top 5 worst and best Miami neighborhoods (Hialeah, Miami Beach: versus, shop, statistics)

[Cougar Beach]

Originally Posted by Tungsten_Udder

The point is exactly that those things are NOT equal. Most crime victims are not "random" people.

I don't know what you know of statistical analysis, but some things can be held equal. If mugging or car break-in are what you're worried about, there is some probability that a particular person will be victimized in a particular spot in South Beach, in a particular spot in Liberty City, and in a particular spot in downtown Boston.

I was just looking at crime stats compiled by the Boston Police Dept., and they were broken down by district and area, covering homicide, rape, robbery, aggravated assault, burglary, larceny, and vehicle theft. Similar statistics for Miami-Dade are available here: Miami-Dade County - Miami-Dade Police

I understand that things like murder and embezzlement are usually perpetrated by people known to the victim, but it is pretty clear that the question is about random crimes. Are you implying that there's no correlation between aggregate robbery statistics, and a random person's chance of getting robbed?

[Tungsten_Udder]

Originally Posted by Cougar Beach

If mugging or car break-in are what you're worried about, there is some probability that a particular person will be victimized in a particular spot in South Beach, in a particular spot in Liberty City, and in a particular spot in downtown Boston.

That probability is in no way reflected by the stats, because it depends on factors not in the stats.

I understand that things like murder and embezzlement are usually perpetrated by people known to the victim, but it is pretty clear that the question is about random crimes. Are you implying that there's no correlation between aggregate robbery statistics, and a random person's chance of getting robbed?

No. I'm not trying to imply anything, but state what I've stated. I'll state it again: the statistics do not give you the data you'd need to arrive at a probability that you will be a victim.

I have other issues with it regarding probability in general, but to get into all of that in any detail would require a book.

[Cougar Beach]

Originally Posted by Tungsten_Udder

That probability is in no way reflected by the stats, because it depends on factors not in the stats. No. I'm not trying to imply anything, but state what I've stated. I'll state it again: the statistics do not give you the data you'd need to arrive at a probability that you will be a victim.

I have other issues with it regarding probability in general, but to get into all of that in any detail would require a book.

Are you saying that statics are useless unless they take into account every detail of every factor? That is what is required to obtain a precise probability for a given situation.

Do you think that the reliability rating of a car (e.g., average number of repairs per thousand miles, at a set of cost ranges) is completely useless?

You're basically saying that no real world statistics convey useful information. That is not correct. You are saying that there is no correlation whatsoever between the crime statistics of a locality, and your chance of being a victim of crime in that locality. You are saying that a statistical relationship with a correlation coefficient of less than 1.0 conveys no useful information. Once again, that is not correct.

Think about it. Do you really think that on the average, parking your car overnight in an area with a very high car theft rate versus one with a very low theft rate, no valid inference could be drawn when one of the locales is chosen, without knowing the car model, the precise angle it's parked from the curb, etc.? If you think that, then you must think that insurance companies that have different rates for different areas are committing a fraud. If you can make your case, you'll be able to retire for several lifetimes on the money you'd get from the class action lawsuit you could file.

[Tungsten_Udder]

Originally Posted by Cougar Beach

Are you saying that statics are useless unless they take into account every detail of every factor?

Why do you keep asking if I'm saying something other than what I'm saying?

The statistics tell you how many people were murdered, raped, robbed, etc. in an area.

What they do not tell you is how likely it is that you'll be murdered, raped, robbed, etc. in that area. They wouldn't have to list every possible variable to tell you that (as if "every enumerable variable" were really a tenable concept anyway), but they'd have to present more than what's presented--and they'd have to present variables that would be practically very difficult to obtain data for.

Do you think that the reliability rating of a car (e.g., average number of repairs per thousand miles, at a set of cost ranges) is completely useless?

No, and I think that instantiations of a type of car are far more similar than people are.

You're basically saying that no real world statistics convey useful information.

No. That's not at all what I'm saying. That's a straw man to avoid looking at what I'm saying.

You are saying that there is no correlation whatsoever between the crime statistics of a locality, and your chance of being a victim of crime in that locality.

I already denied that above.

Do you really think that on the average, parking your car overnight in an area with a very high car theft rate versus one with a very low theft rate, no valid inference could be drawn when one of the locales is chosen, without knowing the car model, the precise angle it's parked from the curb, etc.?

