Controlling Weather and Economies
For me, it is a matter of degree in sensory input, control output and feedback systems.
Weather:
Is predicting the weather impossible? Would it be possible if one had enough sensors? Would it be
possible if one had enough windmills which could be controlled to actually change the weather? How
about if you could explode atomic bombs inside hurricanes? How about if you had huge reflectors on
satellites and could direct sunlight to any spot desired?
Picture the following: You are watching a slug trying to ride a bicycle. It keeps falling over. You are
tempted to say that it is a chaotic system and too unstable to predict or control.
But we know that if you put a five year-old on that bicycle, she would ride it quite nicely.
In a similar manner, my early attempts to control the weather via solar reflectors on satellites would seem
like a failure. Here I am trying to ‘ride’ a hurricane like a slug would try to ‘ride’ a bicycle. Initial
impressions would be that it is a chaotic system and that I am wasting my time.
OTOH, I am not so sure about what is actually chaotic and what simply appears to be chaotic.
Consider the following. I create a robot which is not intelligent, but which simply takes my commands
and memorizes them. I “teach” the robot to ride a bicycle down a hill inside a building. I then show that I
can have the robot repeat the trip down the hill 100 times and it does it perfectly. Now I perform the
experiment and have a butterfly land on the shoulder of the robot while it attempts to ride down the hill
for the 101st time. The robot may well fail to keep the bicycle in balance that last time. This is because
the robot is not using any feedback but only repeating the same pattern over and over. Such systems
are more subject to “butterfly effects” than closed-loop systems.
Economy:
If one had a model of the economy of the world and everyone used credit cards and you could follow
every transaction…AND
You could see when any given commodity became overpriced such as real estate 40% over AND
You could control interest rates for various items such as not allowing interest on mortgages to be tax
deductible when housing became overvalued from the trendline by more than 15% AND
You could also control the discount rate and pull oil out of the strategic reserve when overpriced (as
compared to humans which didn’t) ETC.
I am not so sure that you cannot control the system better than simply blaming the butterflies.
In addition, you have some control over the humans by using advertisements and incentive programs.
Additional stability to the system is granted by encouraging the use of long term futures contracts. One
can buy and sell 2016 oil futures. This means a producer can sell the oil in his well before starting to
drill. As we have discussed, the more oil tankers that exist, the more stable the price of oil becomes
because of traders willing to take advantage of contango to buy spot and sell futures.
In general, stability tends to improve as one increases sensors, effectors and control feedback systems.
don
Computer simulations and modeling
There are countless variables that you would have to keep in sync. Letting one get away could spell
disaster, as prescribed in chaos theory. However, there is an even larger issue. I'll relate it to oil and
gas:
Companies in oil and gas normally talk about volatility, and try to model it using simulations for planning
purposes; since drawing a conclusion using historicals is worthless. The most commonly used model
that I'm aware of is running Monte Carlo simulations. The issue is that human judgement is applied at
the beginning (entering variables) and end of the simulation (analyzing results). The simulation will draw
the same directional pattern no matter how many thousands of times you run it - but the final conclusion
is all dependent upon human judgement. This judgement has proven to be flawed (per your Greenspan
example). As a result, you can forecast, plan, and run as many computer simulations as much as you
want. The only way you'll get close is by sheer luck.
FedEx's most aggressive planning model never showed oil going above $100/barrel. Goldman Sach's
had oil at $200/barrel. CBOT traders had oil prices at Dec 31, 2008 to settle at $150/barrel. All three
started with the same base data, then applied judgement to it.
Markets are self-correcting. We were due for a big plunge, but didn't realize it until it started occurring.
We can reform all we want, but it will happen again. Greed can't be regulated.
Here's a fun game to test out variability of results:
Ask 20 people, "How much is a lot?" Put a percentage on it. You'll be amazed at the results.
