A Fraud is Coming to a Close

Something which has long bothered me about the global warming debate is how "skeptics" are so quick to cry, "Fraud!" about... well, practically anything. I discussed this recently where an organization made a list of hundreds of pieces of work they took (partial or full) credit for as part of applying for a grant. The list included over 500 items, and it turns out approximately 25 of those items should not have been included. "Skeptics" yelled and screamed about how this was criminal fraud that should be prosecuted.

That's nonsense of course. Nobody was able to show any evidence the inclusion of those extra items was done with the intent to mislead as opposed to having been a simple mistake. Nobody was able to show the inclusion of a small number of extra items in one document submitted along with an application could have had any effect on whether or not the grant was awarded. In other words, nobody was able to show this was anything more than an embarassing mistake.

At the same time, these same "skeptics" are happy to either overlook, promote or even defend criminal fraud when it suits their purposes. I'd like to discuss that today because I find it offensive these "skeptics" have robbed me of $100,000.

To be up front, there was never any chance I would actually get $100,000. That's the fraud. You see, last year a man named Douglas Keenan created a web page announcing a contest in which people could win $100,000. I discovered this contest due to a post at the blog Bishop Hill. The announcement said:

There have been many claims of observational evidence for global-warming alarmism. I have argued that all such claims rely on invalid statistical analyses. Some people, though, have asserted that the analyses are valid. Those people assert, in particular, that they can determine, via statistical analysis, whether global temperatures are increasing more that would be reasonably expected by random natural variation. Those people do not present any counter to my argument, but they make their assertions anyway.

In response to that, I am sponsoring a contest: the prize is $100 000. In essence, the prize will be awared to anyone who can demonstrate, via statistical analysis, that the increase in global temperatures is probably not due to random natural variation.

The file Series1000.txt contains 1000 time series. Each series has length 135 (about the same as that of the most commonly studied series of global temperatures). The series were generated via trendless statistical models fit for global temperatures. Some series then had a trend added to them. Each trend averaged 1°C/century—which is greater than the trend claimed for global temperatures. Some trends were positive; others were negative.

A prize of $100 000 (one hundred thousand U.S. dollars) will be awarded to the first person, or group of people, who correctly identifies at least 900 series: i.e. which series were generated by a trendless process and which were generated by a trending process.

Each entry in the contest must be accompanied by a payment of $10; this is being done to inhibit non-serious entries. The contest closes at the end of 30 November 2016.

The file Answers1000.txt identifies which series were generated by a trendless process and which by a trending process. The file is encrypted. The encryption key and method will be made available when someone submits a prize-winning answer or, if no prize-winning answers are submitted, when the contest closes.

Note the lack of any comment about how people can know this is for real. $100,000 is a lot of money. If somebody was actually willing to give it away in a contest, it is unlikely they'd just make a webpage and say so. It is far more likely they would deposit the money in an escrow account and create legal documentation of the terms of the contest and rules under which the money would be awarded to someone.

That didn't happen here. All that happened here is some guy on the internet said he'd pay people $100,000 if they could win a contest he created. Just some guy on the internet. That's all. I can make a webpage where I say, "I will pay $100,000 to anyone who hangs a 100 foot Led Zepplin banner on the Eiffel Tower. I doubt anyone would believe me. Yet, "skeptics" believe this guy... just because?

It gets worse though. Look at that announcement. Tell me, if you wanted to enter into this contest, how would you? You wouldn't. You couldn't. There's no way to. Douglas Keenan actually announced a "$100,000 contest!!1!" in such a sloppy manner there was no way to enter it.

After I pointed this out, the webpage was updated to instruct users to contact Keenan at a particular e-mail address with their entries. However, remember the announcement said:

Each entry in the contest must be accompanied by a payment of $10; this is being done to inhibit non-serious entries.

This is what the web page was updated to say:

Contest entries should be emailed to me (doug dot keenan at informath.org). Each entry in the contest must be accompanied by a payment of $10; this is being done to inhibit non-serious entries.

There is no explanation of how a person would submit their entry fee. That is, despite Keenan updating his webpage because he had failed to provide information on how to enter his $100,000 contest, he still failed to provide enough information for anyone to enter his contest. I contacted Keenan about this, writing:

I had been considering submitting an entry for your contest, but originally I couldn't find any information about how to enter. Now that you've posted your e-mail address, I still cannot see how anyone is supposed to submit the entry fee you say must accompany their entry.

Could you clarify how people are supposed to participate in your contest given there is currently no way to submit an entry under the stated rules?

