Climate Audit

Buddenbrook

Yeah, Bad Astronomy suddenly got hundred of votes out of nowhere. CA were about to catch BA and then suddenly BA were 600 votes ahead. I think a statistician might conclude that when you get steadily 3000-4000 votes during the course of number of days, you won’t suddenly get 500+ votes in mere minutes (in less than a hour I think) without a contributing extra factor.

I can think of only two explanations:

1. Someone posted a link to a very popular liberal/AGW forum/network to defeat those “denialists”. In this case people who have never read BA would be voting BA, which goes against the spirit of the whole thing.

2. Someone cheated to vote multiple times. Shame on him/her.

Bernie

Steve:
I just knew your were not just a puzzle junkie, but actually one competitive SOB.
Let’s go for it.

Jan Pompe

#2

Pharyngula also pulled ahead sharply then too.

#3

I don’t there are many around here that would disagree.

Steve McIntyre

#4. Anybody that plays squash (and I still play league and some tournaments) competes. You still buy beer for your opponents afterwards.

Geoff Sherrington

Out here we put it this way:

“You buy still beer for your opponents afterwards.”

May the force be with you Geoff.

Bernie

Steve & Geoff:
But the beer tastes better when you win!!

Boris

Junk Science has endorsed CA. Congrats or condolences?

Sorry those ideologues are suppressing your votes, BTW.

Sordant G

I support comment 3. But I am also very curious: how do they ensure that people only vote once in 24 hours? I voted; then I deleted most of my cookies and reset my internet address; yet when I went back to the site, they still recogized that it was me. So what technology do they use to do that?

Boris

No, but they will be saved by it obviously. 🙂

Bernie

Anyone:
When it says that Polls close November 8th, does anyone have a sense of at what time exactly? Is there an way of monitoring the voting patterns in the last hours of voting? Is there a way of having them audit the voting – and broadcasting which sites were attracting those given to subverting the voting process? I am involved in a survey of “the best…” and we have to be very careful about identifying groups that are trying to stack the deck.

Steve McIntyre

The Pharyngula vote seems to have collapsed.

andre

yes and for some reason CA is leaping forward. BA 5972, CA 5922

steven mosher

I suggest we plug the daily vote totals into some mannian sausage machine and see if
we get a hockey stick? UC?

Dave Dardinger

What do you have to do to get the stupid browsers to show updated counts? Even when I shut down my browsers, either IE or Netscape or both and then restart it won’t update. It only seems to do it when it allows me to vote again and then sticks with that value no matter what.

welikerocks

re: 24
I posted the links and the whole story this morning on a blog that I belong to that usually has 1,000 of people logged in at one time (its nominated also in another category) and who’s members are at times very interested in this Global Warming issue. Don’t know if that’s the reason, but I hope it helps!

Dave #28, I have a Mac, and IE and Netscape were not working for me anymore at all. Had to upgrade to Firefox, and that seemed to fix all my internet problems.

Luboš Motl

David, it could also be a proxy in your neighborhood. In that case, you must modify the URL to update it, for example open

http://2007.weblogawards.org/polls/best-science-blog-1.php?motl=update1

Or change the irrelevant characters after the question mark. 😉

Sam Urbinto

lol, the bit about SM endorsing CA is a link, and the link goes to… Back to his blog!

So, Steve, how does it feel now to have your blog be an “undeserving mouthpiece for right-wing hackery ?

Whatever the hello that means.

Larry

“Undeserving mouthpiece for right-wing hackery”? I thought the category was supposed to be science blogs. I can see what puts a bee in their bonnets.

Spence_UK

Oops #38 cross posted with Sams comments… you beat me to it, Sam.

MarkW

The few times Steve has allowed his personal politics to slip through, it’s quite obvious that the term “right wing” would never be applied by anyone whot thinks for a living.

Just goes to show you how knee jerk many liberals are getting these days.

Hoi Polloi

Vox Populi!

I guess the grapes are sour for PZ Meyers…

Larry

Boris, if they’re the listeners I’m thinking of, BA will be lucky to have a visible bar by the time they’re done (if the server survives it, anyway).

Sam Urbinto

I think this long ago moved out of the realm of blog reader popularity of the blog and into an issue of voting against somebody else more so than voting for somebody.

It’s a battle of ideology!!! Vanquish the other wing!!! Slay the foe!!!

Oh, brother.

MattN

I’m having alot of trouble getting on to vote. The traffic must be immense….

Larry

48, looks like he dunnit. People at a blog like this think that a few thousand is traffic. You find out what traffic really is when somebody like Drudge or Limbaugh turns the firehose on.

Liz

9:27 am PST Finally got in. Then I refreshed, and CA lost a vote. How the heck can that happen?

Doug

I didn’t much care until they started calling us quacks and right-wing mouthpieces, or worse yet, accusations of using complex jibberish to sound like we know what we are talking about. I thought with my three college degrees and daily reading I knew science when I saw it. Science is found here. There may be some right- wing nut-job noise, but it gets drowned out in the equations.

Thanks Steve!

Sam Urbinto

Larry over at The Reference Frame sez:

BTW, it’s been confirmed that pharyngula has conceded and instructed their minions to vote early and often for BA. That explains why their numbers aren’t rising.

And yes, it’s for political reasons.
Larry R. | 11.07.07 – 5:59 pm | #

——————————————————————————–

Interesting. The WA site is impossible to get onto, and rumor has it that this has made it onto talk radio. If so, this could be a slaughter.

I don’t think Steve wanted this to get political like this, but if you look at pharyngula and BA, they started it, and said that people should vote for BA as a vote against “denialism”. So if they get clobbered, I’m not going to feel bad about it.
Larry R. | 11.07.07 – 6:44 pm | #

——————————————————————————–

steven mosher

RE 58. This is fun

Dave Dardinger

I see what you mean, it’s very slow connecting to them and then all it gives me is a shot of the last results I’d had, not even the main index. Anyway, while this site isn’t really political, just about every other skeptic site is, so what can you do? While I suppose the weblog awards people could try punishing CA, the fact is that most of Rush’s listeners are also online and read plenty of blogs. It’s just that they’d never heard of the weblog awards as I hadn’t until Steve mentioned it. So the votes which may come to CA because of it are legitimate according to the contest rules and should stand. Of course, now that it’s out in the open as a political contest, who knows how many votes either side might pick up?

Spence_UK

Anyway, while this site isn’t really political, just about every other skeptic site is, so what can you do?

This is part of the problem, my thoughts were “hey these sites must have scientists at them, if it is possible to explain to them there is more to this site than the average sceptic site, they might take an interest”

More fool me 🙂

It is easier for them to simply lump CA in with other sceptic sites that focus more on the politics. Of course, this line of reasoning works something like this:

Premise. Political right-wing blogs use bad science to cast doubt about the work of climate scientists

Premise. ClimateAudit casts doubt about the work of climate scientists

Conclusion. Therefore, ClimateAudit is a political right-wing blog

Of course, PZ is helpfully illustrating the anti science reasoning method Affirming the Consequent, so that when his commenters see it on other anti science sites, they will recognise it 🙂

cbone

Re: 54,55

No fears your votes are safe. From the web awards page (just below the vote tallies):

Please note that to reduce the load on our server when you cast your vote the counters are incremented on your display and the vote is recorded in the database. If you refresh your browser it will appear as if your vote has disappeared. That is not the case, it’s simply a caching issue. Your vote is safely recorded and everyone else sees your vote. If you close your browser completely the next time you load the poll the totals displayed will include your vote.

Larry

64, the problem is, if you take a disinterested evidence-based approach, you get branded a “skeptic”, and then a “denier”. The only thing worse than that is for a believer site (*cough, realclimate, cough*) to pretend to be an objective science site.

Truth be known, this is as close to an objective site as they get.

Larry

It doesn’t help that some bright boy decided to do this in flash. Flash is powerful, but efficiency isn’t one of it’s strengths.

steven mosher

RE 70. idjits

Jason L

The voting site appears to be down. Is anyone else having trouble reaching it?

Andy

I hope the people voting for CA from hearing about it on talk radio take the time to actually peruse the site. Especially since Steve Mc put the recap of the Wegman/North hockey refereeing up just in time.

My hat’s off to the regulars, though. CA was doing great even before JunkScience and talk radio threw in support.

Larry

It’s not down, but you may have to retry 4 or 5 times.

What I don’t understand is why I have to close firefox and reopen it to get a refresh. If I just refresh, I get a big gray blank background. Some flash quick, maybe.

Larry

77, Steve specifically avoids politics. Show me another climate blog, on either side of the issue, that does that.

Internal organizational politics of The Team notwithstanding.

Spence_UK

Re #71

I took a quick look at your search. I did wonder if some of the bad astronomy links dated back to last year when they were also in the running, but I noticed that of the 20 found with the climateaudit search, several of them were actually suggesting not to vote CA…

Google search estimates the number of hits, and overestimates with BA, there are around 38 links, and some of them are from last year – although they are consistently supportive, he does not have the same political smear campaign against him that has been orchestrated against CA.

paul

Can’t get through to the site from here in NZ

Spence_UK

It could be a DDOS designed to maintain BA’s lead

Larry, I don’t think anyone is going to go to those sorts of lengths. Remember there are a lot of contests taking place on that server – not just the science one – if any of those contests got a link from a high readership site, I suspect all would suffer. As you point out, more likely a problem with the way the software is set up.

It has been interesting following this little episode in CAs life, but a new MATLAB contest has just started up, so I’m off to spend my time on an entirely different unproductive and time consuming venture 🙂 Apparently we’re gene splicing this time around. Fun for geeks!

Larry

86, I agree that it’s not one of the more likely explanations, but I’ve seen it happen.

STAFFAN LINDSTROEM

#15 Boris …Get Carter! Both the original movie(1971) with Michael
Caine and: Bob Carter: “Climate Change. Is CO2 the cause?” Ending
Part 1: 3 fat polar bears are barbecuing a chicken…NTS they
don’t look catatonic at all, perhaps they drink gin and tonic
L…O…L…Met prof Carter in Stockholm briefly last year…
Compared to him Al Gore is still “the Bore” And I recommend
Tubesucker to download the clips and Flashplayer to play them
as they freeze when streamed..!! Oops! Prof Carter is on YouTube
as mentioned earlier on another thread. If Bob is right we’ll
also freeze in the next 10-30 years…
As admitted on another WAF threads on CA voting for CA has
been a little “fraudulent”…who could have guessed??

