Originally Posted by Vandal-72

It's like testing our coin's randomness by flipping it, keeping a running total and stopping the instant our heads percentage is 60 or 65%.

If this doesn't happen early on, it is very unlikley it ever will (unless the coin is biased). As the population gets larger the mean frequency will rapidly converge on 50%. There is little chance the population mean will rise above 60% after the first hundred. The propabilities fall rapidly: 1.8% for 100 flips, 0.2% at 200, 0.02% for 300 flips. By 500 its 4.6 per million. At 1000 flips it's 1 in 7 billion.

And if it does happen early on, the statistical tests will show you that the outcome is not very significant.

So yes, the if the first 10 flips of a coin comes up all heads or all tails, it is very likley biased (p < 0.002). There is only a 1 in 512 chance that if you keep flipping, the coin will start showing 50:50 on average, as an honest coin must do if you flip it long enough.

Originally Posted by Mikebert

If this doesn't happen early on, it is very unlikley it ever will (unless the coin is biased). As the population gets larger the mean frequency will rapidly converge on 50%. There is little chance the population mean will rise above 60% after the first hundred. The propabilities fall rapidly: 1.8% for 100 flips, 0.2% at 200, 0.02% for 300 flips. By 500 its 4.6 per million. At 1000 flips it's 1 in 7 billion.

Exactly! Your single four sigma storm equalling proof of warming is analogous to someone who only flips a coin ten times and accepts the 60% heads results as proof of bias. Further flips show the coin is not biased and further storm data collection may show that there is no warming effect! You can't declare bias (warming) early in the test nor based on a single storm.

And if it does happen early on, the statistical tests will show you that the outcome is not very significant.

So yes, the if the first 10 flips of a coin comes up all heads or all tails, it is very likley biased (p < 0.002).

That is only true if you stop flipping! If the subsequent 9,990 flips bring the average back to 50:50, then your declaration of bias is wrong.

You do realize that any particular ten flip sequence has exactly the same probability of happening as ten heads right?

There is only a 1 in 512 chance that if you keep flipping, the coin will start showing 50:50 on average, as an honest coin must do if you flip it long enough.

Part of this is my fault but you are mixing two different coin situations.

Within 10,000 flips, there will likely be a sequence of 10 heads in a row somewhere in the dataset. That's analogous to your four sigma storm. You want to declare the first four sigma storm you see as due to warming just like someone who gets the ten head sequence early in the dataset (it doesn't have to be the first ten) might be amazed. But within the entire dataset the sequence is not evidence for anything beyond normal probability and your first four sigma storm might just be normal probability within a larger complete set.

The only way to statistically support a claim of bias (warming effects) is to compare complete sets of data. No single storm is a set. And you absolutely can't stop the collecting once you get a four sigma storm! That's statistical fraud.

Originally Posted by Marx & Lennon

So boiling all this down to the essence, we can't know anything for sure until we know everything for sure.

Not even then. There is no such thing as knowing everything for sure in science. The whole point of statistics is that "for sure" doesn't really exist in nature.

This reminds me of the argument that Keynes made about classical economics, that long after the storm has passed and seas are once again calm, we can learn the solution to the problem. This is a reactionary mindset.

Misconstruing the point under discussion is not helpful.

Science is a discipline, but one of limited value if it can't be used to prevent disasters.

It can. But, all conclusions in science, all, are subject to some level of uncertainty and revivable given new data. If you are asking for 100% certainty then you'd be better off listening to Eric ramble on or join a religion. Just don't expect either of them to be right or even testable very often.

That means making decision based on incomplete data and risking potential error.

It can' t be used proactively and be any other way.

Incomplete data does not necessarily equal insufficient data.

Originally Posted by Vandal-72

Exactly! Your single four sigma storm equaling proof of warming is analogous to someone who only flips a coin ten times and accepts the 60% heads results as proof of bias.

No it not. 10 coins coming up 60% heads is an extremely likely occurrence and hardly meaningful.

That is only true if you stop flipping! If the subsequent 9,990 flips bring the average back to 50:50, then your declaration of bias is wrong.

That’s not gonna happen. What is >99% likely to happen is you would continue to roll a lot more heads than tails.

You do realize that any particular ten flip sequence has exactly the same probability of happening as ten heads right?

Of course I do. You don’t really have a grasp on this do you?

Within 10,000 flips, there will likely be a sequence of 15 heads in a row somewhere in the dataset. That's analogous to your four sigma storm. You want to declare the first four sigma storm you see as due to warming just like someone who gets the 15 head sequence early in the dataset

I made a one change to make the analogy more exact. 15 heads in a row has similar probability to a 4 sigma event.