It depends on just what you're saying that we do know. If it's only the car theft rate, then yes, I think we couldn't determine a cogent probability that one's car would be stolen in that area.

If you think that, then you must think that insurance companies that have different rates for different areas are committing a fraud.

Non sequitur.

[Cougar Beach]

You seem to think that I'm not aware of the information on this page, the FBI's disclaimer on ranking localities based on the crime statistics they collect; I am in fact aware of them: Variables Affecting Crime - Crime in the United States 2007

> What they do not tell you is how likely it is that you'll be murdered, raped, robbed, etc. in that area.

They don't tell you precisely how likely you will be a victim. They tell you how likely, on the average, you would have been a victim during the period that the statistics were collected. You seem to be saying that they won't even give you a gross idea, because it's all too complicated.

The statistics have a predictive quality, so there is a correlation (a word I keep emphasizing) between what the past statistics where, end what the current statistics will end up to be. Of course there can be non-linearities if a large geographical of social change occurs between sampling periods (e.g., a high wall is built, separating neighborhoods, or a natural disaster occurs, requiring the evacuation of a large segment of the population). That addresses the temporal issue.

> No, and I think that instantiations of a type of car are far more similar than people are.

Okay, so the utility (i.e., predictive accuracy) of car repair stats is greater than that of crime stats, where the crime stats address the likelihood that you will be a crime victim. That doesn't mean that the utility is zero.

> Non sequitur.

It absolutely does follow. If you think that car accident, vandalism, and theft statistics based on geography have no useful predictive value, then the variations in insurance rates that are based on it are invalid, and thus fraudulent. You can take into account a person's driving record, value of his car, etc., but if those things remain constant (i.e., if you get insurance quotes for you and your car in various areas), you are still subject to the effect of the general case in those areas. That is what is meant by "all else being equal". There could be any number of factors.

In other words, for you and your car, you will pay more for the same level of insurance in Miami Beach than you will in Concord, New Hampshire. That is because there is a higher probability of theft, vandalism, accidents (with a relatively high likelihood that the other party won't have valid insurance) in Miami then in Concord.

So it is for crime statistics. You are basically the same person you are in Amherst, NH as you are in Opa-Locka, FL, but it is more likely that you be a victim of a crime in Opa-Locka than in Amherst. If you walk around in one place covered with expensive jewelry, but don't in the other place, of course you will affect the probability. But if you dress the same in both places, the statistics obtain. That is the point.

I hope you now understand what is meant by "all else being equal". The point is tht the "all else" things average out as the number of statistical samples are increased. I also hope that you understand that if there is a .002% incidence of robbery in an area, that it doesn't mean that the chance that you will be robbed is .002%. It does mean that you are more likely to be robbed there than in an area where the incidence of robbery is .001%. In fact, it's roughly twice as likely.

I've explained it every way I can possibly think of. If you still disagree, then we will have to agree to disagree, unless a true expert in statistics shows up in this thread.

[Tungsten_Udder]

Originally Posted by Cougar Beach

They don't tell you precisely how likely you will be a victim. They tell you how likely, on the average, you would have been a victim during the period that the statistics were collected.

No they don't.

You seem to be saying that they won't even give you a gross idea, because it's all too complicated.

Right, which is different than there being absolutely no correlation between crime statistics and probabilities of some people being victims.

The statistics have a predictive quality, so there is a correlation (a word I keep emphasizing) between what the past statistics where, end what the current statistics will end up to be.

I have no problem with that. However, that's different than the statistics telling you how likely it is that you will be a victim. Because the victims are not mostly random.

where the crime stats address the likelihood that you will be a crime victim.

I'm not agreeing that the crime stats address the likelihood that you will be a victim.

It absolutely does follow.

No, it doesn't, because you can think that fraud doesn't hinge on the reasons that auto insurers choose to have different rates for different areas.

insurance rates that are based on it are invalid, and thus fraudulent.

"If x is invalid, then x is fraudulent" <---that statement is one I assign "F" to. There are more problems with that than that statement being false in my opinion, but that's sufficient. It's a non-sequitur.

So it is for crime statistics. You are basically the same person you are in Amherst, NH as you are in Opa-Locka, FL, but it is more likely that you be a victim of a crime in Opa-Locka than in Amherst.

Again, I do not agree that that's the case. It's not that I'm unfamiliar with the standard views on this, it's that I think they're wrong. We're just scratching the surface why I think they're wrong, but that's partially because I'm not about to write a book for you on a message board.