Sam Urbinto

Based upon reading the other replies since, I’d think my guess was right. It does appear to be just lots of regular traffic. If a lot of folks without Flash come in, that’ll be a load in and of itself.

And many servers (and/or server software, say Apache) aren’t tuned (or tweaked) very well.

Has anyone emailed them?

Oh, and if this is normal traffic, what happens if this gets out on other forms of media? Ouch!

Now that I think about it, we keep it fairly apolitical here, but this is a pretty political subjecty, so it’s obvious there’s two camps with battle lines drawn, regardless of the focus or behavior of the site itself. (Much like radio hosts for example as I mentioned vs the audience) So it couldn’t help but spill over, it pretty much started that way.

John A

I can’t even get to the awardswebsite. It seems to be crushed by the traffic.

It’s quite amusing to have it happen to someone else for a change.

Larry

I’m just not seeing evidence of huge traffic in the numbers. The site is weak.

dirac angestun

BA 6500
CA 6245

Bernie

Look out for the HS tomorrow afternoon!!

Stephen Richards

I can’t get in from France. I’ve been trying all night

D. Patterson

Just like IRL, can’t get into the polling place to vote.

John A

Can I recommend this comment on Pharyngula?

Jeremy Friesen

Mark R:

No good comes from that sort of assumption. Yes, web polls can potentially be subverted, and yes, it could just as easily happen in favour of CA as in favour of BAB. It’s best to leave this issue at the roadside. Suspicions, unless they can be proven, are useless.

Mike B

#63

This must be fascinating to a statistician.

Actually, it is about as boring as it gets for a statistician. Internet polls give you self-selection
bias, ballot box stuffing (‘bots and all that java stuff I’ll never understand), and websites crashing.

Not a representative sample to be found anywhere.

Larry

109, I could be persuaded either way, but the fact that these science blogs have taken over and passed Malkin and Hufpo in the best blog category strongly suggests that there’s more going on here than just the random effects of people doing what they otherwise would have done.

So it’s probably nothing, but I’ll keep my eye on the ratio anyway, tyvm.

Larry

110, exactly. And the fun for the statistician is to try to infer those things from the patterns.

steven mosher

Cement: Drying.
Boris: writing.

Steve Moore

RE #106:

I read the whole string, and got the impression there are a lot of folks over there who heavily invest in arguments from authority:
“My authority is more authorative than your authority! So there!”

It only takes one “authority” to skew policy if that one has the ear(s) of power. Refer Lysenko.

steven mosher

RE 119.

The embedded link was a non starter for me ALL DAY TODAY. and even though I havent voted
in two days (Honest) I still get the boot from them.

woodentop

#119 that worked for me too, strangely… though the site loaded slower than a slow thing moving very slowly.

Bernie

There have been no additional votes registered in the last 45 minutes, hopefully there checking to see who has been stuffing the ballot box.

MattN

6:20pm, almost 300 back now….

MarkR

The difference between the CA and BA vote has settled to being very close to 270 votes all the time.

Leonidas Lakedaimonian

And now the poll list is empty, as if their database had crashed.

chris

I voted once several days ago. Every time I went back to the site, including just now, it said I couldn’t vote as I had last voted less than 24 hours ago. Should I get paranoid at rigging? Is Diebold involved?

Larry

BA’s taking off like a bat out of hell, CA’s doing pretty good too, but slowly falling behind.

Larry

All links seem to work now; I think they had to reboot one of their servers or something.

Larry

Gap = 321 and rising…

David Ermer

Steve,

I’ve voted from every computer I own (I have multiple Linux boxes that I run optimization simulations on), but the gap isn’t closing.

Good luck, this site deserves to win. It is one of the few places I been that actually does real old school science.

steven mosher

RE 102. Surrender then.

Dan White

After reading the comments on the other “science” weblogs regarding CA being political and junk science, I was at first astounded (gobsmacked, some might say) at the hubris on display.

The big picture is that CA wasn’t even on the map last year, and is now in contention of winning the award (such as it is) against a “team” of opponents. I think next year’s results will be very telling. There’s a saying that goes something like, “A lie makes its way half way around the world before the truth can get its pants on.” I think this poll illustrates how the truth is beginning to catch up to the lie.

Great job Steve Mc!

Larry

It’s wee hours in Europe, late evening in eastern North America, early evening in western North America, and day in east Asia and Australia. This means the Euros get one more crack at it, as do the eastern North Americans. Western North Americans, not so much. It’s hard to say how it’s going to play out, but I think the gap will narrow, but not enough.

Bernie

#145
Brian

You are right, this is silly on a huge number of levels. On the other hand, when you experience the invective and disdain from those at other sites, it always feels good to give them a poke back – childish perhaps, but not malicious. It is also true that skeptics and contrarians (not cynics!), like many here, just love a good race, especially where the underdog looks like it is giving the favorites a run for their money.

Kristen Byrnes

There’s nothing silly about this, it’s what the thread is about.

steven mosher

RE 151. A scholar and a comic. Only in australia.

Jeff C.

Interesting, the Blog readability test http://www.criticsrant.com/bb/reading_level.aspx ranked the education level required to read the competitors as follows:

Climate Audit – college, undergrad
Pharyngula – High School
Junk Sciencs – High School
Bad Astronomy – unable to rank (that can’t be good)

Some others
Real Climate – High School
Watts Up With That – High School
Dr. Pielke Sr’s old Site – College, post graduate

Mark T

Interesting, Jeff…

Mark

Georg Dimitrov

Just saw CA ahead !

Tomas S

CA still in first place right now.

someone

The CA vote has been steadily rising at a rate of about one a second for a considerable time. A very consistent, and i’m emabarassed to say suspicious, state of affairs. I don’t think anyone (on either side) should come out of this feeling good.

Anthony Watts

154 Jeff C. Interesting indeed. I ran a couple of CA’s typical detractor sites:

bigcitylib.blogspot.com – Junior High School

rabett.blogspot.com – College, Undergrad

scienceblogs.com/deltoid/ – High School

As for my own ranking of WattsUpWithThat, “high school” I’m very pleased. Because my strategy is the same as when I did weather and science on TV, i.e. make the subject matter understandable for the broadest number of viewers.

Chris Christner

Someone said:

The CA vote has been steadily rising at a rate of about one a second for a considerable time. A very consistent, and i’m emabarassed to say suspicious, state of affairs. I don’t think anyone (on either side) should come out of this feeling good.

I’m sure James Hansen would reply that it doesn’t matter if the voting is rigged or otherwise usufruct up, because it was equally as messed up last year. So the data may be suspect, but it can be adjusted for accuracy. Hence the current results are completely valid.

Peter Hearnden

The CA vote has been steadily rising at a rate of about one a second for a considerable time. A very consistent, and i’m emabarassed to say suspicious, state of affairs. I don’t think anyone (on either side) should come out of this feeling good.

Indeed. it’s a seriously compromised vote.

Buddenbrook

To be fair, it looks quite likely now, that, in all probability, someone has cheated on behalf of CA too.

Guess you can’t avoid these things with polls like this. There will be always someone. One person is enough.

But still, outside of these surges of votes, CA have been the most popular.

Kristen Byrnes

I remailed all of my lists about an hour ago. Good luck Steve!

Anthony Watts

There may be other things at work. Newsbusters.org has placed an endorsement for CA on their main page and it is wakeup time on the east coast.

As a result of that and other endorsments, there may be a blog and website push going on by supporters:

GOOGLE
Results 1 – 10 of about 1,030,000 for “vote for climate audit”.
Results 1 – 10 of about 30,700 for “vote for climateaudit”.

Jan Pompe

#170

To be fair, it looks quite likely now, that, in all probability, someone has cheated on behalf of CA too.

Why? Isn’t the American East coast waking up about now and firing up their computers (for breakfast) about now and wouldn’t you expect a surge about now?

GrG

#176
So everyone on the East Coast has exactly the same opinion on climate matters? Sure.

John A

13:30 GMT

CA: 9232
BA: 8751
Pharyngula: 5353
JS: 3532

What caused the sudden rush for CA? Do we have a Hockeystick-shaped voting profile?

John A

I think I’ve found a flaw in my plan for blog domination.

MarkR

#180 John A. My theory wot is mine, is that BA has been artificially pumped up, till today, and somehow the plug has been pulled.

MrPete

I too have noticed that the polls go up in “clumps”. Watching for the last few days, I suspect it is partially tied to the time of day that each group of readers first became aware of the contest. Every 24 hours or so after that, the blog will receive a burst of votes.

On that basis, don’t be too surprised if CA falls behind again later today: BA has had a daily surge, and I’m sure part of it is simply the day’s “catch” of repeat votes. Sorry, I didn’t know what time the vote would end and paid no attention to exactly when the surge hit.

Bernie

JohnA:
At least you could have give me (#100) credit for the HS line!! Though frankly I thought the HS would be wielded by anti-CA forces. The pattern of voting is odd in that almost 33% of the CA/BA votes have appeared in the last 12 hours – moving from a combined 13032 at 5:15 PM yesterday to a combined 18353 as of 8:45 AM EST. The increase in BA during that period was 2207 or 33% and for CA was 3114 or 49% – that looks suspiciously lke someone ballot stuffing for CA. (Who and why are obvious questions – and it might not be a CA supporter but a CA detractor.) My hope is that the competition organizers are as good as their word in cleaning out ballot box stuffers regardless of for whom they are voting.

Steve McIntyre

#186. That’s a good point, Pete. CA has been relatively stronger in the mornings than in the afternoons. MAny BA votes will come back into play in the afternoon.

Bernie

GrG:
What brought you to the site? I think most of us skeptics here are still skeptics and do not take anything at face value until we have checked the data – when we can get access to it. Specifically, I am suspicious of anything that allows for multiple voting and leaves open the possibility of easy fraud.

Dave Dardinger

re: #180

That’s it! As soon as the sun rose on the US East Coast it warmed the atmosphere, teleconnecting to the temperature-like bars on the voting results. And since CA is the only of the site in the running which is primarily about temperatures, naturally the effect is greatest for CA. I knew it was something simple like that!

Dave Dardinger

Oops, that should have been #189 in my previous message, to give credit where it is due.