Suppose you look at the first hundred flips and lo, a sequence of 15 heads appears. Sure it could be random luck. But what is the probability that the first 100 flips would have 15 heads in a row? About 1 in 385. Granted it could be luck, but the odds are against it. Yes you can continue to roll ad infinitum, but you can always do more experiments too. At some point you want to draw conclusions.

Originally Posted by Vandal-72

The only way to statistically support a claim of bias (warming effects) is to compare complete sets of data. No single storm is a set.

What do you mean by "complete set". There will always be more storms in the future and more experiments you could run. The point of statistical analysis is to draw valid conclusions about the universe of all storms or experimental results by analyzing a properly collected sample from this universe, not the entire universe.

I think part of the problem here is communication. Suppose the historical collection of strong storms shows a distribution of strengths that when transformed shows normality, an example of this would be a lognormal distribution. Suppose you establish two periods and analyze the strengths of all strong storms and global temperatures duirng each period. Say there are three 30 years of temperatures and over 100 storms.

You average the data and find that one period has the higher average temperature and higher average storm strength. You perform statstical analysis of the two populations and establish a statistically significant difference (p<0.05) between the two samples for both temperature and storm strength. What do you conclude from this?

You could conclude there is no evidence of warmer temperatures or stronger storms. After all you could wait another 30 years and collect another set of temperatures and storm strengths and do the test again. You you feel this is necessary? Why or why not?

Suppose you do perform another evaluation 30 years later. And this one also shows average temperatures and storm strengths that are higher from the previous 30 year period. And again the results are signficant at the p<0.05 level.

You could conclude there is no evidence of warmer temperatures or stronger storms. After all you could wait another 30 years and collect another set of temperatures and storm strengths and do the test a thrid time. You you feel this is necessary? Why or why not?

After all the fact that you had two successful demonstrations could have happened by chance, just as you could roll 5 heads in a row twice in succession. If the null case is valid, if did this enough times you will find many examples when the analysis would came out as it did.

So when would you draw a conclusion? After the first 30 years, 60 years, 90 years, later?

Originally Posted by Mikebert

No it not. 10 coins coming up 60% heads is an extremely likely occurrence and hardly meaningful.

What part of analogy is hard for you to accept? The point is that you are trying to accept a storm as proof of bias despite not looking at all of the context. It is statistical fraud.

That’s not gonna happen.

You don't know that. You can only say it is unlikely.

What is >99% likely to happen is you would continue to roll a lot more heads than tails.

And if you didn't? But, you with your four sigma storm aren't even going to do the honest thing and keep flipping. You are trying to declare a single event as sign of system wide bias and that is simply dishonest.

Of course I do.

You have to check these basics when faced with anyone who doesn't understand what Grubb's Test is actually used for and fails to understand how checking the incoming data as it is collected and declaring significance before it is all is fraud.

You don’t really have a grasp on this do you?

No I get it. Are you still trying to claim that you can conclude statistical significance from a single data point? Care to calculate what your degrees of freedom will be for such an analysis?

I made a one change to make the analogy more exact. 15 heads in a row has similar probability to a 4 sigma event.

Really? Storm wind speeds are normally distributed? You aren't just making stuff up?

Suppose you look at the first hundred flips and lo, a sequence of 15 heads appears. Sure it could be random luck. But what is the probability that the first 100 flips would have 15 heads in a row? About 1 in 385. Granted it could be luck, but the odds are against it. Yes you can continue to roll ad infinitum, but you can always do more experiments too.

No need for that. It is likely that several short tails sequences probably balanced out the heads sequence. How many 100 flip sequences that include a 15 head sequence end up with an average that is four standard deviations from 50%?

Am I just supposed to ignore your reduction of the sample size to 100 from 10,000? With measurements that include only one of two possible results, detecting bias requires a much, much larger sample size. To detect a coin that is only 1% biased would require over 250,000 flips to detect it statistically.

A 20% bias (heads = 60%) will not require that many but it will require more than 100 flips. It's actually 400 flips for a z-score of four sigma confidence.

At some point you want to draw conclusions.

And that's your problem. You want to draw your conclusion as soon as you get a single, low probability, result. That's called cherry picking and it is dishonest. You are purposefully searching through the storms for something that, while unlikely, is not impossible and instead of assessing whether the unlikely result is due to chance you declare bias. Given a very large sample size, "unlikely" events do happen with regular frequency. They aren't really unlikely at all. The only way to tell if the "unlikely" really was a special occurrence and not a routine result from many trials is to run your analysis on the entire sample size. Why is this so hard for you to grasp?

What do you mean by "complete set". There will always be more storms in the future and more experiments you could run.

