year the odds so to speak reset to the same probability. In the case of the New Madrid fault or any another fault line every year a quake doesn't happen contributes to the probability for the following year in the sense that forces continue to build along the plate line.

[terstorm1012]

Hi, thanks, great discussion..what I was referring to was Return Period, or recurrance interval, which leads to a lot of misinterpretations of what it actually means.

http://en.wikipedia.org/wiki/Return_period
http://www.nhc.noaa.gov/HAW2/english/ba ... turn.shtml

[Toadstool]

Thanks for the great replies Downdraft and Terstorm. I guess we all pretty much agree. It just annoyed me when I was watching something on the Science Channel last night when it said that NYC was overdue for a hurricane, and I couldn't put reason to their rhyme.

[Downdraft]

Hi, thanks, great discussion..what I was referring to was Return Period, or recurrance interval, which leads to a lot of misinterpretations of what it actually means.

http://en.wikipedia.org/wiki/Return_period
http://www.nhc.noaa.gov/HAW2/english/ba ... turn.shtml

As with any engineering study your right a Return Period has to be calculated. For instance you want to build a bridge over a particular span of geography. You calculate what the highest water level ever recorded there was and what is theoretically possible. Then you figure out how much money you want to spend based upon the risk analysis and proceed. We all do the same thing when we buy something like a reinforced garage door built to withstand hurricane winds. The door rated at 120 mph is cheaper than the one rated at 140 so you decide how much risk for how much benefit versus how much cost. Do I buy the $10.00 fire extinguisher or do I buy the $100.00 one? What's your house worth to you based upon how long it takes the fire department to get there.

I too get very irked by anyone that uses the "overdue" term as it relates to a hurricane making a landfall in a particular area. It has no scientific or statistical basis and in fact, is usually used by those that wish they would get hit, or the media to stir up hype.

[orion]

Good discussion everyone. One way I use when I teach Stats that seems to help the kids understand is this:

Probability has no memory... no history. If you toss a fair coin for the 5th time and already have 4 heads, the coin has no 'memory' of how it landed before and the probability is still 1/2.

Think about people who fly. If I was a first time flyer, would I want to get on a plane with a guy who has flown 100 times? Is he 'due' to crash because he has flown safely so many times? Of course the answer is no, since probablilty has no history.

~orion

[Deputy Van Halen]

Good discussion everyone. One way I use when I teach Stats that seems to help the kids understand is this:

Probability has no memory... no history. If you toss a fair coin for the 5th time and already have 4 heads, the coin has no 'memory' of how it landed before and the probability is still 1/2.

Think about people who fly. If I was a first time flyer, would I want to get on a plane with a guy who has flown 100 times? Is he 'due' to crash because he has flown safely so many times? Of course the answer is no, since probablilty has no history.

~orion

Actually, this is only true if it has been proven that the two events in question are independent of each other. I'm not sure that this is the case with hurricanes. There may be factors that carry over from one season to the next that can affect the landing probabilities. But if so, I'll bet that they would make locations that are supposedly "due" less likely to be hit now.

[Toadstool]

Good discussion everyone. One way I use when I teach Stats that seems to help the kids understand is this:

Probability has no memory... no history. If you toss a fair coin for the 5th time and already have 4 heads, the coin has no 'memory' of how it landed before and the probability is still 1/2.

Think about people who fly. If I was a first time flyer, would I want to get on a plane with a guy who has flown 100 times? Is he 'due' to crash because he has flown safely so many times? Of course the answer is no, since probablilty has no history.

~orion

Actually, this is only true if it has been proven that the two events in question are independent of each other. I'm not sure that this is the case with hurricanes. There may be factors that carry over from one season to the next that can affect the landing probabilities. But if so, I'll bet that they would make locations that are supposedly "due" less likely to be hit now.

I agree that there may be other factors which make the return rate regular, but for cities like NYC we don't have many data points so I don't think there is even a supposed link. If there are factors, it would be great to discover and quantify them and get those hurricane models working. TD4 is still going to Mexico, right?

[Downdraft]

Yeah it's going to Mexico they are overdue. Just kidding with ya!

[Toadstool]

Yeah it's going to Mexico they are overdue. Just kidding with ya!

lol, good stuff.

[terstorm1012]

This is a great thread and I'll be sending it around to various people who like to say "We're so overdue for [insert natural event here]."