Re: OT: interesting global warming quote found elsewhwere

Martin Brown wrote:
James Arthur wrote:
On Jul 26, 5:31 pm, bill.slo...@xxxxxxxx wrote:
On Jul 26, 1:06 pm, James Arthur wrote:

bill.slo...@xxxxxxxx wrote:
James Arthur wrote:
[Climate modeling] is like the stock market and the many, many "models" developed
to profit from it. They all go bust.
"Past performance is not an indication of future results."
--Wall Street disclaimer
And this is for a closed, mildly chaotic system with a
well-established long-term upward trend, with a limited
set of variables, all available with great precision,
in real-time. The entire system is "instrumented,"
documented, and logged on-line.

Sure, but any predictive model of the stock market feeds information
to the participants in the market, who proceed to try and exploit,
thus destroying the model's predictive capacity.

No, I'm speaking of any number of super-secret attempts by
Wall Streeters to hire PhD statisticians, super computers,
and try to beat the street with models.
They absolutely do not share info with the market.
PBS (Public Broadcast System, public, non-profit TV in the US) has
documented at least one such attempt.

As soon as they buy shares, they share information with the market.

That notion doesn't follow. To the extent it's true, it would
increase profits.

Not necessarily - primarily it tends to increase volatility.

Buying/selling stocks shares what information? That the
stock is worth buying/selling.

Buying drives prices up. Since you already bought (based
on your successful model), a rising price benefits you.

Vice versa on the sell side.

So, to the extent you've shared information, it benefits
you, contradicting Bill's explanation.

But I was illustrating the perils of longer-term models of
chaotic systems; maybe he means short-term.

There are other interactions and complications. That was
my point: the stock market, far simpler than climate,
defies prediction except over short spans. Not centuries.
Not even six months.

Even this simpler system has subtleties the modelers can't
capture. And the global climate modeling problem is worse.

Wasn't a rash of hurricanes predicted the year after
Katrina? And the next year? Wrongly? And weren't
they wrong, even updating mid-season? They were.
Why? Because their models /don't/ predict.


Here's another analog: we know about plate tectonics,
measure strain and movement, know the history and
long-term trends, yet cannot and do not try to predict
earthquakes.

/That's/ science. Seismologists have models, but
do not presume to predict what they cannot predict.
Despite huge pressure to do so. Kudos to them.

Cheers,
James Arthur
.

Relevant Pages

  • Re: OT: interesting global warming quote found elsewhwere
    ... They absolutely do not share info with the market. ... And the recent bounce in oil prices owes much of the overshoot to speculators driving the price ever higher and product scarcer. ... The researchers have to be extremely careful about what they disclose because the mere mention of increased risk of an earthquake in obvious locations can affect property values and in a hyper litigious society like the US it gets hairy. ... The Japanese are still trying pretty hard at earthquake prediction. ...
    (sci.electronics.design)
  • Re: Bailout alternatives
    ... Presumably if all these finance institutions bankruptcy was ... allowed (which is what keeps normal market ... ... Selling money for profit? ... Investment banking will henceforth need to be conducted as non-profit, ...
    (sci.econ)
  • Re: OT: interesting global warming quote found elsewhwere
    ... Wall Streeters to hire PhD statisticians, super computers, ... They absolutely do not share info with the market. ... defies prediction except over short spans. ... Actually the stock market overall is more predictable over very long investment timescales otherwise there would be no point in putting money into it. ...
    (sci.electronics.design)
  • Re: Mac Mini Clones Redux...
    ... But trading a higher percentage of profit on a per unit basis for a larger percentage of the market could net them a profit greater then what they have now - by an order of magnitude no less. ... Why should Apple care about marketshare, as installed platforms, which is pretty much all the Windroids seem capable of spouting. ... Ruin Apple and make little or no dent in the 'marketshare'. ...
    (comp.sys.mac.advocacy)
  • Smart Money
    ... could have put his money in a hedge fund and waited. ... Carl Wolfenden, the acting exchange manager of Intrade, explains the logic ... The market pricing is ... But the political prediction markets didnâ??t begin with high-rolling ...
    (rec.audio.opinion)
Quantcast