Degrees of freedom. The results from a tiny sample size give you almost no confidence no matter how extreme a single event might have been (Grubb's Test properly used). And a sample size of one storm is just nonsensical. If you want to detect warming effects on storms, you state that you are going to look at the next five years of storms ( a set of data for which you could not possibly already know the results) and then do exactly that. In fact, you might even have to have more than five years of data depending on the amount of variance between and the total numbers of storms. You don't go in and say, "well 2013 had a monster storm, let's run an analysis on 2013 storms versus the past."

That's just plain dishonest and you should damn well know that! It's exactly the same dishonesty that deniers use when they declare that warming has stopped for the last 13, 14, or 15 years because they purposely start their "analysis" with 1997, the warmest year on record.

The point of statistical analysis is to draw valid conclusions about the universe of all storms or experimental results by analyzing a properly collected sample from this universe, not the entire universe.

A single storm is not a sample. Stopping your data collection after the first big storm is not properly collecting, either. You seem to speak the words but your four sigma storm hypothetical belies your lack of understanding of what they actually mean.

Originally Posted by Mikebert

I think part of the problem here is communication. Suppose the historical collection of strong storms shows a distribution of strengths that when transformed shows normality, an example of this would be a lognormal distribution. Suppose you establish two periods and analyze the strengths of all strong storms and global temperatures duirng each period. Say there are three 30 years of temperatures and over 100 storms.

You average the data and find that one period has the higher average temperature and higher average storm strength. You perform statstical analysis of the two populations and establish a statistically significant difference (p<0.05) between the two samples for both temperature and storm strength. What do you conclude from this?

You could conclude there is no evidence of warmer temperatures or stronger storms. After all you could wait another 30 years and collect another set of temperatures and storm strengths and do the test again. You you feel this is necessary? Why or why not?

Suppose you do perform another evaluation 30 years later. And this one also shows average temperatures and storm strengths that are higher from the previous 30 year period. And again the results are signficant at the p<0.05 level.

You could conclude there is no evidence of warmer temperatures or stronger storms. After all you could wait another 30 years and collect another set of temperatures and storm strengths and do the test a thrid time. You you feel this is necessary? Why or why not?

After all the fact that you had two successful demonstrations could have happened by chance, just as you could roll 5 heads in a row twice in succession. If the null case is valid, if did this enough times you will find many examples when the analysis would came out as it did.

So when would you draw a conclusion? After the first 30 years, 60 years, 90 years, later?

Moving the goalpost.

None of this has anything to do with your original claim that a SINGLE STORM is evidence of warming.

"So, if a category 1 or higher tropical storm with wind speed that was 4 sigmas above the mean wind speed for category 1 or higher tropical storms historically occurred, to explain the event as a simply a random outlier out of the normal population such storms could be ruled out as beyond a shadow of doubt (a scientist would not make that call, a policymaker or judge would). Far more likely is that a special cause was operative (here the scientist would agree). Since there is no reason to believe that higher surface temperatures cannot lead to stronger storms this would be a very likely candidate for the special cause."

Just noticed something really ridiculous, how exactly do you get a four sigma wind speed in a category one storm? Winds of that strength would make the storm a category three or four storm!

Originally Posted by Vandal-72

Moving the goalpost.

No it was part of an approach I would follow up on. Don't bother with it now, I think I have found a better way to express what I mean.

Just noticed something really ridiculous, how exactly do you get a four sigma wind speed in a category one storm? Winds of that strength would make the storm a category three or four storm!

You don't. But it says category 1 or higher.

After you posted about the coins, I think I see your objection.

The key is the following observation: If a 4 sigma storm happened in 1835, or 1910, it would be not be significant. But if the same storm happened in the next few decades it would be signifcant.

The reasoning is as follows. In the first case, there is nothing that a priori distinguishes one period from another, they are all the same. So the fact that a very rare event happened during one of many indistinguishable periods is unremarkable because since there are many periods available for it to happen, and it is going to happen in one of them.

However recent decades are not indistinguishable for the other periods. Scientists have reason to believe that humans have perturbed the climate in ways that might make very strong storms more likley. Since this perturbation is of recent vintage. Decades falling after when the potential for the perturbation effect was realized are different from all the decades before. We can test this notion by predicting that a super storms (storms that would easily break all records from before we started monioring) should happening more often and then watching to see if this is indeed the case.

This is like suspecting than a coin is heavily biased towards heads, undertaking an experiment in which the coin is flipped 10 times to see if a very unlikely percentage of heads is seen.

Let's say the belief of the potential for stronger storms began in the 1980's. In any case efforts to monitor storm strength by satellite only began then. Suppose a superstorm happens every 600 decades on average. What is the probably of such a storm happening in one of the ten decades after 1980? There are ten decades, so ten opportunities for this event of 1/600 probability to happen. The probability it will happen in one of them is 1/60. If one actually happens, it would reject the null hypothesis that strong storms are not more likley in the post-1980 period than before. We would conclude the hypothesis that human climate perturbation has made super storms more likely is supported at the >98% signficance level.