Sachin Desai
sachin6@hotmail.com
C: 312-404-6172
________________________________________
From: donbot@comcast.net
Subject: Computer simulations and modeling
Date: Wed, 11 Feb 2009 17:55:50 -0600
Sweden
If you want to model how infectious diseases spread, you need a decent simulator to see how the
various coping strategies pan out. Your simulation needs to take into account the population, its age
and gender distribution, where people live and how far they travel from home to work and which people
share homes.
But making this data realistic would be hard. After all, would anybody willingly agree to have their real
data entered into such a simulation?
Actually yes. Swedes. All nine million of them.
Yep, the personal details of the entire Swedish population have been used to create what must be the
world's largest and most realistic computer simulation of the way infectious diseases spread.
Lisa Brouwers at the Swedish Institute for Infectious Disease Control and buddies have built a simulation
called Microsim in which every member of the Swedish population is represented with details including
their sex, age, family status, school, workplace and their geographic location at these places to within
100 metres.
That makes for potentially fantastic simulations but it also raises extraordinary questions over privacy.
The data is only minimally anonymized: each individual is given a random identifier but otherwise their
personal data is intact.
Given that the team is combining data from three different sources, this doesn't sound like nearly
enough protection.
But Brouwers must know what she's doing. Or at least be praying that the rest of Sweden doesn't find
out what she's done.
- - - - - -
In today’s hearings before congress, one congressman asked each CEO if they could tell how many
more loans they made because they received TARP money. Some of them weren’t in the business of
making retail loans, so those banks don’t count. My point is that the banks that said that they could tell
how many more loans they made because of TARP must have a model or simulation of their bank.
Otherwise they would not be able to access theoretical information like “what if” questions. Normal
accounting will tell you only what happened. Simulation accounting will tell you what happened and what
would have happened if this or that had been different.
My point is that we don’t have the software yet to simulate the whole economy including all of these
banks. The FRB has programs to simulate the economy, but they are not as complete and detailed as
what I am proposing. To create a more satisfactory model would cost billions of dollars. I claim that it
would be worth paying $25 billion (I don’t think it would cost that much) if necessary because then
congress and the FED could say, “what if we have these items in the stimulus package”, followed by
changes until it was optimized according to the model being simulated. Obviously the model will not be a
perfect representation of reality. Nevertheless, it would be so much better than operating without such a
model that we could easily save hundreds of billions of dollars by using the model.
Consider the following: If you read Greenspan book, you will find that he refers to using an economic
model. The problem is that the model was mostly in his head. He now admits that an error in his model
was that his model assumed that housing prices would never go down nationwide year over year.
(Oops) If the model had been in a computer rather than in his head, two advantages would be clear.
One is that all of the members of the Federal Reserve Board could debate whether the parameters were
correct or not. The other advantage would be that even after Greenspan retired, we would still have the
model. Consider taking that most excellent Greenspan model and correcting it by allowing that housing
prices CAN go down year over year. If we now had such a model in a computer at the FRB, we would be
in better shape to know what to do next.
There are things that humans are better at and there are things that computers are better at. People
making decisions on economic stimulus packages by testing ideas against a computer model would be
far superior to congress debating the subject and coming to a political compromise between the various
models in their respective minds.
I would like to see a supercomputer which is tied into the bank accounts of every corporation and
individual in the US much like the Swedes are tying into medical information in the above example.
In an interview with Peter Jaworski (The Journal, Queen's University, March 15, 2002—Issue 37, Volume
129) Milton Friedman said that ideally he would "prefer to abolish the federal reserve system altogether"
rather than try to reform it, because it was a flawed system in the first place. He later said he would like
to "abolish the Federal Reserve and replace it with a computer."
That, in my opinion, is where we should be headed.
don
Chemtrails (See C2C 12-07-2011)
Geo-engineering programs use chemicals disbursed from airplanes (Chemtrails) and other means to
allow corporations to manipulate natural systems including weather and soil, which in turn will lead to
marketing opportunities for goods and services tailored to the new environmental conditions being
created.