Bernie

Hey John, you think I am going to share my winnings with you!

Hasse@Norway

Hey! CA leading by some 900 votes 🙂 This is just like the Indy-Pats game of last week. Pats trailing until the last 10 minutes just to keep things interesting!

Abu Liam

CA received a nice plug on Ice Cap/Newsbusters that might help explain it

Deborah

A lizard at Little Green Footballs asked other lizards to support you.

Buddenbrook

Good, skepticism is the way to go.
I shall patiently await the new views of those that previously complained about BA & PZ. The surges over which they were so furious are starting to appear rather minor compared to this morning’s explosion. So, I must ask, where is thy furor now?

Agree totally. It’s only what is fair and logical. Everyone can draw their own conclusions about the validity of internet polls, unless there is ID verification, which polls like this of course never will have.

Still I’d maintain that outside of the voting surges CA got the most votes.

cbone

Down the stretch they come! CA has a 932 vote lead at 9:30 AM EST. 7.5 hours remain in the voting. (polls close at 12PM GMT)

Andy

CA just crested the 10,000 vote mark. Wow!

Re #207, Steve as usual hit the nail squarely on the head. I’d add that the group running the Weblog Awards at least takes this “fun” process seriously enough subject the results to audit (think “internal audit” in the business world). That’s more than we can say for the Team.

John A

14:50 GMT

CA: 10507
BA: 9536
Pharyngula: 5417
JS: 3618

Sudden mysterious surge for CA continues. Becoming increasingly skeptical about this vote.

Calm down, John. It’s just a popularity contest.

Bernie

Steve:
I would not request an audit of such an award either. I agree totally with your point – the truth content of significant statements need to be carefully and openly evaluated. I guess my view is that the burst of votes is unlikely to come from CA readers but people who
have been drawn here (or to BA) to make some other point that is independent of whether this is the Best Science Blog. Because
this particularly internet voting process is subject to a large number of possible influences that detract from any reasonable meaning that can be attached to the “Best Science Blog”, the competition is “silly” though still fun. In terms of real outcomes, I would be interested in seeing how the process has impacted traffic on CA both during the competition and, say, four weeks from now. Clearly this thread has attracted a couple of new commentators. Let’s see if it helps add substantive contributors.

jeremiah

i’d say the surge was mostly down to the drudgereport/newsbusters/coleman story. the drudge report gets 17 million visits a day – you’re bound to see a surge from a link like that

Hoi Polloi

Yep, that’s why I asked yesterday if there’s an increased traffic on ClimateAudit. For sure a number of voters have taken the time to actually visit CA, as I did visit Pharyngula, which unfortunately was quite a disappointing experience.

Steve McIntyre

test

Steve McIntyre

Bad Astronomy is surging much more strongly than CA right now.

I looked at 10:49 Eastern and, as compared with John A’s earlier look., it was
10649 – 10507
9898 – 9536

I blinked and BA had 100 new votes while CA had 16. So this isn’t over by any means.

Larry

221, I’m wondering if the poll taking database system itself is introducing some lags into the accumulation of the totals that make our snapshot trends useless. The way these are piling up doesn’t seem natural to me.

Larry

That was strange. 221 became 222, and a new 221 came out of nowhere. It looks like we have similar syncronicity issues at the CA server.

tpguydk

#219: I agree. This site does hands on science and so few do, and for that I greatly appreciate its existance. Hopefully you’ll get more traffic here and more people will want to ask questions and learn something.

I have to say I’m actually a recent skeptic, and it’s learning about the IPCC process and reading this site that’s made me very skeptical of what they say the past was like.

M. Jeff

Another conspiracy theory:

It is not coincidental that the current 33.2% vote for climateaudit.org is the same as the 33.2% faction, (my estimate), of scientists who are true believers in the totality of the CO2 AGW theory, a percentage vote that will convey absolute consensus status on the fact that climateaudit.org is the Best.

Larry

219

You’ll never convince the awful commenters at Pharyngula with facts and analyses, but as soon as they discover there is a new winner they’ll change tune.

The power of the perceived consensus. Those are called sheeple, and there are too many of them who think they have original thoughts by following the herd.

Steve: C’mon. This poll won’t change anyone’s mind.

fFreddy

Who did you vote for, Peter ?

Ethan Shepard

229: Larry,

You’re right about the IP log to an extent. The work around is pretty easy. I’m not “engaged”, but am tempted – if just for fun. Whether CA or BA wins is largely irrelevant, as the process has given CA and Steve a nice venue – a needed venue. That is a win!

Larry

What I find a little hard to explain, and IT guys please chime in, is that we’ve had several incidents of their server choking, followed by a very rapid run up in BA’s numbers. From what I can determine reading the thread, the earlier run up in CA’s number weren’t preceded by a server freeze. Is this basically correct, and if so, what might be going on?

iamnoone

224 “I’m wondering if the poll taking database system itself is introducing some lags into the accumulation of the totals that make our snapshot trends useless. The way these are piling up doesn’t seem natural to me.”

I’d imagine the total is updated in batches. Otherwise, you’d have to lock the database everytime someone tries to vote. Chances of that working successfully is inversly proportional to volume.

Larry

240, which would explain why things move in fits and starts. It looks like they update batchwise, but then have some interpolation algorithm to make it appear like it’s happening in real time. That would explain the peculiar behavior perfectly. I’ll bet that’s what they’re doing.

Larry

CA pulling ahead. Almost 600 now.

Larry Sheldon

‘Tis sad to see nominally sensible people being distracted by a “poll” that is about as un-reliable and meaningless as a poll can possibly be.

This is not even a “popularity contest”.

(For the record, CA is the first blog I started “watching”, and one of the ones (of the ones I have looked at) that I still watch.)

Let’s get back to important stuff.

Tom Vonk

Europe has voted – that’s what happens when the early morning comes in US East .
And in Europe I know many people (rather higher level of scientific understanding) that read and sometimes even post on CA while BA is not so well known .
I am aware that this statement has no statistical value , it is just a snapshot from my personnal surrounding .
However if it could be generalised , I’d say that CA gets an easy lead in Europe .
Don’t know about US .
I have voted right now .

Larry

246, 247, that’s why the blog is organized into threads. If this annoys you, go find something that doesn’t. Just like any poll or election, this kind of trend forecasting is interesting stuff in it’s own right.

And since you’re not paying me, I don’t have to be doing what you insist. If you have employees, go boss them around.

Jean S

BA has the poll embedded to a post.

Mark T.

Almost as immature as people coming in and whining about how immature it is. This is nothing more than a light-hearted break from all the other action going on in this arena. That there exists a separate thread to do such venting is testament to Steve’s willingness to keep these discussions out of the normal science threads (yeah, even Unthreaded). 😉

Mark

Iain

In the last 60 minutes BA has gained 1540 votes, while CA gained 1564. I can’t help but think we might be watching the battle of the ballot stuffers. Which makes it decidedly uninteresting.

Larry

Why do people make the effort of coming into a room, yawning, and then leaving?

steven mosher

Can we get nested comments?

http://meidell.dk/archives/2004/09/04/nested-comments/

GrG

#238 – Larry

What I find a little hard to explain, and IT guys please chime in, is that we’ve had several incidents of their server choking, followed by a very rapid run up in BA’s numbers. From what I can determine reading the thread, the earlier run up in CA’s number weren’t preceded by a server freeze. Is this basically correct, and if so, what might be going on?

Are you associating the server freeze with BA’s “run up”? Why can’t it be associated with CA’s “run up”? Wouldn’t a DDoS attack seem more plausible after an increase – to prevent the other side(s) from voting?

And why was CA’s morning surge due to “East Coasters logging on” or newsbusters/drudge (their other surges are also A-OK); yet BA’s surges are all due to something “more sinister”. Of course, people want their side to win, but be objective, please. Some people on here on being rational about it, but others will do a 180 when the opposite of the situation presents itself.

Larry

263, if you read subsequent comments, there’s a plausible answer to most of those questions; having to do with how the system does database commits. And that was a question, not an accusation. It’s still not completely clear just exactly triggered their server freezes, but there are many possible sinister and non-sinister explanations.

Just as with climate science, you can’t tell which way the train went by looking at the tracks.

Rick Ballard

Como se dice, “Waste of time”?

Si dice: “Votate senza freni o vi spacco la testa tutte!”

‘Course, that’s just a rough translation…

Luboš Motl

Dear everyone, I think that you should finally notice that the table shown on the Awards page is not real-time. It is updated only occassionally. I don’t claim that it follows that there is nothing unfair about the numbers but the things you call strange are simply not strange.

Larry

266, This has been a learning experience. As Luboš said, the actual database updates aren’t real-time, and so the ratios will change abruptly, as the flash display interpolates results. So the most likely explanation for an abrupt change in the ratio is a database transaction.

GrG

#265 – Larry, I agree.

#268 – Steve, the emphasis on the sinister part was mine in order to show its absurdity; I should’ve used italics, I suppose, so my appologies. So hold up with trying to straw man me, please.

Of course people can change their opinions; but flip-flopping is something else, and that’s what people here should be careful not to do. The only side I advocate here is that of being rational.

Jim B

CA – 14021
BA – 12924
———-
CA Up 1097
4 Hours to go!

Rick Ballard

Luboš,

Really, really not strange when campaign ads continue to show up at various other blog venues.

Gunnar

>> the actual database updates aren’t real-time

Well, database updates occur as they are submitted. It’s probably just that the web page itself is cached in memory, and only refreshed once in a while.

Jeremy Friesen

The comments on the Pharyngula site have been very interesting… mostly well thought out reasonable arguments leading to a few people changing their position and agreeing to look at CA for what it is.

Jay Currie

Steve,re 268, it is quite possible that the CA/BA race can blow by races for “bigger bloggers” simply because there are a lot of people for whom CA – one way or another – matters a lot. If you are on the Team or support the IPCC’s view of the data, CA is trouble. And, with any number of political parties having climbed aboard the AGW bandwagon, that sort of trouble has huge ramifications.

Web polls are meaningless but they attract attention. Win or lose CA is now a good deal more famous than it was last week. People who had bought the “consensus” without really thinking about it are reading things which give them pause.