In this case, a single storm does establish the case as much as analysis of a group would.

Vandal, here is my initial thought process when I posted about the four sigma storm.

I assumed that the warming issue is going to be settled one way or another in the next 4-5 decades. Over this period there will be N storms that are candidates for superstorms. That is, N chances for a superstorm to happen. N is your sample size and it is much larger than one.

Assume a superstorm happens as one of the N future storms that will occur over the next 4-5 decades, and this is a signficant (p<0.05) event. The probability that any one of these N storms is a super storm is Np, where p is the probability that any one storm is a superstorm. If this is to be signficant Np < 0.05. Thus p < 0.05/N.

I guessed that over the next 4-5 decades there should be hundreds of candidate storms (which I thought of as "strong storms"). Assume an N of 500. This means p < 0.0001. I used the term "four sigma strom" to refer to a storm for which p is at least this small. Four sigma is a 1/31000 probability for which N could be as large as 600.

Thus a superstorm is defined as a storm with a strength that is expected to occur in about 1 in 30000 strong storms. A strong storm one strong enough so less than a dozen or so of them happen each year on average. This would give an N in the neighborhood of 500. The strength of strong storms would not be normally distributed. By transforming the data, you may find that the transformed values are normally distributed. For example if the transform is the natural log, and the transformed data is normally distributed this is called a lognormal distribution. Other transforms may be used that give other types of distributions. The statistician doing the analysis would make an appropriate choice. They would then analyze all the strong storms from before 2014 to establish the probability distribution for historical strong storms. This would be used to estimate the strength level that corresponds to a p small enough so Np < 0.05 (one would likely choose a provisional N larger than what one will end up with to be conservative).

Now the study would not be completed until 2060 or so. As you say the dice would continue to be rolled. And they would, I never meant to imply that one would stop assessing storm strength after a big storm happened. However the result is statistically signficant if any one of the N storms is a superstorm. As soon as a superstorm happens then this condition is met. Only if a superstorm does not happen and the null case is not rejected do you have to wait until 2060 to come to this conclusion.

Originally Posted by Mikebert

Vandal, here is my initial thought process when I posted about the four sigma storm.

I assumed that the warming issue is going to be settled one way or another in the next 4-5 decades. Over this period there will be N storms that are candidates for superstorms. That is, N chances for a superstorm to happen. N is your sample size and it is much larger than one.

Assume a superstorm happens as one of the N future storms that will occur over the next 4-5 decades, and this is a signficant (p<0.05) event. The probability that any one of these N storms is a super storm is Np, where p is the probability that any one storm is a superstorm. If this is to be signficant Np < 0.05. Thus p < 0.05/N.

I guessed that over the next 4-5 decades there should be hundreds of candidate storms (which I thought of as "strong storms"). Assume an N of 500. This means p < 0.0001. I used the term "four sigma strom" to refer to a storm for which p is at least this small. Four sigma is a 1/31000 probability for which N could be as large as 600.

Thus a superstorm is defined as a storm with a strength that is expected to occur in about 1 in 30000 strong storms.

So, rare but not impossible in a pre-warming world.

A strong storm one strong enough so less than a dozen or so of them happen each year on average. This would give an N in the neighborhood of 500. The strength of strong storms would not be normally distributed. By transforming the data, you may find that the transformed values are normally distributed. For example if the transform is the natural log, and the transformed data is normally distributed this is called a lognormal distribution. Other transforms may be used that give other types of distributions. The statistician doing the analysis would make an appropriate choice. They would then analyze all the strong storms from before 2014 to establish the probability distribution for historical strong storms. This would be used to estimate the strength level that corresponds to a p small enough so Np < 0.05 (one would likely choose a provisional N larger than what one will end up with to be conservative).

Now the study would not be completed until 2060 or so. As you say the dice would continue to be rolled. And they would, I never meant to imply that one would stop assessing storm strength after a big storm happened. However the result is statistically signficant if any one of the N storms is a superstorm. As soon as a superstorm happens then this condition is met. Only if a superstorm does not happen and the null case is not rejected do you have to wait until 2060 to come to this conclusion.

Alright, I'm following your reasoning better now. You still have a problem with declaring significance on a single event. Your defining a superstorm as a rare but still possible outcome of normal, non-warmed, storm generation. A statistical test for significance requires more than a single event. What if you get only one single superstorm and all the rest of the storms during the time period are weaker than the pre-2014 baseline? Can you still conclude that warming caused the superstorm when the rest of the data indicates that warming actually weakens storms? If however, your superstorm is mixed in with a higher than baseline occurrence of large storms, then it is the accumulated data of all the storms that gives your conclusion significance not the presence of a single storm.