Next
For me, it is a matter of degree in sensory input, control output and feedback systems.
Weather:
Is predicting the weather impossible? Would it be possible if one had enough sensors? Would it be
possible if one had enough windmills which could be controlled to actually change the weather? How
about if you could explode atomic bombs inside hurricanes? How about if you had huge reflectors on
satellites and could direct sunlight to any spot desired?
Picture the following: You are watching a slug trying to ride a bicycle. It keeps falling over. You are
tempted to say that it is a chaotic system and too unstable to predict or control.
But we know that if you put a five year-old on that bicycle, she would ride it quite nicely.
In a similar manner, my early attempts to control the weather via solar reflectors on satellites would seem
like a failure. Here I am trying to ‘ride’ a hurricane like a slug would try to ‘ride’ a bicycle. Initial
impressions would be that it is a chaotic system and that I am wasting my time.
OTOH, I am not so sure about what is actually chaotic and what simply appears to be chaotic.
Consider the following. I create a robot which is not intelligent, but which simply takes my commands
and memorizes them. I “teach” the robot to ride a bicycle down a hill inside a building. I then show that I
can have the robot repeat the trip down the hill 100 times and it does it perfectly. Now I perform the
experiment and have a butterfly land on the shoulder of the robot while it attempts to ride down the hill
for the 101st time. The robot may well fail to keep the bicycle in balance that last time. This is because
the robot is not using any feedback but only repeating the same pattern over and over. Such systems
are more subject to “butterfly effects” than closed-loop systems.
Economy:
If one had a model of the economy of the world and everyone used credit cards and you could follow
every transaction…AND
You could see when any given commodity became overpriced such as real estate 40% over AND
You could control interest rates for various items such as not allowing interest on mortgages to be tax
deductible when housing became overvalued from the trendline by more than 15% AND
You could also control the discount rate and pull oil out of the strategic reserve when overpriced (as
compared to humans which didn’t) ETC.
I am not so sure that you cannot control the system better than simply blaming the butterflies.
In addition, you have some control over the humans by using advertisements and incentive programs.
Additional stability to the system is granted by encouraging the use of long term futures contracts. One
can buy and sell 2016 oil futures. This means a producer can sell the oil in his well before starting to
drill. As we have discussed, the more oil tankers that exist, the more stable the price of oil becomes
because of traders willing to take advantage of contango to buy spot and sell futures.
In general, stability tends to improve as one increases sensors, effectors and control feedback systems.
don
Computer simulations and modeling
There are countless variables that you would have to keep in sync. Letting one get away could spell
disaster, as prescribed in chaos theory. However, there is an even larger issue. I'll relate it to oil and
gas:
Companies in oil and gas normally talk about volatility, and try to model it using simulations for planning
purposes; since drawing a conclusion using historicals is worthless. The most commonly used model
that I'm aware of is running Monte Carlo simulations. The issue is that human judgement is applied at
the beginning (entering variables) and end of the simulation (analyzing results). The simulation will draw
the same directional pattern no matter how many thousands of times you run it - but the final conclusion
is all dependent upon human judgement. This judgement has proven to be flawed (per your Greenspan
example). As a result, you can forecast, plan, and run as many computer simulations as much as you
want. The only way you'll get close is by sheer luck.
FedEx's most aggressive planning model never showed oil going above $100/barrel. Goldman Sach's
had oil at $200/barrel. CBOT traders had oil prices at Dec 31, 2008 to settle at $150/barrel. All three
started with the same base data, then applied judgement to it.
Markets are self-correcting. We were due for a big plunge, but didn't realize it until it started occurring.
We can reform all we want, but it will happen again. Greed can't be regulated.
Here's a fun game to test out variability of results:
Ask 20 people, "How much is a lot?" Put a percentage on it. You'll be amazed at the results.
