« Fighting the pro-capitalist state | Main | Interest rates & the 1% »

September 04, 2014

Comments

Stevenclarkesblog.wordpress.com

A very good point but it raises lots of interesting questions.

In what situations/contexts is simplicity useful and where is complexity better?

How can we distinguish between them in advance?

What is it in the nature of each context that makes it more suited to simplicity or complexity? Is it because some relationships are simple and depend on a few variables - considering more adds no useful information but can lead you astray? Others are complex and simplicity omits important variables.

Jean-Marc Liotier

As Albert Einstein has been paraphrased to say: “everything should be made as simple as possible, but no simpler” !

In general, there may be decreasing returns in adding variables, but is that also the case in chaotic systems ? It seems to me that complexity is a part of making the model robust to chaotic reality.

Deviation From The Mean

"that simple models often out-perform expert judgment"

The problem with this is that simple models rely on a body of knowledge. So what appears to be a simple model is in fact something created on the back of past brain power. Even to be conceived the simple model relies on a past body of accumulated knowledge. Simple models do not spring into existence from nowhere.

If a simple model didn't account for this external knowledge it would stop working after a short period of time as history would have made it moribund.

Good management practice often doesn't do away with judgement but simply devolves it to people who actually know stuff (and then to the manager this looks like control and feedback rules because they are not doing anything!). But I do agree with your attacks on managerialism.

Yes

I don't agree that Yes voters see independence as the simple answer to Scotland's tricky problems. People are being asked to vote in a binary contest. They are being asked to consider that voting No will free them from having to deal with the tricky problems that would be posed by independence.

rogerh

Rule of thumb works well most of the time. Practitioners who intelligently tinker with systems often get good results ahead of the theorists. Those who tidy up systems and remove waste get results for a while at least. History is full of charlatans, mountebanks and frauds for whom simplicity is anathema - so last year - but who just sometimes hit on a good idea. Progress seems to emerge from this primordial soup of messy half-right ideas.

Which brings us to the political-media class, always on the lookout for stealing a march, gaining an advantage, getting an article published, being invited to a conference. Success in their world has nothing to do with success in the real world, so pity the poor practitioner being pestered with yet another nostrum and thinking 'Oh no, not again' but with the bonus of a slight insecurity - 'but what if it works'.

Matt Moore

Best post of yours that I've seen so far.

Have you read "Simple rules for a complex world"?

Do you have an earlier post that explains your version of Marxism? I'm finding it hard to reconcile your blog content to the worldview of other Marxists I have know.

astaines

Nicely expressed. I do a lot of modelling for rather different purposes, and I'm always tugged, as is my discipline generally, between parsimony and what we call 'realistically complex models'.

As a life long fox I prefer robust big models, to fragile small ones.

ajay

The entire school experience consists of learning simple explanations, and then learning that those were wrong and that the real explanation is more complex. (No, atoms aren't really little solar systems with electrons whizzing round protons...)

No wonder we're predisposed to thinking the correct explanation must be the complicated one.

The comments to this entry are closed.

blogs I like

Why S&M?

Blog powered by Typepad