Play Bunny Play
From Frank and Theresa Caplan, The Power of Play, about what makes for success in creative professions
To create one must have a sense of adventure and playfulness. One needs toughness to experiment and hazard the risk of failure. One has to be strong enough to start all over again if need be and alert enough to learn from whatever happens. One needs a strong ego to be propelled forward in one's drive toward an untried goal. Above all, one has to possess the ability to play!Ms. Rabett calls Eli's lab his toy chest. He mutters about her stash taking over the house. Live safely, die bored.
Posts
21 comments:
This is a kind of love the people attacking scientists really don't understand. In classic conspiracist fashion, every investigation, every prediction, every discovery is seen either as the advancement of the evil government conspiracy or the triumph of the rebel truth-sayers.
I'm pretty political, but hopefully not that political. Sometimes I want to say "Hey, man, we can go back to snarling at each other in just a minute. There's some cool science happening; let's enjoy it before we try and determine what it means for whom."
Speaking of taking over the house, I would direct the Bunny's eyes toward this. Look carefully at the listed personnel.
Not nearly as amusing as the papers. You gonna tell Tamino or should Eli.
Hmm, hadn't even looked at those. Looks like a couple of obscure geology topics and... inevitably I suppose... our old friend Nic. You go ahead and pass along the Tamino-bait, but I'm more curious as to how the hell Morner of all people was allowed into such a role in an EGU pub.
The belief that risk will save us from boredom sounds to me like a symptom of Despair.
Another famous place is the Tuvalu Islands, which are supposed to soon disappear because they’ve put out too much carbon dioxide. There we have a tide gauge record, a variograph record, from 1978, so it’s 30 years. And again, if you look there, absolutely no trend, no rise.
- Dr. Nils-Axel Mörner
url="http://www.psmsl.org/data/obtaining/rlr.monthly.data/1452.rlrdata"
rlr=read.table(url,sep=";",na.string="-99999")
lm(rlr[,2]~rlr[,1])$coeff[2]
#rlr[, 1]
#2.286868
# "Next!"
Can someone please tell me why this 'cheat sheet' got me to within 0.002c from a month out on the UAH Global figure published recently?
I genuinely do not know.
But to be THAT lucky, first shot?
http://i1291.photobucket.com/albums/b550/RichardLH/uahtrendsinflectionfuture_zps7451ccf9.png
Fig 1.
http://i1291.photobucket.com/albums/b550/RichardLH/UAH-Comparisonofcascadedlowpassfilteroutputmonthsrunningaverage_zpsf1883e44.png
Fig 2.
Richard Linsley Hood (and yes, that name is real).
Sorry: That should have been 'this simple mathematical cheat sheet'
RLH
"I'm more curious as to how the hell Morner of all people was allowed into such a role in an EGU pub."
Could be related to the editor-in-chief being from the algerian petroleum institute ...
@RLH: normal distribution of the residuals
RLH has already tried his mathurbation at Tamino's and is banned there, I guess he's trying virgin territory over here ...
Bunnies may be a lot of things, but virgin ain't usually one of them.
Ron Broberg said...
@RLH:
"normal distribution of the residuals"
12/7/13 9:26 AM
That will be interesting then as I was unaware of any residuals when using simple averages. Like you would do to the whole thing in you just averaged all the figures, but this 'circuit' splits it into timebands.
Are you aware of bandpass splitter circuits in analogue. Used all the time. This is the direct digital equivalent.
I would love to know the mathematical or logical reason why it not correct. A genuine response about the maths.
I am just trying to get an explanation if that is not too much.
Anonymous dhogaza said...
"RLH has already tried his mathurbation at Tamino's and is banned there, I guess he's trying virgin territory over here ...
12/7/13 9:56 AM"
Well actually he is still trying to get some answers.
I DO realise that it could all have been luck, bur are you really, REALLY, sure?
I mean, it is only averages, that is all.
To be crystal clear about the method I am using and so that you may try it for yourselves the procedure is simple.
Take a column of energy related figures,
Create a cascaded set of running averages (output of sets the input range of the next) in further cols alongside.
Step up by using: next = round(previous *1.3371).
Start from 2 if you like but un-zip it to get the maximum span window that the data can provide.
So you now have smoother and smoother data down over time as the input has dissipated into the system.
Just like any other power circuit. After all, that is what we are dealing with.
So then you can either plot the first graph or, as is normal, you subtract between pairs of columns to get a 'zero' referenced signal in the bassband/timespan tio get the second graph.
Almost all engineers who deal with power would know and recognise that circuit and its outputs.
What scientific reason is there for it not being true for climate?
And if you really believe that the span numbers that come from that series co-incidentally includes,
12,28,37, and 49
which are year, month, lunar orbit 'lock' and 4 years (solar lock) and thus has $ direct orbit figure in it then you do start to wonder. I mean, how many points do you have to get that are potentially orbital related for you NOT to be as least slightly curious.
@RLH: was unaware of any residuals when using simple averages
Your residuals
"Ron Broberg said...
@RLH: was unaware of any residuals when using simple averages
Your residuals
12/7/13 8:07 PM "
You know, I thought that plotting the 'noise' and demonstrating that it has a normal distribution around the center line was good practice. Must have forgotten something, somewhere.
Oh, I know, because I foolishly used a start point of 12 (it was just a trial at first) for the series. It wasn't until quite recently that I realised that the series actually starts from 1, but an average of 1 sample is....
And so the first stage on the current display has some 'noise' that is digital.
If I had started from 1 then that distribution would be even better.
If you wish to prove me wrong, my working are on the net both as graphs and description.
RLH
You know, I thought that plotting the 'noise' and demonstrating that it has a normal distribution around the center line was good practice.
Indeed, it is.
Must have forgotten something, somewhere
Indeed, you did.
I was unaware of any residuals when using simple averages.
And now I have linked those two concepts for you: residuals ~ noise
If you wish to prove me wrong, my working are on the net both as graphs and description.
Ah. No.
I have no wish to prove you wrong.
I thought I was helping you to understand the answer to the following question ...
Can someone please tell me why this 'cheat sheet' got me to within 0.002c from a month out on the UAH Global figure published recently? I genuinely do not know. But to be THAT lucky, first shot?
Hopefully, you now see that your luck, while random, was guided. Guided by the shape of "noise", "noise" you have already calculated.
So let me get this right, the 'noise' which is weather, by normal distribution, and predicting, not this month, but 18 months. out.
OK. SO the bet is on.
I predict that the UAH figures will be higher until Dec2013 and then they will stay lower for a period of about 18 months.
Now if I get THT even close to right....
As validity for my engineering credentials, perhaps you would look though these drawings and comment.
http://s1291.photobucket.com/user/RichardLH/story/77871
Post a Comment