Anthony Watts

RE283 I think you speak of “pollarization” 😉

What a day, my shiny new laptop went DOA in the last few hours and I left it open and on the voting page at webawards. Another reason to hate flash 🙂

Aileni Noyle

No longer an argument – CA is the Best. Congratulations… I did read that right, didn’t I ? My eyes are a bit off today.
A.

Gunnar

>> Count how many votes are cast per second. That would be a lot of disk thrashing if they were individually written. Now observe how the numbers behave. It’s pretty clear that there’s some sort of economization going on. It makes a lot more sense when you look at it that way.

It’s interesting how your comments look when you get into an area that I know a whole lot about. They appear ignorant, no offense intended. Yet, your post projects confidence in what you write.

Votes/second? A modern relational database can make 38k updates in fractions of a second. There is no way that this voting contest is overwhelming any database that I know of, so your comments appear Mannian.

There is no client side app, other than the browser. Unless you’re talking about the little animation that “appears” to be counting votes, but this is simply a visual gimick. There is the database, a web server, and the browser. The web server is caching pages. Periodically, it refreshes the page content from the database, which IS updated in “real time”.

Larry

284, let’s see if we can correlate CA’s traffic with temperatures (on different time scales, of course).

Larry

Gunnar, that is a flash app. Try it with a browser without flash. There’s no way in hades that those counters are being driven that quickly through the internet. Something’s definitely being done on the client side. The rest of it’s conjecture.

But just watch the numbers. That’s not the pattern of real-time voting.

PeterS

It’s worth wondering if a large anti CA block vote might actually be cast for CA. The outcome would be CA winning the poll and later being publicly discredited for ‘cheating’. The discredit would have a lot more impact-value at undermining CA’s claim to ‘honesty’ than the winning.

Larry

295, add that to the possibility list. Yes, that’s entirely possible. The reverse is also possible, too. And both at the same time.

I still think what we’re looking at primarily is the propagation of electronic chain letters, both through email and through websites.

Bernie

JohnA
Did you notice the amazing periodicity in the data? I bet that it is also correlated with CO2 emissions from automobiles? The more CO2 the higher the interest in CA. What say you add a 7-day filter and see what happens.

Larry

CA just broke 15,000. This is obviously more than just the CA gang.

Jim B

295 – No, it won’t happen because this is not a life and death fight for CA. Steve would just have to say the cheating was wrong and whoever did it was wrong, and he’s in the clear. If it gets out that weblogawards.org can be easily hacked, it’s their bread and butter, if they lost their reputation they lose revenue, if a hacking issue came up my guess is they would keep it quiet.

SolarPowered

This lurker has been commending this site and the work of John Daly to our Australian journalists for years. Alas, the absence of understanding, education, political expedience or the greening of the bureaucracy has failed to make a dint in the absolute nonsense that spews forth from our media.
The media science guru (Dr. Karl) declared AGW must be true as there is more papers for the theory than against it.
Our learned institution distanced themselves from Professor Bob Carter as he did not represent the views of that institution.
Some half-wit at Griffith University in the carbon IPCC trough declared how smart the team was, after all they had solved the ozone hole (same comment made by (bear man pig) Gore here in Aus on the day of a record thinning of ozone over the antarctic)!

Climate Audit should receive this puny accolade and the contributors should bask in the glory with the support of the astronomy geeks. When Royal Societies are choking science and Nobel prizes are being handed out for terror every bit helps!

Steve McIntyre

CAn someone help me find a reference: the argument is made from time to time, that , if the Hockey Stick were wrong, the situation would be much worse due to greater climate sensitivity. Can anyone locate a quote for me?

iamnoone

There is no client side app, other than the browser. Unless you’re talking about the little animation that “appears” to be counting votes, but this is simply a visual gimick. There is the database, a web server, and the browser. The web server is caching pages. Periodically, it refreshes the page content from the database, which IS updated in “real time”.

they’d have to lock the “total” field to avoid the “lost update”. it would be easier to write the vote to a temporary table and update the total field from that table via a periodic process. The temporary table would have records with AUTO_INCREMENT ids to keep them unique. The database will ensure the ids are unique by using locking code that’s way more robust than anything a flash author could write. The temporary table would also allow for recovering votes after crashes. It’s extremely tough locking filesystems and databases without getting race conditions that cause lockouts. I don’t know how they’re doing it, but I wouldn’t waste my time writing code capable of locking out 45 submits till the current one is complete.

Papertiger

They gave Al Gore a prize for saying; the fight is over, the science is decided, a consensus has been reached, there is no argument.

They called Steve a crank on the Wikipedia.
Then they sent out their goon squad to squash the Climate Audit.
Over 6000 cyber attacks every week.
Knocked him offline after he proved Hansen had screwed the pooch.

The Awards committee put him on the card as an afterthought, touting the fight as a chance for a “nobody” to become a “somebody”. The current and past champions dismiss him as the anti-science blogger.
Pharyngula and Bad Astronomy joke among themselves about which one should get the title this year, “You want it again BA? – No you take it this time.”

The match is supposed to be easily won by Bad Astronomy, but someone forgot to tell Steve, who sees this as his shot at the big time.
With a small band of followers.

DING DING DING DING

http://www.discoverynet.com/~ajsnead/allsongs_1/rocky.html

“..it really don’t matter if I lose this fight. It really don’t matter if this guy opens my head, either. ‘Cause all I wanna do is go the distance. Nobody’s ever gone the distance with the hockyteam, and if I can go that distance, you see, and that bell rings and I’m still standin’, I’m gonna know for the first time in my life, see, that I weren’t just another bum from the blogisphere. ”

Thirteen thousand votes and counting. Does that sound like the argument is over to you, Mr. Gore?

…he shook up the world. … he changed history.

STEVEN! STEVEN! STEVEN! STEVEN!

IT AIN’T OVER MR GORE – NOTHING IS OVER.

Bernie

I just went to BoingBoing. No facts, no data, just “all knowing” opinions.

Dev

The results displayed by the flash applet come from a separate text/xml file from the weblogawards site. They are NOT embedded in the flash swf code.

Steve McIntyre

#312. That’s what I was looking for. It’s too bad that Connolley has gotten so petty at wikipedia; he often has interesting takes on things.

Mot Normalt

Okay all introverted lurkers! (This is my third post ever here)
Let’s show all these “skeptics” what a “skeptic” with motivation can do..
Call a friend right now, apologize for all the climate talk you have polluted the conversations with, and promise to shut up (except when you are asked spesifically), if they just give a vote to CA right now.

MOSHPIT

re 288. black hats are behaving themselves. Only an f**ktard would hack a blog vote.

Somebody build this counterfactual to Darwin (jim B) a shorter bus to take to school.

My canadian Mohel can keep the tip if he snips.

Iain

#302, and how do you deduce that “it’s clearly CA who is playing games with the votes” from the following voting patterns?


17:00 18:00 19:00 (GMT)
sciguy 12 10 10
JS 96 67 62
ITP 7 7 5
JBS 2 5 1
Phar 83 72 70
BAB 1540 1149 1137
ISW 9 7 5
SB 5 4 5
CA 1564 1726 1142
BA 4 4 6

????

brian

“IT AIN’T OVER MR GORE – NOTHING IS OVER”
That’s just one example of the tone surrounding this contest … the ideology battle lines are even more clearly drawn … while a good showing in this contest certainly brings more attention to CA, it also brings more ridiculousness in my opinion. Hopefully SteveMc won’t have to spend time moderating commenters (both venting loyalists and rabid attackers) when he could spend it continuing to do the analyses he’s been doing.

Larry

326, if you don’t like this thread, don’t come here. Do you make a habit of walking into people’s homes and complaining about the beer cans?

Gunnar

>> they’d have to lock the “total” field to avoid the “lost update”. it would be easier to write the vote to a temporary table and update the total field from that table via a periodic process.

Ok, there is no reason for a total field and no chance of a “lost update”. You presume a stupid db schema. There is a table Vote, with columns BlogID & IPAddress.

When the page renders, it does SELECT count(*) FROM Vote

iamnoone, the rest of your post is non sensical.

>> Gunnar, once again, we’re arguing when we’re in agreement on the fundamentals.

I hardly think so. I’m speculating that this is your response to being strongly challenged.

>> But the animation isn’t irrelevant. I’m speculating that the animation is also doing some interpolation. This is based on the way the gap seems to move in steps.

Oh now you’re speculating, but no hint of that in #280. There is no mystery. The cache expires every so often, could be 5 minutes, could be 30. When it does, the next page request causes it to render fresh. The page then executes SQL statements against the DB. HTML is constructed, along with the data and the embedded flash or javascript. The HTML is saved in the cache, and then delivered back to the requesting browser. When the browser displays the page to the user, it runs the animation. The animation cannot possibly be changing the numbers from what they were when the page was rendered on the server. Subsequent requests from thousands of others receive the cached version until the cache expires again. Haven’t you ever noticed on this site:

“Due to caching, your comment may take up to 15 minutes to be displayed.”

Or were you thinking that everyone sits completely idle for 15 minutes, then everyone writes a flurry of posts in that last minute?

brian

Larry#328…my comment was complimenting CA’s work and my concern it will be compromised…your love-it-or-leave-it attitude seems contrary to the spirit of this blog: to promote skepticism and open discussion. thanks for making me feel welcome.

Larry

Or were you thinking that everyone sits completely idle for 15 minutes, then everyone writes a flurry of posts in that last minute?

I’ve never actually observed this. This would be pretty boring if we had to wait 15 minutes for a response.

Steve McIntyre

MArgin down to 500. It looks like BA will be in the lead in about 15 minutes.

MOSHPIT

re 333.

” gives brian a pink belly”

Dude, you make a fine contribution here. Don’t take your ball and go home.
Leave the ball.

Don’t take stuff on a silly thread seriously mate. .. punches you in the shoulder

MOSHPIT

RE 337. Odds?

dirac angestun

CA 15501
BA 14924

Jeremy Friesen

CA 15879
BA 15389

Larry

Now it’s up to 944. What’s going on?

Gunnar

>> I’ve never actually observed this. This would be pretty boring if we had to wait 15 minutes for a response.

Oh, there is definitely caching going on here, but it’s probably set a lot lower, like 2 minutes or something.

#316, that’s really irrelevant to my description of how it works. Yes, in reality, a page is not one post and response, but is made up of multiple post/responses. Still, the file in question is a cache of the actual data.