If you are basing your four sigma storm strength on post-warming storm strengths then you can't use the presence of such a storm as a test for warming effects.

A single event can simply not be used as a statistical test sample.

Originally Posted by Mikebert

No it was part of an approach I would follow up on. Don't bother with it now, I think I have found a better way to express what I mean.

Fair enough.

You don't. But it says category 1 or higher.

Ok. I see what you are trying to say.

After you posted about the coins, I think I see your objection.

The key is the following observation: If a 4 sigma storm happened in 1835, or 1910, it would be not be significant. But if the same storm happened in the next few decades it would be signifcant.

The reasoning is as follows. In the first case, there is nothing that a priori distinguishes one period from another, they are all the same. So the fact that a very rare event happened during one of many indistinguishable periods is unremarkable because since there are many periods available for it to happen, and it is going to happen in one of them.

However recent decades are not indistinguishable for the other periods. Scientists have reason to believe that humans have perturbed the climate in ways that might make very strong storms more likley.

They can only suspect that this is possible. To actually "know" would require a statistical analysis of a large dataset of storm strength distributions pre and post-warming. Once that has been done, your superstorm test is redundant and you are in effect engaged in confirmation bias.

Since this perturbation is of recent vintage. Decades falling after when the potential for the perturbation effect was realized are different from all the decades before. We can test this notion by predicting that a super storms (storms that would easily break all records from before we started monioring) should happening more often and then watching to see if this is indeed the case.

A single, record event was and still is possible despite the perturbation. It is the context, all the post-warming storms, that allows us to conclude that warming had something to do with it. What if the record storm is a single anomaly embedded in decades of very, very weak storms? Does warming increase storm strength?

This is like suspecting than a coin is heavily biased towards heads, undertaking an experiment in which the coin is flipped 10 times to see if a very unlikely percentage of heads is seen.

Ten flips is no where near enough flips for any sort of conclusion to be drawn.

Let's say the belief of the potential for stronger storms began in the 1980's. In any case efforts to monitor storm strength by satellite only began then. Suppose a superstorm happens every 600 decades on average. What is the probably of such a storm happening in one of the ten decades after 1980? There are ten decades, so ten opportunities for this event of 1/600 probability to happen. The probability it will happen in one of them is 1/60. If one actually happens, it would reject the null hypothesis that strong storms are not more likley in the post-1980 period than before.

No it wouldn't. What if the next 800 decades see not a single superstorm? You are stopping your data collection when you get what you are looking for again. That's statistical fraud. If however, you got 4 superstorms in two decades, that set of data might be significantly different from the pre-1980 baseline. But, you must absolutely document more than one event.

The probability of a royal flush in stud poker is 1/649,740. Do you immediately call the dealer a cheater if your opponent gets that hand? What if they get two in a row? Which situation has actual statistical evidence of cheating?

We would conclude the hypothesis that human climate perturbation has made super storms more likely is supported at the >98% signficance level.

No. More likely can not be deduced from a single event. More likely implies that it occurs more often than before. "More" requires an excess of one.

Your degrees of freedom will be n-1 or n-2 depending on what statistical method you use. A single result would mean you must divide by zero in your test, undefined.

In this case, a single storm does establish the case as much as analysis of a group would.

Simply not possible. A single event is not a pattern. Statistics is analysis of patterns.

The following is written by a leading environmentalist who has always supported Mr. Obama. But it appears that he is awakening to the fact of how mere words can be deceiving.

Obama and Climate Change: The Real Story


The president has said the right things about climate change – and has taken some positive steps. But we're drilling for more oil and digging up more carbon than ever

by Bill McKibben

Two years ago, on a gorgeous November day, 12,000 activists surrounded the White House to protest the proposed Keystone XL pipeline. Signs we carried featured quotes from Barack Obama in 2008: "Time to end the tyranny of oil"; "In my administration, the rise of the oceans will begin to slow."

Our hope was that we could inspire him to keep those promises. Even then, there were plenty of cynics who said Obama and his insiders were too closely tied to the fossil-fuel industry to take climate change seriously. But in the two years since, it's looked more and more like they were right – that in our hope for action we were willing ourselves to overlook the black-and-white proof of how he really feels.


If you want to understand how people will remember the Obama climate legacy, a few facts tell the tale: By the time Obama leaves office, the U.S. will pass Saudi Arabia as the planet's biggest oil producer and Russia as the world's biggest producer of oil and gas combined. In the same years, even as we've begun to burn less coal at home, our coal exports have climbed to record highs. We are, despite slight declines in our domestic emissions, a global-warming machine: At the moment when physics tell us we should be jamming on the carbon brakes, America is revving the engine.