Sachin Desai
sachin6@hotmail.com
C: 312-404-6172
________________________________________
From: donbot@comcast.net
Subject: Computer simulations and modeling
Date: Wed, 11 Feb 2009 17:55:50 -0600
Sweden
If you want to model how infectious diseases spread, you need a decent simulator to see how the
various coping strategies pan out. Your simulation needs to take into account the population, its age
and gender distribution, where people live and how far they travel from home to work and which people
share homes.
But making this data realistic would be hard. After all, would anybody willingly agree to have their real
data entered into such a simulation?
Actually yes. Swedes. All nine million of them.
Yep, the personal details of the entire Swedish population have been used to create what must be the
world's largest and most realistic computer simulation of the way infectious diseases spread.
Lisa Brouwers at the Swedish Institute for Infectious Disease Control and buddies have built a simulation
called Microsim in which every member of the Swedish population is represented with details including
their sex, age, family status, school, workplace and their geographic location at these places to within
100 metres.
That makes for potentially fantastic simulations but it also raises extraordinary questions over privacy.
The data is only minimally anonymized: each individual is given a random identifier but otherwise their
personal data is intact.
Given that the team is combining data from three different sources, this doesn't sound like nearly
enough protection.
But Brouwers must know what she's doing. Or at least be praying that the rest of Sweden doesn't find
out what she's done.
- - - - - -
In today’s hearings before congress, one congressman asked each CEO if they could tell how many
more loans they made because they received TARP money. Some of them weren’t in the business of
making retail loans, so those banks don’t count. My point is that the banks that said that they could tell
how many more loans they made because of TARP must have a model or simulation of their bank.
Otherwise they would not be able to access theoretical information like “what if” questions. Normal
accounting will tell you only what happened. Simulation accounting will tell you what happened and what
would have happened if this or that had been different.
My point is that we don’t have the software yet to simulate the whole economy including all of these
banks. The FRB has programs to simulate the economy, but they are not as complete and detailed as
what I am proposing. To create a more satisfactory model would cost billions of dollars. I claim that it
would be worth paying $25 billion (I don’t think it would cost that much) if necessary because then
congress and the FED could say, “what if we have these items in the stimulus package”, followed by
changes until it was optimized according to the model being simulated. Obviously the model will not be a
perfect representation of reality. Nevertheless, it would be so much better than operating without such a
model that we could easily save hundreds of billions of dollars by using the model.
Consider the following: If you read Greenspan book, you will find that he refers to using an economic
model. The problem is that the model was mostly in his head. He now admits that an error in his model
was that his model assumed that housing prices would never go down nationwide year over year.
(Oops) If the model had been in a computer rather than in his head, two advantages would be clear.
One is that all of the members of the Federal Reserve Board could debate whether the parameters were
correct or not. The other advantage would be that even after Greenspan retired, we would still have the
model. Consider taking that most excellent Greenspan model and correcting it by allowing that housing
prices CAN go down year over year. If we now had such a model in a computer at the FRB, we would be
in better shape to know what to do next.
There are things that humans are better at and there are things that computers are better at. People
making decisions on economic stimulus packages by testing ideas against a computer model would be
far superior to congress debating the subject and coming to a political compromise between the various
models in their respective minds.
I would like to see a supercomputer which is tied into the bank accounts of every corporation and
individual in the US much like the Swedes are tying into medical information in the above example.
In an interview with Peter Jaworski (The Journal, Queen's University, March 15, 2002—Issue 37, Volume
129) Milton Friedman said that ideally he would "prefer to abolish the federal reserve system altogether"
rather than try to reform it, because it was a flawed system in the first place. He later said he would like
to "abolish the Federal Reserve and replace it with a computer."
That, in my opinion, is where we should be headed.
don
Chemtrails (See C2C 12-07-2011)
Geo-engineering programs use chemicals disbursed from airplanes (Chemtrails) and other means to
allow corporations to manipulate natural systems including weather and soil, which in turn will lead to
marketing opportunities for goods and services tailored to the new environmental conditions being
created.
Next
Envirobot Loves Nature, oh yes he does
 | |  | |  | |  |