He chose not to respond to my e-mail. I received a request for $10 via Paypal some time after I sent the e-mail, but that doesn't answer my question. I couldn't even know it was from him. Anyone could have made a Paypal account and listed their name as "Douglas Keenan" to try to scam people. These sort of things led me to write a blog post saying:

But that brings us back to the real problem of all this, the $100,000. That is such an absurd number. I don't know how much money Douglas Keenan has. Maybe he really does have $100,000 he can just throw away. I don't know. What I do know is the idea people can be sure he'd give it away just because he says so is absurd.

The reality is if somebody wins this competition, Keenan doesn't have to give them the prize money. If he had put the money in an escrow account and set up terms of payment declaring the money would go to the winner of this contest (if any), then yes, we'd have reason to believe winning the contest would actually result in winning the money. But just because Keenan says so?

Seriously?!

I don't know Keenan. Maybe he's a standup guy. Maybe he really would pay $100,000 out of his own pocket on nothing more than his word. Maybe he wouldn't though. To me, he's just some guy on the internet. I read a story published by a news organization on the internet, and I'm skeptical. I read some guy telling me I can win $100,000 on the internet if I pay $10, and the first thing I'm thinking is, "Is this guy a Nigerian prince?"

Mind you, I'm not saying Douglas Keenan is running a scam. What I am saying is he has done absolutely nothing to engender any trust. The idea people should pay him money simply because he says he is running a contest is absurd. People who run $100,000 competitions do a lot of things to try to make them trustworthy, but Keenan hasn't done any.

Which was made worse by the fact his announcement said:

The file Series1000.txt contains 1000 time series. Each series has length 135 (about the same as that of the most commonly studied series of global temperatures).

Which is nonsensical. People routinely use monthly data as far back as 1880. When I pointed this out on the blog post announcing this contest, saying:

Global temperature series are not 135 data points long. They're much longer. They would only be as short as Doug Keenan is claiming if one were restricted to annual temperatures, a completely unreasonable restriction nobody genuinely interested in studying global warming would apply.

Keenan responded:

@ Brandon Shollenberger

Regarding the length of the time series, the main temperature series relied upon by the IPCC, in its most-recent Assessment Report (AR5), begins in 1880 and is annual.

Which is complete and utter bunk. People don't restrict themselves to examining annual data. Monthly data is used all the time. Keenan's claim here was a ridiculous fabrication that should have gotten him laughed out of the room. That didn't happen. "Skeptics" seemed to take no issue with this absurd claim.

As pathetic and untrustworthy as Keenan and his contest were appearing, I couldn't be certain it was a scam. That is, until Keenan decided to start lying in incredibly obvious ways. You see, Keenan's contest involved people trying to identify which of 1000 series were created via a "random" process and which had a linear trend embedded in them. An obvious first step in trying to do this was to make a histogram of the trends in each series. Here is one a user named Magma created:

11_22-Magma-1

As you can see, there are clear patterns in the data. This suggested a path one could take that might result in winning the contest. Keenan saw people discussing strategies like this and couldn't allow it. As such, he decided to change the contest so this strategy wouldn't work. He has this posted on his site:

22 November 2015
The generation of the 1000 series relies on the generation of random numbers. That presents a difficulty, because current computers do not generate truly random numbers. There is a widely-used method of addressing the difficulty: use a computer routine that generates numbers that seem to be to random (i.e. fake it). Numbers generated by that method are called “pseudorandom”.

A computer routine that generates pseudorandom numbers is called a “pseudorandom number generator” (PRNG). PRNGs have been studied by computer scientists for decades. All PRNGs have weaknesses, but some have more serious weaknesses than others.

The Contest was announced on 18 November 2015. Shortly afterward, a few people pointed out to me that the PRNG I had used might not be good enough. In particular, it might be possible for researchers to win the Contest by exploiting weaknesses in the PRNG. I have been persuaded that the risk might be greater than I had previously realized.

The purpose of the Contest is to test researchers' claimed capability to statistically analyze climatic data. If someone were to win the Contest by exploiting a PRNG weakness, that would not conform with the purpose of the Contest. Ergo, I regenerated the 1000 series using a stronger PRNG, together with some related changes. Note that this implies that the files Answers1000.txt and Series1000.txt were both revised.

The 1000 regenerated series were posted online four days after the Contest was announced—on 22 November 2015. (Each person who submitted an entry before then was invited to submit a new entry, with no fee.) When the Contest closes, the computer program for the original 1000 series and the encryption key for the original Answers1000 file will be posted here—together with the program and encryption key for the regenerated series.

This is a lie. I can't speak to the PRNG Keenan used or any issues it might have, but if Keenan posts the code he used to create the original and the new data sets, it will be different. It is indisputable he modified the code he used to generate these series. You can see for yourself by comparing the original and updated, posted here for consideration. For a simple proof, consider how the original file began:

-0.229 -0.293 -0.385...