You could argue that private industry, not the White House, has driven that boom, and in part you'd be right. But that's not what Obama himself would say. Here's Obama speaking in Cushing, Oklahoma, last year, in a speech that historians will quote many generations hence. It is to energy what Mitt Romney's secretly taped talk about the 47 percent was to inequality. Except that Obama was out in public, boasting for all the world to hear:


"Over the last three years, I've directed my administration to open up millions of acres for gas and oil exploration across 23 different states. We're opening up more than 75 percent of our potential oil resources offshore. We've quad*rupled the number of operating rigs to a record high. We've added enough new oil and gas pipeline to encircle the Earth, and then some. . . . In fact, the problem . . . is that we're actually producing so much oil and gas . . . that we don't have enough pipeline capacity to transport all of it where it needs to go."

Bottom line

It has: Despite brave opposition from groups like Tar Sands Blockade, Keystone South is now 95 percent complete, and the administration is in court seeking to beat back the last challenges from landowners along the way. The president went ahead and got it done. If only he'd apply that kind of muscle to stopping climate change.

http://www.commondreams.org/view/2013/12/17-8

Yes, we're drilling for more oil etc and will be until it runs out. This is the tiger than every president since Jimmy Carter is either riding or is unaware of.

That is, American public keeps demanding more and more energy so that we can not only maintain our way of life, but keep on improving it, and God help the administration that tells us we need to cut back. He or she would be out of office in an instant, to a chorus of boos, hisses, and accusations of hating freedom, prosperity, and security. Obama knows it. Everyone above the level of village dogcatcher knows it.

The only time a president can offer "blood, sweat, and tears" and get away with it is in a war in which we're visibly and in-your-face fighting for our lives. The War on Terror has not risen to that level, pray it never will. And any warning that we face real limitations ahead and had best prepare for them is seen as not only pessimistic, but as near-treason.

In short, his hands are tied here, as would be anyone's - except those who truly believe it's gung-ho and full speed ahead and that we are truly producing this oil rather than using it up at a high rate of speed. Their hands are not tied because they're with the program all the way!

Originally Posted by The Grey Badger

Yes, we're drilling for more oil etc and will be until it runs out. This is the tiger than every president since Jimmy Carter is either riding or is unaware of.

That is, American public keeps demanding more and more energy so that we can not only maintain our way of life, but keep on improving it, and God help the administration that tells us we need to cut back. He or she would be out of office in an instant, to a chorus of boos, hisses, and accusations of hating freedom, prosperity, and security. Obama knows it. Everyone above the level of village dogcatcher knows it.

The only time a president can offer "blood, sweat, and tears" and get away with it is in a war in which we're visibly and in-your-face fighting for our lives. The War on Terror has not risen to that level, pray it never will. And any warning that we face real limitations ahead and had best prepare for them is seen as not only pessimistic, but as near-treason.

In short, his hands are tied here, as would be anyone's - except those who truly believe it's gung-ho and full speed ahead and that we are truly producing this oil rather than using it up at a high rate of speed. Their hands are not tied because they're with the program all the way!

The only nitpick I'd have with the above is that you might be improperly disparaging the village dogcatcher.

But this is why I so often reprise the notion of values, that a stubborn clinging to one's existing way of life drives many crises. With 20 20 hindsight, bedrock principles of civilization like the divine right of kings, slavery and so many others have been recognized as living past their time by some, but too many will cling to that which has always been. Global Warming is just this crisis's issue, or one of them.

It is easy to despise the bad conservatives who clung to the past in prior crises, the people who never got a chance to write history books. How does one understand them? Try accepting that we are equally clinging to past immoral and disastrous life styles in this time. We are being human.

In the Industrial Age, crisis wars have required around four years for the winner to fully mobilize, figure out the tactics of the current generation of weapons, and commit to the level of savagery required to make it clear that utter devastation is necessary and inevitable if culture change is to be forced. It has been ugly, but victory against mere human opponents has been possible.

If we delay a global warming crisis until the immediate future's threat is clearly greater than the benefits provided by maintaining the old values, it will be too late. The damage will already have been done. Mother nature carries an awful lot of 'momentum'.

Yet I'm not seeing significant evidence of a spiral of rhetoric (let alone a spiral of violence) leading up to confrontation on environmental issues.

The Badger has the right of it.

Yes, I owe the dogcatcher an apology.

In fact, if anyone is aware that we're living in a contracting economy, the dogcatcher will be, as he tries to catch the Jones' abandoned pooch - the family moved away and couldn't bring the dog, so ...

I've acquired all three cats more or less that way.

Originally Posted by The Grey Badger

Yes, we're drilling for more oil etc and will be until it runs out. This is the tiger than every president since Jimmy Carter is either riding or is unaware of.