It’s interesting that when confronted with an unknown, humans will immediately begin to construct speculative theories about it, or accuse someone of foul play. I guess we really don’t like being ignorant about how something works, myself included.

cbone

1060…

I’m guessing that there are several thumbs on both sides of the scales here…Anyway, as they say in hollywood: Any PR is good PR. The folks who have an open mind and are willing to challenge their beliefs will find this site to be intellectually stimulating. The rest, well they will get scared off by the actual science content and instead go back to their snarky social babble sites.

Steve McIntyre

back down to 700 votes or so; then over 1000 again. I accept the comments from some computer guys that we’re seeing caching effects, but I for one don’t understand the pattern, if any, and would welcome any explanation of the seeming stutter steps.

Larry

Over 1100 now. There’s some strangeness going on.

Gunnar

I don’t see how the conclusion

The votes are not updated in real time it seems.

follows from

I updated the page and CA had 15800 votes I updated it second time 10-15 seconds later, and CA had 16100 votes. And then about 45-60 seconds later 16,500 votes.

If we had 24k hits yesterday, it’s certainly more on this last day of the poll, with conservative blogs all over pushing CA. I could well imagine 35k hits, mostly new people who are pre-disposed to vote for CA. This 35k is not spread over 24 hours, but clumped into prime browsing hours, which could be as little as 5 hours. That could be 6000/hour average. Your report of 15800 was at 1:38pm, and at 3:08, I read 16,672. That’s 872/1.5 hours. Hardly extraordinary.

pk

Holy moly, the gap is up to about 21000 now.

Jason D.

Silver lining: of the new visitors, most will never come back, but some will stay and learn something.

Kristen Byrnes

Steve,

Funny, I’m seeing you ahead by over 2,000

pk

Steve, you’ve just about reached the combined total of Pharyngula and BA from last year.

Gunnar

>> but I for one don’t understand the pattern, if any, and would welcome any explanation of the seeming stutter steps.

Maybe I’m not understanding. What is puzzling you?

Larry

You see these huge, almost instant jumps, followed by long periods of relative stability. Again, and again. It could be explained by word getting to various media and web outlets, but it sure seems like a system quirk. Maybe it’s a combination of both.

Larry

367, the CA lead jumped by 1000 in under a minute.

A Hax0r

A hax0r might be backing you up 😉 Who knows …

PS: Heard the others (BA) have backups also! It’s a backups’ battle.
Either way this blog will win! By ranking or publicity…

Rick Ballard

Maybe Kristen has a rather wide social network – this innertubes thing can be pretty amazing.

Luboš Motl

CA is ahead of BA by 2000, 35 minutes before the deadline. A possible hypothesis is that the whole awards software is fake and creates artificially high numbers and fake dramatic changes of fortune in order to attract people, voters, and advertisers. 😉

Dave Dardinger

So what is there? Something like a half hour left in the voting? Still anyone’s game the way things are going though CA was ahead 1800+ last I looked.

Larry

Hour and a half left. The CA server clock is wrong.

Kristen Byrnes

# 375 LOL Rick, 😉

Gunnar

>> so you’re sure that this is 100% an artifact of the way their site works?

Yes, considering that:

Almost all high volume sites are cached.
The vote site is, especially obvious when you look at the total. I checked 10 minutes ago, it was 45,096. I checked just now, it was 45,096. The page is cached. DB updates are occuring as people vote, just the page is held staticaly in cache for a while.
The vote changes referenced do not seem extraordinary given normal web traffic. Most people have a hard imagining how many people there are. Take off from big city (LA effect is best), and it’s amazing to see neighborhoods as far as the eye can see. And of course, the web is global. Just hope that drudge doesn’t link CA, since that would be indistinguishable from a DOS attack.

I did just notice the server time is different that EST, so my previous post should have read:

Your report of 15800 was at 1:38pm, and at 2:08, I read 16,672. That’s 872/.5 hours = 1700/hour. Hardly extraordinary.

paul

The sudden surge in votes for BA was the 2500 IPCC scientists who reached a consensus on AGW, a block vote.

Jim B

OK, some of you guys don’t understand flash, and this poll program is written in flash. Flash is cached on YOUR computer the cache is also held there under your temporary internet files. Jumps in the poll is not on their computer it’s actually on yours, that’s why it can be hacked in the first place, it’s a client side system. So no one will see the same cache and some programs will not cache the flash at all. If you have no cache, you should see a nice smooth increase in the poll, I do. I use firefox and have no problem watching vote by vote come through. So the cache your seeing is not there’s it’s yours, but YES there are bots running on this system. How do I know because I’ve made them before and can see their patterns.

This is not my doing, but I know it’s being done.

Alan Johnson

#386 Steady on, Lubos;when GMT was invented, London was where it was at, so to speak. Besides, it upset the French, so we felt obliged, really.

Steve Moore

snip

An obviously well-considered comment by someone very familiar with ClimateAudit.

Dave Dardinger

Ah, Moondancer, welcome!

I see you still don’t understand that this isn’t a right-wing site, despite it’s regular population containing a large number of right-wingers (myself included – for clarity). Steve McIntyre isn’t right-wing and neither are quite a number of others as well. Moreover, he studiously snuffs out partisan political bickering. No liberal is going to be asked to turn in his birkenstocks before entering. All that’s asked is a scientific POV.

Doug

Note to liberals who feel this is a right wing conspiracy.

It is not. I have followed it for years and send money when I can.

I am a flaming liberal. When Bush got reelected I moved to Canada. I favor drastic cuts in fossil fuel use. Not because of any proof that CO2 is going to fry us, but for all the other environmental reasons.

The CO2 issue is being worked out. I don’t think it is as bad as feared. People like Steve are working on it, and hopefuly our policies will be based on good science.

Peace to you.

Jim B

Well 30 minutes to go and CA leads by about 1500 votes. It’s a done deal. Well if this proves nothing else it proves Steve’s fans know more about computers than Bad Astronomy’s fans.

fFreddy

Well 30 minutes to go and CA leads by about 1500 votes. It’s a done deal.

As it says on the votes page:

RESULTS ARE NOT FINAL UNTIL WINNERS ARE ANNOUNCED! Each poll is checked during the voting and after polls close for excessive voting from individual machines. If excess voting is found it is noted and the votes are removed.

It ain’t over till it’s over …

Larry

I wish it was Pharyngula in second place. Then I could tell them to pray for a miracle.

Larry

That fat lady better start two-stepping it in those stelletos. I’m suspecting that both sides are tapped out.

Andy

Re #397

I am a flaming liberal.

Wonder how many tons of GHGs a flaming liberal throws off 🙂

Larry

I spoke too soon. BA just picked up about 200 in a minute.

Georg Dimitrov

#406,

Don’t be afraid to win. It’s quite clear that there is also a huge amount of ballot stuffing for BA, and my guess is that both balance out.

dirac angestun

CA 19750
BA 18631

with about 12 minutes to go.

Larry

As I said, Steve’s prediction of a total of over 50,000 has already happened, but unless something changes, it’s unlikely CA will make it to 20,000, and almost impossible for BA to do so.

Mot Normalt

True tele-connections!

Dennys McAullay

Also votes for JunkScience stopped growing a while ago, allowing more growth to CA. It’s not unlike Pharyngula votes going to BA.

Philip_B

You are missing the obvious here. Multiple positive feedbacks are causing runaway voting.

That comment is not as ironic as it may sound.

Otherwise, I’m not seeing any caching at all. I’d say the observed caching is due to internet providers and not from the source. BTW, I started seeing surges in CA votes relative to BA, 24 hours ago, then BA would surge back. As I said, multiple positive feedbacks.

Larry

My bad again. CA just passed 20,000.

Predicting the future is hard.

Jim B

20,000!

Jeremy Friesen

Larry said:

As I said, Steve’s prediction of a total of over 50,000 has already happened, but unless something changes, it’s unlikely CA will make it to 20,000, and almost impossible for BA to do so.

To my delight, you were mistaken 🙂
CA 20098
BA 18806

5 minutes to go.

Larry

I’m calling it for CA.

tpguydk

#398 yeah i’m a flaming liberal too…but I think steve’s in the right about this and Mann et al are not..

Jason C

#373 made a point.
Even if a bot voted for you you can still claim that a climate blog was oppressed by haters as removed votes based on logs will always be under doubt.
Weblogz should make the script safe server-side or witness what just happened.

Regards.

cbone

ITS OVER!

Final, unoffical, tally:

CA:20,131
BA:18,894

Ivan

Victory!

MrPete

ok, this is VERY strange.

CA stopped at 20131.
BA continues to climb. I saw 18897… then 18902, 18903, 18904…

something is leaking?!!

Bob Thompson

Congratulations!! Steve’s the winner by over 1200 votes

Jim B

Well this is interesting polls have been closed now for 3 minutes and BA STILL has votes coming in about 1 every 10 seconds like clock work, now that’s a bit for ya! It some how is keeping the session open pretty nice.

The Fat Lady

It’s oooooverrrrrrrr
It’s oooooverrrrrr
It’s ooooooooverrrrrrr

Anthony Watts

Ok everybody, lets get back to work, shows over.

Preliminary congratulations to Steve and the entire CA community, pending adjustment of the “raw” data. 😉

Larry

443, I just noticed that. Either their clock is way off, or they have some very strange database dynamics. We need to keep refreshing until the numbers stop changing.

Jim B

My guess the there is a macro going it opened 50-100 screens and is now doing the voting the macro to vote, is you wanted any more oblivious proof of a bot at work just continue to watch BA’s count go up, TURN YOUR BOT OFF!

That’s funny!

Larry

Uhh, guys. That’s correct. CA is frozen at 20,132, and BA keeps clicking up. This is starting to look like not a glitch.

Robert Dennis

Kristen Byrnes posted a nice note on the BA blog at the close of the voting. Her decency and sense of perspective are in marked contrast to PZ Myers’s comments on the same thread.

Robinson

Don’t panic Larry, it’s caching behaviour 😉

Sam Urbinto

MOSH!!!! lol, can we talk about the ice melting and solar power….

“I’m wondering if the poll taking database system itself is introducing some lags”
I was thinking that the entire time. It depends on the type of data storage they use (it could be flat files, right?) and how often they do a query on it and how their voting software itself interacts with it. And traffic levels and blah blah blah. Whatever.