Compare that to how the new file begins:

-0.23 -0.39 -0.37...

Anyone can see the new file rounds to fewer digits than the old file did. The more precise data is, the more accurate analysis performed on it can be. By rounding to only two decimal places, Keenan destroyed knowledge and reduced the power of any analysis performed to try to win his contest.

This goes far beyond rounding. Remember this histogram I showed you above? We can create a histogram for the original data set and the new one and compare them. Here is a crude example I made last year while discussing this:

11_23_overlaid_keenan

The green shows the trends of series in the original data set. The pink shows the trends in the new data set. The purple shows where they overlap. That Keenan changed the code he used to create these series is made obvious by fact there are systematic differences in where these colors show up.

We can make this easier to consider by realizing these histograms are symmetrical. Whether the trend in question is positive or negative has no effect on anything. That means we can take the absolute value of trends and effectively "fold" the chart over. That gives us this crude chart:

11_23_overlaid_keenan_abs

As you can see, the original (pink) data set had a spike in trends well away from the central values. The new (green) data set does not have this spike. Instead, it has a signficantly larger number of series with a near-zero trend. This shows Keenan altered his data set in response to people identifying a pattern in his data which might let them win his contest in order to destroy that pattern.

The same thing can be seen in many other properties of these data sets. Here is a crude chart I made showing the lag-1 autocorrelation coefficient for these data sets. Again, the original data set can be seen in pink. This time, the new data set is in blue (purple still indicates the overlap:

11_23_overlaid_keenan_acf

The same thing can be seen in other lags, with partial autocorrelation coefficients and many other statistical properties. It is simply beyond dispute Keenan altered his data set in a way that increased the difficulty of his contest after people had publicly discussed how they might attempt to win his contest. That's completely dishonest. Even worse, Keenan said:

The 1000 regenerated series were posted online four days after the Contest was announced—on 22 November 2015. (Each person who submitted an entry before then was invited to submit a new entry, with no fee.)

Keenan took people's money for a contest then changed the nature of the contest without even offering a refund. That is fraud. It is literally criminal fraud. Keenan took money with the promise of a service then changed the terms of the agreement without anyone's consent. That is defrauding people.

Make no mistake, Keenan changed his data set with the specific intent to make his contest more difficult after having taken people's money. I don't know how many people had submitted an entry fee at that point, but the encoded list of entries shows 54 people have entered this contest so far. That's only $540. It's not a great deal of money. It is, however, money that was clearly taken under false pretenses as Keenan lied about the nature of the changes he made to the contest. This is 54 people Keenan criminally defrauded (unless people submitted multiple entries, in which case there would be fewer).

You won't hear a peep from any "skeptic" complaining about this. The contest was promoted on multiple "skeptic" sites. The proprietor of the largest of these even defended Keenan's lies via e-mail, explicitly stating I do not have his permission to publish his e-mail but:

That language you used in your post might be considered libelous under the circumstances.

I won't publish the e-mails Watts and I exchanged on this issue since he went out of his way to say I don't have his permission to, but I'll quote a bit of my responses:

That the data sets are significantly different is indisputable. You indicate you've read my post, which should mean you are aware of the issue. You've seen the histogram plots showing the changes. I'm attaching versions I've made myself which confirm the results, so there is no doubt that they are in fact what one gets with the data sets. I could show the results of other tests I've run confirming changes in the statistical properties of the data set. It's not even needed though. Any examination of the data files would show the changes are obvious, as the original file began with:

-0.229 -0.293 -0.385...

But the new file begins with:

-0.23 -0.39 -0.37...

While individual data points would be expected to change between runs since there is RNG involved in the creation of series, there is no legitimate explanation for reducing the number of significant figures reported in the files. Doing so reduces the amount of information present in the data, making the challenge more difficult to complete.

Nobody can possibly compare the two data sets and reasonably believe they were created via the same means. The process used to create these series was clearly changed in a way which makes the challenge more difficult. That was the central point of my post and e-mail to you, one you have somehow failed to even acknowledge. I don't understand that, but regardless, you are defending this dishonest and illegal action.

If you want to continue doing so, that's your choice. I cannot compel you to speak out against dishonesty from people you know. Or anyone for that matter. All I can do is emphasize how abundantly obvious what Keenan did is and suggest you not try to imply legal dangers with remarks like:

That language you used in your post might be considered libelous under the circumstances.

To get people to not point it out. Because honestly, suggesting a person who pointed out an illegal activity may be at risk of a libel suit for pointing out a person has broken the law is just...

And:

Are you kidding me? This isn't true at all. The data is not "'more random' than before to prevent a shortcut that has nothing to do with the challenge." Or at least, if it is, that has nothing to do with the differences I've highlighted. Being "more random" wouldn't cause major, systematic changes in statistical properties like the linear trends present in hundreds of series.