That is, American public keeps demanding more and more energy so that we can not only maintain our way of life, but keep on improving it, and God help the administration that tells us we need to cut back. He or she would be out of office in an instant, to a chorus of boos, hisses, and accusations of hating freedom, prosperity, and security. Obama knows it. Everyone above the level of village dogcatcher knows it.

The only time a president can offer "blood, sweat, and tears" and get away with it is in a war in which we're visibly and in-your-face fighting for our lives. The War on Terror has not risen to that level, pray it never will. And any warning that we face real limitations ahead and had best prepare for them is seen as not only pessimistic, but as near-treason.

In short, his hands are tied here, as would be anyone's - except those who truly believe it's gung-ho and full speed ahead and that we are truly producing this oil rather than using it up at a high rate of speed. Their hands are not tied because they're with the program all the way!

This is why truly great leaders had the courage to go against the grain. Sure, they put themselves in the margins but were brave enough to withstand the boos for the sake of humanity. What many of us don't understand is that sacrifice for the good of the whole is what eventually saves us.

This current Gilded age needs bravery, not conformity to whatever is popular and safe. Without courageous leaders who are willing to lead with abandon, the world will go further into the hell of a Global Warming with the permission of the cowardly.

Originally Posted by Deb C

This is why truly great leaders had the courage to go against the grain. Sure, they put themselves in the margins but were brave enough to withstand the boos for the sake of humanity. What many of us don't understand is that sacrifice for the good of the whole is what eventually saves us.

This current Gilded age needs bravery, not conformity to whatever is popular and safe. Without courageous leaders who are willing to lead with abandon, the world will go further into the hell of a Global Warming with the permission of the cowardly.

It's not just the boos. Obama is getting those for the things he's done that in an earlier age would have passed a lot more easily. It's the prospect of being impeached - or even lynched, or called a crackpot, and the leadership goes then to someone who is "with the program" of drill and burn, drill and burn, rinse and repeat, until the end.

You can't even talk about some of this stuff with the fashionably ecologically correct! "La-la-LA, I don't HEAR you! Poor old Mom, she's losing it."

Originally Posted by The Grey Badger

You can't even talk about some of this stuff with the fashionably ecologically correct! "La-la-LA, I don't HEAR you! Poor old Mom, she's losing it."

I suppose the same could be said for those who refuse to see that their idol made campaign promises that he never intended to keep. And, that this kind of denial is a gross codependency that enables our politicians to keep abusing the citizens. So in that respect, they become a big part of the problem. Our future generations thank them.

Originally Posted by Vandal-72

If you are basing your four sigma storm strength on post-warming storm strengths then you can't use the presence of such a storm as a test for warming effects.

Well of course not. The null hypothesis (the one I am trying to reject) is that there is no effect of warming. This means the post-warming storms and the pre-warming storms belong in the same population and so have the same distribution. So the four-sigma strength is based on pre-warming strengths.

Originally Posted by Vandal-72

No it wouldn't. What if the next 800 decades see not a single superstorm?

Support you did the group analysis you prefer and got a significant result. Then the next six groups show no significant result. This can happen too. This is why a p value is assigned to it.

The probability of a royal flush in stud poker is 1/649,740. Do you immediately call the dealer a cheater if your opponent gets that hand? What if they get two in a row? Which situation has actual statistical evidence of cheating?

No. But what if the recipient of the hand had announced immediately before that he was going to get a royal flush?

Your degrees of freedom will be n-1 or n-2 depending on what statistical method you use.

You wrote “A 20% bias (heads = 60%) will not require that many but it will require more than 100 flips. It's actually 400 flips for a z-score of four sigma confidence.”

What did you use for degrees of freedom in this calculation? I suspect none. From Wiki:

Originally Posted by wiki

Statistics is closely related to probability theory, with which it is often grouped. The difference is, roughly, that probability theory starts from the given parameters of a total population to deduce probabilities that pertain to samples. Statistical inference, however, moves in the opposite direction—inductively inferring from samples to the parameters of a larger or total population

In the example above you are doing probability analysis, the first bolded analysis. Degrees of freedom do not apply because you know the parameters of the coin flip population ahead of time. Degrees of freedom arise when you are inferring population parameters from a sample.

The 4 sigma analysis is also a probability analysis. The null hypothesis (the one being tested) holds that the perturbation has no impact. Therefore storms afterward fall into the same population as those before. The "before" population can be analyzed with N-1 degrees of freedom (N = # historical storms) to estimate its parameters. Given the parameters, the probability of the super storm occurring can then be estimated. If the probability is low then the null hypothesis is not supported. That is, it is unlikely that the “before” distribution is valid for the “after” situation.

Now in order to do this you do need a large sample of “before” data in order to have a large degree of freedom) so that the sample parameters closely approximate the population parameters.