All, we’re just often commenting on the possibilities. Is it possible it’s normal? Is it possible something strange is going on? Give the people guessing at causes a break. You can bring up a point without believing it yourself, as a matter of discussion, too. “It could rain tomorrow” doesn’t mean I think it will. Not everyone here is as even-handed or open-minded as everyone else, too. Chill out, this is a lighthearted change of pace.

Proxy war indeed.

I think it’s just the way the system happens to be working, myself. Bursty.

Something Anthony said:

“Benthic Bacteria Emit Noxious Gases”
Spencer sent that to Limbaugh and he misinterpreted it as real (until Spencer probably got ahold of RL during the break and said no no no I sent that to you telling you it was a hoax (something like ‘I find it hard to think somebody went to all that trouble for a hoax’ which RL took as being ‘it has to be real because nobody would go to all that trouble’ It should have been “This is a hoax, and I find it hard anyone would go to all this trouble for one!”) So that’s just sometimes how things get “usufruct up”

joeschmo1of3

CA has been going up as well. The votes are still tallying. The 1200 vote lead seems to be holding.

Will C.

CA now at 20242, so we’re still climbing too.

WOOHOO!!

TerryB

To all at CA (again -I first posted this on the wrong thread,what a div!)

10.20pm UK time and its looking close now – but I think you might just pip it.

I’ve been a non-contributing lurker here for ages now and most who’ve contributed – whatever side they’re on – should be congratulated for having an open, honest exchange of views to the benefit of science.

Not quite the Nobel that Gore-Al (Jor El) got, but still, well deserved.

Lets hope this makes Steve’s excellent work become more prominent and helps bring about more honest and diligent research when undertaking poxy – sorry, proxy – reconstructions with fancy al-gore-rithms. Sorry, couldn’t resist that.

Whatever the final result, well done.

Dave Dardinger

It could be that it comes from various ISPs caching some page updates and their eventually being released with datestamps indicating they are valid. Much like the situation with Military ballots in Florida in the US 2000 Presidential Election.

Of course we could start a conspiracy theory that it comes from ChiCom operatives who intercept messages to the US and search them for illegal contents.

Larry

If you want homebuilt geothermal, better plan on finding a titanium heat exchanger at some surplus yard, and plan on cleaning the scale out of it regularly. Oh, and you’d better live in geyser country, too.

pk

I just voted on one that I hadn’t previously voted on and the votes went up by one. They must be certifying the votes based on the date/time stamp in the database.

Judith Curry

WOW. I’ve been following the weblog contest, and even voted. Congrats to Steve and the Climateauditors! CA definitely deserved to win, this is by far the most dynamic science blog on the web.

Steve: Thanks, Judith.

Larry

I’m not saying that anybody got robbed, but I’m just puzzled that the votes keep dribbling over an hour after it’s over. The internet isn’t that slow.

Jason C

People their voting script is flawed and bots can still vote.
Locking is only done client-side (in Flash) but the polls are still opened to low-level voting.

http://2007.weblogawards.org/get_poll.php?poll_id=117 -> Voting still open.

IL

As a professional scientist I also heartily congratulate Steve and the many excellent contributers. I hope the final announcement will reflect the final poll position – thoroughly deserved in my view. I wish more scientists would look at what is done here and respond like Judith Curry and interact in a mature, dare I say scientific way, rather than the extreme and hateful vitriol characterized by Myers et al.
Long may you help keep climate science honest Steve.

Bob Koss

They must still be count the absentee ballots.

steven mosher

481. I put a T in every vote I cast. There are no absent ts.

“will pun for money”

Sam Urbinto

Judith Curry rocks! Nice post over there on Phar.

Numbers just a bit ago:
CA: 20,631
BA: 19,219

Just refreshed:
CA: 20,632
BA: 20,673

WTF? I can see them going down if votes are considered bogus, but going up more than 1400 within minutes? (actually, it’s pretty much a tie at 37.5/37.6%) So far, BA keeps going up 1 every refresh, CA is not moving. Something is weird. I think there is a back-door; or did they throw out votes they should have kept and addthem back? WTF. Seems bogus.

Anyway, the “Best Blog” is 24,908. The next closest is at 7,544, so both CA and BA beat them resoundingly. OTher than that, well yeah.

We already knew Steve was going to be gracious regardless.

Sam Urbinto

We should wait until the winners are annouced. But come on now, regardless if the poll is bogus or not, or meaningful or not, or important or not, or the greatest thing in the known universe; how does a blog jump 1400 votes, hours after the voting is over?

I always thought it was just the way the voting was going, but that was when it was open. Now that it’s closed, I am not so able to think it’s just by chance…..

I wonder what’s going on.

Philip_B

Unbelievable, BA has just past CA in votes. It’s about 50 votes ahead.

Larry

489, That’s what I said. Welcome to King county. Oh, looky there! Another box of uncounted ballots!

Philip_B

I think it more likely someone has found a way around their ballot closure deadline. I expect the Weblog site will revise BA’s numbers down.

Larry

493, 494, that’s the thing. If this happened when voting was active, it could be a glitch. 3.5 hours after??? That’s no glitch. It’s an exploit.

Larry

Paul, the numbers Steve put up in the “left/right” thread are pretty close.

Larry

Imagine that. The numbers aren’t increasing any more. Just like in King county.

Larry

What West Florida fiasco? If you want monkey business, try WA state governor, 2004. There’s a case study in monkey business.

Dennis Wingo

I have a screen capture right after the Polls closed at 2:00 pm Las Vegas time that showed CA ahead and the latest one that shows CA behind and an addition of almost 2,000 votes for BA.

steven mosher

cached pages pages with altered time stamps?

Richard Sharpe

Votes are still trickling in … over four hours after the cutoff …

Jan Pompe

I’m wondering how in the past hour some 3 hours after voting has closed BA pulled ahead from near 1500 behind.

Jason L

EST…………..CA……….BA……Delta
(The previous post had a typo. The first 4:44 PM should have been 4:16 PM).

3:59 PM …18828….17294…1534
4:16 PM …19259….17798…1461
4:44 PM …19735….18575…1160
4:53 PM …19783….18755…1028
5:03 PM …20131….18907…1224
5:17 PM …20242….18983…1259
5:31 PM …20247….19060…1187
5:54 PM …20555….19144…1411
6:22 PM …20613….19175…1438
6:34 PM …20615….19181…1434
8:20 PM …20634….20681….-47

Larry

Ok. That makes me feel better that it’s all hosed up all around. No problem…

theduke

Reverted: so which horse are you rooting for in the home stretch?

Kristen Byrnes

I recieved emails from 2 friends in Australia who said they voted 3 – 4 hours after polls closed and the totals still went up. I voted in 2 other categories a few minutes ago and kept track of the numbers. We’ll see what’s up after they announce the results.

MrPete

Tech support confirms:

* Cached flash “voting tool” still gets “votes” in — and they show up. BUT they will not be counted.
* They are adamant that post-close votes will not count
* They’re still working on validating results

I’m sure they’ve learned a few lessons for next year!

dirac angestun

From http://2007.weblogawards.org/polls/best-science-blog-1.php

“Each poll is checked during the voting and after polls close for excessive voting from individual machines. If excess voting is found it is noted and the votes are removed.”

My reading of that is that votes cast after polls close may be recorded, but will be removed from the final tally. That should put CA back in front by over 1200 votes.

Steve McIntyre

I’ve just come back from squash league. When I left, the polls were closed and we were 1400 votes in front. Now we’re 50 votes behind. Sort of like a MWP temperature reconstruction.

rk

Regardless of who wins, I think this is significant. 20K votes is way more than very famous blogs got in their respective categories. The fact that blogs that were anti-CA referred people to vote against CA and still it is a photo-finish. (yes, I know that pro-CA blogs referred too)

If the #’s hold up, it is a good moment for CA/Steve Mc. (ok, not as good as the the GISS correction or the Barton hearings, but hey)

Congrat’s to everyone on this blog.

Luboš Motl

After reading the comment #512 with the numbers, I probably won’t believe an explanation of the weird evolution that assumes that all numbers are refreshed simultaneously and everything is clean.

It’s not just the delta that is changing weirdly. More importantly, the ratio of rates of votes is just completely crazy. For example, between 6.34pm and 8.20pm, CA receives less than 20 votes while BA receives about 1500.

The ratio is about 75. There can be groups of people that vote after an e-mail or a web comment. But I just find it unlikely that all these people are so synchronized that every interval has a pretty clear winner, by orders of magnitude.

There’s no proof but it is hard to assign too much importance to such data.

dirac angestun

Then many happy returns as of yesterday.

Jim B

Seriously though they should release the IP’s of the people or organizations who gamed the system. I for one would like to know who they are.

Jason L

For what it’s worth, someone at BA wrote:

Mike on 08 Nov 2007 at 11:04 pm

The site administrator said that they did not program the site to stop counting votes after the polls closed but that the votes that came in after 2 PM Vegas time would be corrected back. CA won according to the media outlet at the convention where the results were announced live.

I wouldn’t trust any numbers until the official count comes out.

Anthony Watts

I can still vote! See my post at http://www.wattsupwiththat.com

It appears Firefox has the ability to vote where IE doesn’t.

Gunnar

>> Jumps in the poll is not on their computer it’s actually on yours, that’s why it can be hacked in the first place, it’s a client side system.

Jim B, Of course, browsers cache as well, but that doesn’t mean that the server isn’t cached. In my post where I was seeing the total votes not change for 5 minutes, I was doing a Shift-refresh, which certainly avoids local cache and requests a new page from the server. You’re confused if you think that client side caching, or flash is at all related to “why it can be hacked”. The flash is only animating the poll results client side. The actual voting was just a radio button post back.

I’m not saying that poll results were not hacked, but I think it has nothing to do with flash. One way to hack the results would be to put a vote submittal in the page load of a high volume site. That way, everyone who simply viewed Daily Kos or Malkin would vote without knowing it. That’s why internet polls are silly.