I know these changes were made because I am capable of doing things like looking at simple histogram charts and seeing when there are significant differences. I'll note you didn't comment on the two I sent you or offer an explanation as to why they are so dramatically different. You didn't even ask questions or seek additional information so you could try to better understand the situation.

I also know these changes were made because I am capable of understanding simple ideas like, having more information makes it easier to solve problems. You claim the changes to the data sets were merely to make the series "more random," yet I think everyone can understand rounding numbers to two decimal places instead of three does not involve any randomness. All it does is reduce the amount of information present for people trying to solve the challenge.

As you can probably guess from these excerpts, the discussion was... strange. It eventually ended with Watts saying this as his final remark:

Got a chance to look on a decent screen. I’ll ask Doug why the difference exists.

He never contacted me about this issue again. He never re-visited the blog post he ran on his site promoting this contest. He never bothered to tell anyone, "Hey, remember that contest I promoted on my website? It turns out the guy behind it secretly changed the contest to make it more difficult then lied about it. Yeah, sorry, my bad. Maybe in the future I shouldn't promote fraudsters in their criminal activity."

You won't see Watts say that. You won't see him say anything like that. I don't think you'll see Andrew Montford, the guy who runs Bishop Hill where I discovered this contest, say anything like it either. I don't think you'll even see Ross McKitrick say anything like it even though he's a fairly big name in the "skeptic" field who co-authored multiple papers with Steve McIntyre, one of the people "skeptics" respect the most. (Of course, you won't see McIntyre fix a blog post despite having known it had a serious error for the last three months now, an error called into question before he even wrote the post.)

What should we make of all this? I don't know. I think it's become abundantly clear "skeptics" as a whole don't have any integrity or interest in the truth. I won't re-hash that point any further. If you want to see what I have to say on that subject, I suggest you read this post in which I burned any bridges I might have had with the "skeptic" community.

I don't think it hurts to re-hash that issue. I think the total lack of skepticism from "skeptics" merits a great deal of attention and commentary. That's not the point of this post though. The point of this post is today's date. Today is November 30th, 2016. Douglas Keenan said:

The Contest closes at the end of 30 November 2016 (UTC), or when someone submits a prize-winning answer, whichever comes first.

When the Contest closes, the computer program (including the random seed) that generated the 1000 series will be posted here. As an additional check, the file Answers1000.txt identifies which series were generated by a trendless process and which by a trending process. The file is encrypted. The encryption key and method will also be posted here when the Contest closes.

And:

The purpose of the Contest is to test researchers' claimed capability to statistically analyze climatic data. If someone were to win the Contest by exploiting a PRNG weakness, that would not conform with the purpose of the Contest. Ergo, I regenerated the 1000 series using a stronger PRNG, together with some related changes. Note that this implies that the files Answers1000.txt and Series1000.txt were both revised.

The 1000 regenerated series were posted online four days after the Contest was announced—on 22 November 2015. (Each person who submitted an entry before then was invited to submit a new entry, with no fee.) When the Contest closes, the computer program for the original 1000 series and the encryption key for the original Answers1000 file will be posted here—together with the program and encryption key for the regenerated series.

A year ago Keenan said this contest would end today, after which he would post the material needed to replicate his data sets. I wrote today's post to draw attention to that. I think people should all be interested to see the code Keenan used to create these two data sets and how it was modified. I don't think he'll post it. If he does post it though, I am absolutely certain it will show exactly what I've said - Keenan altered the statistical properties of the data set used in his contest in a way which made it more difficult for anyone to win.

As a final note, because "skeptics" let him get away with that, I was robbed of the chance to win this contest. I am confident a person could beat the contest within a reasonable number of guesses (spending no more than $1000 on entries) in its original form. It's only because Keenan lied and cheated to prevent people from winning his contest that that wasn't possible.

Keenan can prove I'm wrong. All he has to do is post code that creates these two data sets which is the same save for what PRNG it uses. If he cannot do that, then he is a liar and fraud who robbed me of a chance at $100,000.

A chance he probably lied about in the first place because let's be real, he was never going to pay anything out. This was just a scam.

5 comments

  1. The contest's webpage has been updated to say:

    UPDATE [2016-12-01]. The Contest is now closed. No winning entry was received. The ANSWER is now available. The program that was used to generate the series, etc., together with some commentary, will be posted here later today.

    I guess that means we'll see if Douglas Keenan lied to everybody like I said he did before much longer.

  2. All I've heard is that he finds blogging about the climate debate boring now. A number of people have expressed similar sentiments. I don't think it's anything more than that. I certainly hope not.

Leave a Reply

Your email address will not be published. Required fields are marked *