2011-04-19 15:45:53Gilles is spamming now; really, it's time for him to go
Daniel Bailey
Daniel Bailey
yooper49855@hotmail...
97.83.150.37

Deleted Comments

  1. Gilles at 15:41 PM on 19 April 2011 (Email commenter)
    what's the weather in australia BTW?
    Deleted by Daniel Bailey (Spamming)
    | Edit
  2. Gilles at 15:41 PM on 19 April 2011 (Email commenter)
    hope your employer doesn't mind if you loose time to track my posts on this forum
    Deleted by Daniel Bailey (Spamming)
    | Edit
  3. Gilles at 15:40 PM on 19 April 2011 (Email commenter)
    I'm having breakfast, it's a funny way to start the day :)
    Deleted by Daniel Bailey (Spamming)
    | Edit
  4. Gilles at 15:39 PM on 19 April 2011 (Email commenter)
    yes, you're definitely getting nervous :)
    Deleted by Daniel Bailey (Spamming)
    | Edit
  5. Gilles at 15:36 PM on 19 April 2011 (Email commenter)
    "but they are rapidly exhausting their bag of tricks to hide the incline.
    "

    The incline ? you mean an acceleration ? how much is the acceleration following you ?
    Deleted by Daniel Bailey (Spamming)
    | Edit
  6. Gilles at 15:35 PM on 19 April 2011 (Email commenter)
    are you getting nervous, moderators :) ?
    Deleted by Daniel Bailey (Spamming)
    | Edit
  7. Gilles at 15:29 PM on 19 April 2011 (Email commenter)

    Deleted by Daniel Bailey (Empty Comment)
    | Edit
  8. Gilles at 15:29 PM on 19 April 2011 (Email commenter)
    how about this one ?



    I have a question about this one : how can it be that the real temperature curve is strictly confined just in the intervals of models , just hitting precisely the border on each side, without crossing it , and with no spare room around it - is it possible, statistically speaking, without a careful selection of "the set of models " that have been kept to plot the curves ?
    Deleted by Daniel Bailey (Implied allegation of fraud, dishonest and impropriety)
    | Edit
  9. Gilles at 15:21 PM on 19 April 2011 (Email commenter)
    waaaooouh ... was my whole post really OT ? :/ then shouldn't you also suppress the answers ?

    Well i'm really sorry to be so powerful to spoil your favorite websites, but aren't they precisely devoted to convincing people like me ? if this site is supposed to debunk skeptical arguments, then you should gratefully welcome me to provide so much material to debunk and feed your site, shouldn't you ?

    now you may be surprised that I seem to be skeptical on every topics, but if you think a little bit about it, it is not so surprising. For if there were one single very strong , unavoidable argument for the AGW theory , kind of a Halley's comet, the precession of Mercury perihelia , the Morley -Michelson's experiment, something like that, this would for sure have been presented again and again, make all covers of all magazines, and heralded in IPCC reports. Obviously there is no such thing- just a bunch of indications. So it's quite normal that on each topics, doubts can arise. I'm just here to remind it. A bunch of indications is not a strict evidence - there was a bunch of indications that the sun and the planets revolved around the Earth after all - this was perfectly reasonable , and even, not so wrong even now (we're perfectly entitled to describe the whole motion of the universe in a free fall reference frame centered on the Earth in general relativity....); the only thing is that it is much more useful to describe the motion of planets including the Earth around the sun - the only criterion is "usefulness".

    So I'm nothing against any climate theory - I'm just asking if its usefulness has been proved, practically speaking. For the moment, I don't really see it.
    Deleted by Daniel Bailey (Moderation Complaint AND Trolling: The Gruesome Twosome)
    | Edit
  10. Gilles at 14:30 PM on 19 April 2011 (Email commenter)
    Muoncounter : as i said , numerical models are generally speaking unreliable for predicting internal variability - since this variability stems from highly non linear coupled phenomena which are notoriously difficult to catch precisely. A linear model wouldn't have any unforced variability at all. Of course numerical models are non linear and have their own internal variability, but there is no reason a priori to think it matches exactly the real one (which is the essence of the above paper).

    So arguing that models reproduce the overall curve with the forcing, but not without, relies on the assumption that models would have correctly predicted the evolution without forcings - meaning actually that you assume that their internal variability is correct , and that for instance "not showing long term unforced variability" is a reliable criterion that they really don't exist - which is not true as I said. So you can't use it to exclude other possibilities - that the variance due to unforced variability is larger, and consequently that of forcings smaller, since variances must add.