MrPete

Stats people, how about solving a fun problem?
Seems to me we have expertise on tap here for an interesting exercise in computer fraud forensics.
PLEASE NOTE: this is for fun. Yes, I care about the truth. I don’t honestly care how the final tally turns out — the goal of better publicity is already achieved!
THE SETUP
Suppose you had a time series of data — votes from potentially related data sources (a computer cluster for example). You have time of day and exact sources. You suspect some of the sources may be controlled in an automated way, to generate votes for a (set of) targets.
A simplified version (with no variance)… here are times, sources, targets:
0, x1, a
6, x2, a
10, y1, b
12, y2, a
18, x3, a
20, x4, b
etc…
In this case, I’ve created a pattern with 6 and 10 second voting intervals, and no variation. In the real (messy) data set, I’m sure there is significant variability in the patterns, simply due to internet/computer delays. Even if nobody programmed in any sophistication of random timing.
QUESTION
How would you analyze the data set to tease out any fraud, with good confidence?
I’m guessing that a graph of vote intervals over time, possibly accompanied by fourier analysis, might be informative.
I’m interested in comments on the problem, particularly from our statistics experts!

Gunnar

>> I would dismisse the caching theory, unless some very bad rpograming is on the server side.

As I’ve explained, caching is not a theory, it’s a fact. I’m confused as to what you have observed that you feel is strange and needs explaining.

>> Like considering client time instead of server time to determine if the vote was placed before closig time, all you have to do to hack the voting would be change my PC time and place one vote for each competitin day.

Client time is inconceivable. It doesn’t look like they had a way of implementing “closing time”. They just took the link down, then realized the page was still there. They can always get the result at closing time with a SQL where clause.

>> I[t] may be that the server registers votes in queues that are processed over time, and the display results are outdated by hours, or even days. But for the amount of votes in question I would say that would be out of proportion.

Also inconceivable. There is absolutely no reason to make up obscure theories like this.

Mr Pete,

Even if there was a web farm, the time recorded would be the database server time, so there is no problem with that. There is no reason to resort to statistics to determine something which is discoverable by other more direct means.

L Nettles

The vote page currently shows BA ahead of CA. 20683 to 20638

Steve McIntyre

Hanging chads and everything. They said that the contest would be fun, but who’d have expected this much fun. The link in #546 is interesting. My audit instincts suggest to me that the same mechanism used in the post-closing voting was probably being used pre-closing. In this respect, the asymmetry between CA and BA post-closing votes (1400+ votes) is shall-we-say statistically significant.

Bob Koss

Al Fin has some complimentary things to say regarding Steve and the awards.

Larry

Wasn’t it Gore who wanted to do elections online? Using his innertube thingy? He’d have still lost.

Larry

If nothing else, this has been a really interesting case study in online security. It doesn’t look like they took more than token steps against a hack. And with the front door wide open, and the screen door missing the screen, they just helped themselves.

I’ll remember this the next time I have to worry about kiddies. The whole internet setup is just begging for a breakin. Looking at the link in 546, it was just too easy.

MrPete

Bernie – exactly. I suspect the poll managers are pretty upset, and are digging hard into the issue. It never feels good to be hacked.

My suggestion for analysis of voting patterns is based on the idea that those who gamed the system simulated use of a large set of machines, all behind a small number of IP addresses. The hard part would be to distinguish 1000 legitimate votes from behind the IBM NAT’d firewall, and 1000 fraudulent votes from behind the XYZ.com NAT’d firewall.

Bernie

I think they, the organizers, have to look at a number of the other voting categories as well. A quick look at the numbers for Fark.com (the best online community site) suggests similar issues – given that Fark received more votes than any other site in the competition. The pattern of votes across all the comptitions certainly suggests that a few of the competitions received inordinate attention and the attention of hackers..

Sam Urbinto

Gee. Over at Phary, PZ says (in hello stan palmer) about Steve’s answer to the FAQ question “Does your work disprove global warming?” this way:

That isn’t a position. That’s obfuscatory bafflegab calculated to encourage deniability.

And of course refers to CA people over there with the nice moniker of “Rats with their moldy flecks of rotting garbage.”

Feel the love!

Larry

557, I think they have other problems, as well. The results for best blog were pretty lopsided, and the “winner” was a rather lame, strange blog that nobody’s ever heard of.

Larry

This is semi-hilarious: Steve closed the “left/right” thread to comments, and then some ‘bot added a comment after the close. How could that possibly happen?

StuartR

Congratulations to both CA and BA. the Weblog Award for Science is offically tied for first place. Both get 20,000 votes.

Larry

Sez who? Is that official?

BarryW

Re 567

That’s what it says on the website. My guess is that the voting was so badly hacked that they couldn’t unravel the good votes from the bad for both sites, so they declared it a tie. “Chad” would be proud.

L Nettles

The tie breaker should be which site has posted the most science since the end of the count.

Kristen Byrnes

A tie? That’s silly.

steven mosher

I’m just glad I have a really hot looking cousin. Since a tie is like kissing your cousin.

Larry

Steve agreed to it. So be it. I guess that means Steve is more like Nixon than Gore.

Chris Christner

It’s officially a tie and now I’m concerned about the quantum physics aspect of this.

You’ve got a pro-science site (CA) combined with an anti-antiscience site (BA), doesn’t that mean they’ll both go into a cat state? Or will they simply obliterate each other in a burst of radiant energy?

Might be wise to wear safety goggles when you visit.

steven mosher

This will be an intersting test.

Since it is a tie both sites lay claim to science blog of the year.

Both sites have a traffic pattern PRIOR to the blog awards.

They tied.

Which blog will gain more traction with new users?

that is the tie breaker.

6 months down the road who has a bigger pagehits anomaly.

Ah yes, see if CA page hits can create a hockey stick

bernarda

Several hours ago the polls closed and the results gave Bad Astronomy winning by a few dozen votes. Both of you had between 6000 and 7000 votes.

Now Webblogs lists 20,000 about for each. Why does Webblogs want you to win?

L Nettles

Bernarda wrote:

Several hours ago the polls closed and the results gave Bad Astronomy winning by a few dozen votes. Both of you had between 6000 and 7000 votes.

Now Webblogs lists 20,000 about for each. Why does Webblogs want you to win?

Huh? On what planet did that happen?

Philip_B

Gunnar, take a look at the client javascript. It looks like it’s checking the client system’s clock time against a cookie to determine if you voted in the last 24 hours. A ten year old could hack around that.

pk

bernarda,

The polls closed yesterday at 5:00pm EST. When they closed the votes were CA ~20,200 and BA ~19,000. In the hours after closing, they went to CA ~20,634 and BA ~20,681. The organizers declared a tie and 20,000 votes for each. I’m not sure where you get the 6000 something votes.

Kristen Byrnes

Paul,

He’s too old for your cousin. Besides, he mixed his viagra and rogain and now he looks like Don King. LOLOLOLOL (I’ve been sooooo looking to get back at him for that data chart stunt he pulled)

Kristen Byrnes

Still nothing from Greece (sigh) Not that I’ve had much time with 6 hrs of homework per night.

steven mosher

RE 593. Ah greece. I loved the airport there. It was All I got to see.

Good luck kiddo. Hey my son had to write a response to AIT the other day.
I gave him your site and Surfacestations. He says thanks.

Philip_B

Kevin Aylward posted this comment in response to what I thought was a rather biased piece at Wired News.

conservative blogs have been winning the Weblog Awards for years.

That was true in 2005, but things have changed substantially since then. Liberal (or left leaning) sites did very, very well in 2006 – snapping up most of the political category wins. This year it was fairly balanced and the non-political folks (Post Secret and FARK) won big against political (left and right) sites.

Coverage and links from non-political portions of the interenet are much higher in the last few years because people have seen that the contest is open, fair, and fun.

Wired Gadget blog interviewed me at length this morning about the last minute poll issues. Expect a story this weekend.

reverted

#576: “Since the votes at the time of closing clearly were in favour of CA a tie seems unfair and would seem to reward cheating.”

Since we all know that there was very strange behavior for CA vote-counts *before* polls closed, there is no particularly good reason to cite what you just did as if it is a legitimate excuse for claiming a win for CA. (The fact is, we just don’t know.)

In summary, it appears that:
1) BOTH sides cheated by messing with the voting system;
2) Because of this, it could conceivably be argued that NEITHER side really “deserves” to win; and, finally,
3) All this monkey-business apparently wreaked so much havoc that they were unable to reliably discern between legitimate and bogus votes in order to announce a definite winner.

Now, I don’t think any of the actual blog owners condoned cheating. But, the fact remains that it apparently happened on both sides.

So, a tie seems to be the most generally satisfactory way to proceed from here. Be happy for the win, even if it is a tie. Stop griping.

Congrats to both CA and BAB!

(Yay @ #598, too!)

Steve Moore

Stop griping.

I would quote Lambert (“I like griping”) but we all know what happened to her.

Sam Urbinto

reverted, I think you’re jumping to conclusions on this. There are a lot of folks out there on the Internet in general, and in the blogosphere in particular, that have never been to either site, and have both the means and the will to tickle the system to change the numbers for either side. It’s fairly clear that some exploitation went on, but the only obvious strange thing (which could have been a normal artifact of how they have the voting system set up, don’t forget) was the jump of 1400 votes (rather than the removal of bad votes down). I would be surprised to find that hardly anyone, if anyone, that frequents either site would have either the means or the will to mess with the votes.

You have to remember that up until now, many from either “side” knew of the other blogs. There are a lot of blogs out there. Once some people that are in the circles that frequent multiple blogs learned of the others, it became less of a matter of denizens that frequent the various ones on that list voting for their favorite out of x blogs they know, but that of people voting for the blog they perceived as most matching their views (and that I would say they only had the impression of what they had been told or had only seen on the surface). It changed the dynamic so much, the latter stages were nothing the same as the earlier.

That said, having read Steve for quite a while, and seeing the general way that Phil acted and what he wrote, I would say that neither of the two did OR condoned any cheating. Read what they both write, and take both at face value. I would also say the majority of regulars at either (or both) sites would do things like that. I myself don’t think this is a case of “one blog cheated better than the other” but rather a case of outsiders deciding to influence things “their way”. How that impacted things overall? I don’t think anyone knows, nor do I think it’s worth trying to find out.

I believe the people that ran the poll didn’t have an idea anything like this would happen, and that it’s better to just give it a tie and drop the issue as both not important, and as a messed up situation better dropped than obsessed about.