In many cases you do not have a pre-existing reference population and so have to estimate the parameters for the reference population from a small sample of control experiments. But in this case, as in quality control, you have a good idea of reference population parameters. If you don’t, then you cannot define a four sigma event (remember sigma the population standard deviation, not s the sample standard deviation). The assumption made by the term “four sigma storm” assumes that one can do this. In reality it might not be the case. But you made an assertion that for all reference populations (even ones which are exactly known like coins or cards) one can never infer valid probabilities for single events occurring over a pre-specified period of time.

I am saying that is not true.

Statistics is analysis of patterns.

Statistics is the mathematical body of science that pertains to the collection, analysis, interpretation or explanation, and presentation of data. Significance testing is one aspect of statistical analysis that estimates the probability of outcomes. It is most closely akin to probability analysis.

The article speaks for itself. I guess it takes a brave woman like Warren to do what some other politicians are just too cowardly to do.


Elizabeth Warren Comes Down Hard Against Global Warming, Separates Herself From Hillary Clinton on Climate Change

TransCanada Corporation wants to build the Keystone XL Pipeline to carry oil from Alberta Canada's tar sands to two refineries owned by Koch Industries near the Texas Gulf Coast, for export to Europe; and Hillary Clinton has helped to make that happen, but Elizabeth Warren has now taken the opposite side.

Secretary of State Clinton, whose friend and former staffer Paul Elliot is a lobbyist for TransCanada, had worked behind the scenes to ease the way for commercial exploitation of this, the world's highest-carbon-emitting oil, 53% of which oil is owned by America's Koch Brothers . (Koch Industries owns 63% of the tar sands, and the Koch brothers own 86% of Koch Industries; Elaine Marshall, who is the widow of the son of the deceased Koch partner J. Howard Marshall, owns the remaining 14% of Koch Industries.)


David Goldwyn, who "served as Secretary of State Hillary Clinton's Special Envoy and Coordinator for International Energy Affairs," is yet another lobbyist for TransCanada. So, TransCanada has two of Hillary's friends working for them. Misters Elliot and Goldwyn thus worked intimately with Hillary's people to guide them on selecting a petroleum industry contractor (not an environmental firm, much less any governmental agency) to prepare the required environmental impact statement for this proposed pipeline.


Hillary Clinton as the Secretary of State had already displayed a record of carrying out the policies that were being promoted by her lobbyist friends, when she did everything possible, early in President Obama's first term, to support U.S. funding for the fascist junta in Honduras that perpetrated a coup d'etat on 28 June 2009 overthrowing that nation's progressive democratically elected President, and who then installed their own regime, and promptly placed their country into a continuing violent terror that caused Honduras ever since to be the nation with the highest murder rate in the world. Hillary's lobbyist friend in that particular matter was Lanny Davis, who also is an occasional Fox News contributor.

Bottom line:

This could be a turning-point in Warren's political career. She's no longer at war against only the financial industry corruption that dominates the conservative, Clinton and Obama, establishment within the Democratic Party (and all of the Republican Party) , but she is also at war against their environmental corruption. For yet another example of that corruption: On 2 October 2013, Joe Romm at Think Progress headlined "More Bad News For Fracking: IPCC Warns Methane Traps Much More Heat,"and he reported that, "The Intergovernmental Panel on Climate Change (IPCC) reports that methane ... is far more potent a greenhouse gas" than previously known, so bad it "would gut the climate benefits of switching from coal." And then, just five days after that, Jon Campbell in upstate New York headlined "In Oneida County, Hillary Clinton Touts U.S. Oil-and-Gas Production," and he reported that at Hamilton College, Hillary Clinton praised fracking for methane, by saying, "What that means for viable manufacturing and industrialization in this country is enormous." However, if Warren won't be able to get either Wall Street or the oil patch to finance her political campaigns, then how can she even possibly rise within the power-structure?

Entire article: http://www.commondreams.org/view/201...OlV04.facebook

Originally Posted by Deb C

The article speaks for itself. I guess it takes a brave woman like Warren to do what some other politicians are just too cowardly to do.

I seem to recall there was a time when most of Britain didn't want to fight another world war. During that time Winston Churchill was a very unpopular man. Somehow, though, about the time of Dunkirk, he became the only possible choice.

Welcome, Elizabeth, to the Wilderness Years.

Originally Posted by B Butler

I seem to recall there was a time when most of Britain didn't want to fight another world war. During that time Winston Churchill was a very unpopular man. Somehow, though, about the time of Dunkirk, he became the only possible choice.

Welcome, Elizabeth, to the Wilderness Years.

I hope she's as quotable.

Originally Posted by Bad Dog

I hope she's as quotable.

Unlikely.

Originally Posted by B Butler

Unlikely.

Just curious as to why you think she is unlikely to be quotable. Is it because she , as you indicate, will be exiled from her position like Churchill?