But if you dress the same in both places, the statistics obtain. That is the point.

I do not agree with that. That's not a sufficient factor for determining what you want to determine.

It does mean that you are more likely to be robbed there than in an area where the incidence of robbery is .001%. In fact, it's roughly twice as likely.

That's the standard view. I think the standard view is wrong. I'm not going to get you to agree with me about that on a message board.

I've explained it every way I can possibly think of. If you still disagree, then we will have to agree to disagree,

We should have done that long ago.

unless a true expert in statistics shows up in this thread.

I'm an expert in philosophy, including philosophy of science. I also did a lot of logic and philosophy of mathematics (as those were initially the second and third fields I was pursuing as areas of specialty, but I decided on other fields (ontology and aesthetics) for that eventually). Those are some of the things that I believe are relevant for this (and actually, I believe that my expertise is more relevant for my points than a statistician's expertise would be, because some of this hinges on philosophical points about what statistics does and doesn't tell us), but again, I'm not going to write a book for you on a message board. It's a waste of my time, in my opinion, because like usual for these kinds of discussions, you're reading the fact that I have a different view as being indicative of not understanding the received view, so you're explaining the recieved view in a "But this goes to 11" way, and that's not likely to change.

If you want to go a bit more in detail with this, by the way, I'd do it with you in a chat setting instead. I don't like doing it in a message board setting because it's too easy for parties to ignore various points, gloss over things, etc. in a message board setting . . . and meanwhile the amount of stuff we have to type grows exponentially and eats up a lot of time that I could use to do something far more productive. Just let me know if you'd like to chat about some of this stuff instead. I usually use IRC for that.

[EnjoyTheSilence]

Hialeah = worst.

Hialeah isn't dangerous per se, but you'll come across the worst human scum, so it's best to avoid the place like the plague!

Hialeah = Hellaleah. The murder rate isn't bad, but the people are....it's hell. I challenge you to spend a night in Hellaleah, with the wonderful residents. It's a special place, unlike anywhere else in the United States. While you may not get killed, you'll be sure to encounter true human scum in Hellaleah, whether it be the mainac drivers or just the people in general.

[Cougar Beach]

You gainsay every one of my statements, you define what I say as the "standard view", repeat several times that the standard view is wrong, put it down by calling it the "received view" (as if it's just based on faith), and yet....

Originally Posted by Tungsten_Udder

We're just scratching the surface why I think they're wrong, but that's partially because I'm not about to write a book for you on a message board.

Originally Posted by Tungsten_Udder

I don't like doing it in a message board setting

...can't or won't articulate your reasoning in a couple of paragraphs, all in a lengthy post.

Originally Posted by Tungsten_Udder

I believe that my expertise is more relevant for my points than a statistician's expertise would be

Sorry, not buying it. You are a philosophy expert (whatever that means). I'm an engineer. I don't think we are speaking the same language. I'd rather you point me to a book or other publication that articulates what you are trying to convey. Do we start with Plato?

What we're discussing here is really not a bizarre statistical situation. Perhaps there is a good reason for the "standard view". Epistemologically, I'm not sure that we can say anything about anything.

[EnjoyTheSilence]

Originally Posted by Cougar Beach

You gainsay every one of my statements, you define what I say as the "standard view", repeat several times that the standard view is wrong, put it down by calling it the "received view" (as if it's just based on faith), and yet....





...can't or won't articulate your reasoning in a couple of paragraphs, all in a lengthy post.



Sorry, not buying it. You are a philosophy expert (whatever that means). I'm an engineer. I don't think we are speaking the same language. I'd rather you point me to a book or other publication that articulates what you are trying to convey. Do we start with Plato?

What we're discussing here is really not a bizarre statistical situation. Perhaps there is a good reason for the "standard view". Epistemologically, I'm not sure that we can say anything about anything.

Don't take anything 'Tungsten_Udder' says seriously. He is a snowbird who clings to his faux paradise like a fly clings to a piece of dog s**t on the sidewalk.

[Cougar Beach]

Originally Posted by BCreass

Don't take anything 'Tungsten_Udder' says seriously. He is a snowbird who clings to his faux paradise like a fly clings to a piece of dog s**t on the sidewalk.

I don't know much of anything about him personally, and it's irrelevant anyway. The issue under discussion is not one of meaning and existence, and can be discussed objectively.