I doubt any of the people normally on either/both blogs cheated, and I’m almost positive neither operator did. It simply wasn’t there. The only person that seems to have become upset at this at all is, well, you can tell whom it was.

windansea

3) All this monkey-business apparently wreaked so much havoc that they were unable to reliably discern between legitimate and bogus votes in order to announce a definite winner.

hey the guys who design the smartest vote bots deserve to win the best science blog

just kidding 🙂

Steve McIntyre

while the idea of a tie had been presented by third parties, I proposed an agreed tie to Plait last night and then we both presented the idea to relieved Weblog Awards organizers today. (They still have an outstanding brawl in the Techonology Blog category). The purpose of the contest was to be fun and to draw attention to blogs; it accomplished that. Yes, CA was 1400ahead at 5 pm and yes, bots hacked into the system after 5 pm and altered the votes in favor of BA. Those votes would probably be easy to deal with. But once you started down that road, you’d have to recognize that bots had been active in the votes prior to 5 pm. Sorting those out would be a herculean task for the organizers, and, in my opinion, an unreasonable burden for them given the purpose and scope of the contest. Based on post-closing bot activity, I could guess that pre-closing bot activity would be proportional to post-closing bot activity where the proportion was known but it’s just a guess.

We can take satisfaction in our award and I’ll be proud to display the logo. But readers should also recognize that it doesn’t do CA any harm if Phil Plait also has an awards tag.

John S

# 599

You left out one other possibility. Someone at Weblog Awards cheated and was caught then a tie was offered to quiet things down. I watched the voting all day yesterday. Nothing seemed inconsistent with the tallies throughout the day to me. Every time a new blog or radio talk show endorsed either side, the numbers jumped for that side, I can understand the numbers jumping by hundreds or thousands in just one or two minutes after a blog was endorsed by a radio show. Communication with the Weblog Awards server was direct, so when you voted the tally went up by 1. That is why there was no caching of large numbers of votes. All you had to do was use a different link (ie: the one at Anthony Watts site or the one at CA) to look at the tallies and you got the fresh tally. What was really telling was when the administrators of the Weblog Awards Forum said they were aware that voting was still occuring within an hour of the polls closing, then 2 hours and 20 minutes later the tally for BA jumped by over 1400 votes in less than one minute. It wasn’t until shortly after that when the numbers quit rising. That’s not a bot or a cache problem, it is someone at Weblog Awards fixing the numbers. I would give long odds that the person there was found out and rather than admitting to what happened, the awards people gave first place to both blogs in the hopes that things would quiet down. They may even have used the arm twister “both sides were cheating” to help close the deal. But then, neither CA or BA has commented on the events leading up to or the details of the deal they agreed to yet so we cannot be sure until we hear from them.

Steve: I don;t think that this hypothesis makes any sense. Weblog Awards said:

Someone figured out how to cache a poll and make it keep voting after the poll had been turned off for everyone else. Sean was eventually able to fix that but it took several hours so one (or more) people had access to voting while everyone else was shut out.

Yeah, they had a problem with their set up bu there’s no reason to think that it was intentional/

John S

… and Steve spoke as I typed.

Larry

Someone figured out how to cache a poll and make it keep voting after the poll had been turned off for everyone else.

I’m trying to make some sense out of just exactly what it means to “cache a poll”. I think that’s a convoluted way of saying that someone intercepted their scripts and diddled them. That seems like a really funny way to say that.

Larry

Is Johnny Cash a pole?

John S

Steve’s note in # 609:
“Yeah, they had a problem with their set up bu there’s no reason to think that it was intentional/”

When their site links a blog that tells it’s readers to vote once or twice for BA (ie; cheat), I can’t trust them as far as I can throw them.

mccall

Let’s try this from a legal perspective:
did either of the blog owners leave fingerprints at a “jury tampering” crime scene?
:
:
:
:
:
:
:
:
:
:
:
http://www.huffingtonpost.com/phil-plait/weblog-awards-2007-vote-_b_71762.html

5ubliminal

#609, #617 – You’re way off! Don’t blame them as their only problem was raw incompetence.
Let’s say an outsider (not a reader of any of the two blogs) gamed the system with two thoughts in mind:

1. WLA created a voting script so weak it had to be revealed so people know the real winners might not have been real.
2. Science category was chosen as all the other sites clearly ganged up on the CA blog (regardless of its opinions which I will not question).
And the worst thing is that even the organizers advised voting for BA against CA which is untolerable from a hacker’s point of view.

The hacker came into play about 8 hours before end of competiton kept BA and CA close during competition but favoured CA the entire time.
Actually their script was even worse than the hacker first imagined. It allowed voting after time elapsed and that had to be pointed out hence the final pack of 1400 votes for BA (4 hours after vote closing) came as fatality and changed the winner after closing time.
They were cast because the hacker wanted a tie between the competitors and a draw as a final result (it was not the competitors’ mistake / problem the script was complete crap).

There may have been other gaming the system (little league – with few votes) but only one brought real publicity to both BA and CA blogs. Take advantage of this publicity and give quality to your (new) reader base which should have grown as a result of the controversy.

Regards, 5ubliminal

Duane Johnson

Congratulations Steve M.

I posted on the BA sight last night that I felt declaring a tie would be appropriate. I’m a long-time CA lurker(seldom poster), who thinks that if Plait had more familiarity with the CA site and with the issues involved, the skepticism he shows on extraordinary science claims would cause him to view CA more favorably, if not become an actual supporter of CA. Wouldn’t it be good if that happened?

Andy

Belated Happy Birthday, Steve. It appears as though it was on the 7th, which means we share it, although separated by a few years.

5ubliminal

#626

The only way to prevent this would be to limit the number of votes coming from one IP during a certain period of time.
Assuming that one IP has a network behind is politically correct (as you allow let’s say all exmployees to vote) but you no longer have security on the poll. Zip!

Captchas shows people lazyness and not every1 would have voted as they did.
It’s a no way out situation unless you consider one IP = one Vote in 24 hours.

John F. Pittman

Happy belated birthday.

Gunnar

Ok, it won’t let me post the SQL that checks for 24 hours ago, but you get the picture, as if anyone cares.

MrPete

Email verification has zero ability to eliminate duplicate voting from fraudsters. It’s actually a rather hard problem.

M. Jeff

Since the Best Science Blog competition, the level of civility here seems to have deteriorated.

Gunnar

>> They were put in SQL for a reason.

Oh really? Philip_B, actually, I always use ansi SQL when I can. A couple of years ago, I tried to come up with a standard way to do what I did above with one SQL statement. I’d settle for 2 though, but couldn’t find it. AFAIK, you can’t use a correlated subquery with an insert statement, but I could be wrong, and would love to find out if it’s possible.

Can you try to get your syntax correct at this site: http://developer.mimer.com/validator/parser200x/index.tml

If there is no ansi SQL way to do it, then I’ll continue to use this T-SQL when it’s SQL Server. PL is Oracle.

Subliminal, spell much? Not sure why you launched into ad-hominem. MJW asked for a more robust solution. Since you’re psycho analyzing me, I would speculate that you feel really insecure about the growing prominence of SQL Server & ASP.NET, so you lash out against anyone who doesn’t. #2, the fact that WLA uses PHP, a technology from 1994, is irrelevant, in the context of the question I was answering. #3, email verification was an alternate method, for those that are locked out because someone in their office has already voted. Given that it’s a measure of loyalty, all the better.

>> Email verification has zero ability to eliminate duplicate voting from fraudsters.

MrPete, care to elaborate?

lucia

@Larry–
Given the massive comment deletion between 632 and 633, I don’t think the Gunnar/ 5obliminal conversation is really going to get off the ground. (I have no issues with comment deletion; but who wants to repost explanations about why Gunnar suggests really is not enough or discuss the whole spelling issue? )

Steve: I don’t mind some banter from people who are online, but it doesn’t need to be preserved, so don’t expect it to have a long life expectancy.

lucia

@Steve– No problem. Things had definitely gone well off track! But, with respect to Gunnar’s comment in 637, 5ubliminal had admitted he was argumentative.

5ubliminal

True Philip_B . SQL Server coders have a tendency to scratch the back of their heads taking the long way (you know the route).
Standard SQL rules. KISS (Keep It Simple Gunnar) is the concept you should go by. And join a webdev forum for the show-off, not here. Let those in the field have all the fun!

Gunnar … I can’t spell (I’m not EN speaker) … but you can’t read! MJW says :

As someone who knows next to nothing about web programming

So he did not ask for a solution but a short explanation for a non-tech individual. :)__)

PS: Gunnar … you’re out!

steven mosher

I think the Mohel took more than a snip from Gunnar and Flub blinal.

oar knot

Philip_B

Gunnar, the link below from Oracle explains correlated subqueries with examples.

http://download-east.oracle.com/docs/cd/B14117_01/server.101/b10759/queries007.htm

And get yourself a copy of Date’s book. It really is a gem of clarity.

MrPete

Why is email verification useless for prevention of duplicate voting?

Because anyone with control over an email server can process an essentially infinite number of email addresses. Think of it qas reverse phishing.

A non-techie explanation: a common form of spammer attack is the “dictionary” attack: attempt to send emails to DictionaryWord@somwhere.com, where DictionaryName is every word in a dictionary.

I’ve seen blunt attacks: one per second from the same source, in alphabetical order. I’ve seen sophisticated versions: randomized, with random timing, from a dozen sources… but over time it becomes obvious [if you are looking] that the dozen sources are all pulling from a single dictionary.

Similar tools can be used to generate a million “email validated” votes — (automated, and random reply time) response to each and every verification email.

If a computer can be used to emulate ELIZA (sorry for the four decade old example!), this is a nit.

A good voting system must distinguish between votes from 1,000 people at ibm.com and 1,000 “people” at FakeBigCompany.com, both of which are behind good firewalls — and that’s a tough order.

Gunnar

#648, Philip_B, no problem, and thanks for responding. I could have been wrong. Now, we both know for sure, and I know more about your intellectual honesty.

#649, MrPete, great points all. A capcha in the process might help, but there is always a way around everything, but there is a practical limit to what someone will do to skew a meaningless web poll.

#650 >> Our tendency in almost all situations is to minimize our sense of the uncertainty levels

So true, and to maximize our sense of efficacy. The end result is: we’re certain that disaster is coming, and we know we did it. Human conceit.

Greg

Enjoyed it!
Good review on the whole poll ordeal:
http://www.inoculatedmind.com/2007/11/12/monday-madness-stop-suppressing-the-vote/