    In more technical terms, the models have not proved to simulate correctly the power spectral density at low frequency, so you can't argue that there is a cut-off in their output to claim that this cut-off is real - it has just not be tested properly against observations.
    Deleted by Daniel Bailey (Off-topic)
    | Edit
  11. Gilles at 14:11 PM on 19 April 2011 (Email commenter)
    so as far as I can judge, following your answers, the models are "good" because data lie within the uncertainty range of a set of predictions. This is interesting, because it's enough to widen the uncertainty range by adding wrong models to the sample, to make the whole set "better", following your criteria. So we end up with the conclusion that the quality of a set is increased by adding wrong models to it ?

    I think that i applied this criterion of "the set of astrological predictions" , it would perfectly work - because, you know, astrologists are not totally stupid and tend to predict "reasonable numbers" of catastrophes and happy events, so probably the real number will also lie within the set of predictions.

    Something is obviously wrong with this criterion, so I would propose to go back to the nice idea of DM, that I totally share : models are always wrong,but they must be *useful*.

    But what is *useful* ? I would propose a definition : a model is useful if it allows to reduce significantly the uncertainty of predictions, compared with crude methods. That is if it has a significantly measurable better predictive power than crude estimates.

    But what is a "crude estimate" ? well, I would say : just extrapolating the past trend within the standard deviation. That's really easy, and a lot of people are playing with that on the web.

    So I would expect that the models are "useful" if they do a better job than simply extrapolating past variations. Do they ? heeem... I don't see that in the plots ! actually with arctic ice, extrapolating the known curve when models were published would have given better results than the model.

    And saying : "oh, it's worse that what the model predicted, so the situation will be even worse in the future, " is not justified. Because you don't know the real reason why it is worse, and there is nothing saying that this reason will evolve in the same way as the model - actually extrapolating the 2006 to 2007 shift of sea ice melting would have given very bad predictions for 2010. So the only reasonable thing to say is : of my model isn't that good after all, I wouldn't rely on it to make predictions.
    Deleted by Daniel Bailey (Trolling)
2011-04-19 15:51:39Banned
John Cook

john@skepticalscience...
60.231.60.165

I've shut down Gilles' user account

2011-04-19 19:22:09
Dikran Marsupial
Gavin Cawley
gcc@cmp.uea.ac...
88.108.211.99

 

He appears to be back again.

http://www.skepticalscience.com/news.php?n=701#48803

2011-04-19 21:11:42Time to plug that code loophole.
Daniel Bailey
Daniel Bailey
yooper49855@hotmail...
97.83.150.37

He must be leaving his browser open at work; at the time he was banned he was at home eating breakfast.  So now he's at work using his work computer, where he must be leaving his browser open.

2011-04-19 21:30:23
Daniel Bailey
Daniel Bailey
yooper49855@hotmail...
97.83.150.37

His most recent banned comment (by Dikran) comes up with this IP: 152.77.248.103, 152.

Note that this is not a valid IP.  Deleting every character after the "103" results in an IP based in Grenoble (theoretically his home base), France.

2011-04-19 22:49:14
Dikran Marsupial
Gavin Cawley
gcc@cmp.uea.ac...
88.108.211.99

Feel free to delete any of my posts that have responded to Gilles' comments this morning if it will help tidy things up.

2011-04-20 08:45:46
logicman

logicman_alf@yahoo.co...
86.148.172.104

Gilles is probably using a university terminal, in which case he may well be in breach of their terms of use by spamming this site.

Octet series 152.77.x.x  is assigned to dn series x.ujf-grenoble.fr

http://dnstree.com/fr/ujf-grenoble/colorado/

2011-04-20 11:40:54
Daniel Bailey
Daniel Bailey
yooper49855@hotmail...
97.83.150.37

He's probably at home more than a little bit pissed-off right now that his diversion du jour (see? I can spake Francais) isn't even reachable on his home PC...

Edit: Boy, was I ever wrong.

2011-04-20 13:38:19Like a bad rash, he just won't go away...
Daniel Bailey
Daniel Bailey
yooper49855@hotmail...
97.83.150.37

Still at it:

http://www.skepticalscience.com/news.php?n=694#48957

2011-04-20 14:12:07Fixed my code
John Cook

john@skepticalscience...
124.187.72.92

Gilles cannot post under his banned account now - I've tightened up my code so it checks their status on the fly