Posted on 05/20/2020 11:08:22 AM PDT by ImJustAnotherOkie
I have been asked by a source in Britain to review the Ferguson model code for my opinion. Just so everyone has some idea, the original program used by Ferguson was a single 15,000 line file that had been worked on for a decade and by no means is this remotely sophisticated. I seriously doubt that Imperial College will want to go public with the code because it is that bad. To put this in some perspective, just the core to conduct basic analysis in Socrates is about 150,000 lines of code. It is so complicated, it takes a tremendous amount of concentration to try to see the paths it has available to it for basic analysis.
To try to keep this in traders terms, reviewing the code reveals this is just a stochastic which is INCAPABLE of forecasting high, low, or projected price target expected to be achieved. Any trader knows that a stochastic is a trend following measure not a forecaster of the trend nor a projection tool to say when a market is overbought or oversold. This clearly shows the vast chasm between trading models and academic models where the money is never on the line. The documentation even states:
The model is stochastic. Multiple runs with different seeds should be undertaken to see average behaviour.
Stochastic is simply defined as randomly determined; having a random probability distribution or pattern that may be analyzed statistically but may not be predicted precisely. In other words, they begin with a presumption, and therein lies the FIRST error. Fergusons assumption was wrong, to begin with. Then this mode is so old, they recommend that it be run only on a single CORE processor as if we were dealing with an old IBM XT.
Effectively, you start the program with what is called a seed number which is then used to produce a random number. Most childrens games begin this way. In fact, this is a version of what you would be similar to the game SimCity where you create a city starting from scratch and it simulates what might happen based upon the beginning presumption. There are numerous bugs in the code and the documentation suggests to run it several times and take the average. This is just unthinkable! A program should produce the same result with the same data from which it begins. Therefore, there is no possible way this model would ever produce the same results. In reality, this model produces completely different results even when beginning with the very same starting seeds and parameters because of the attempt to also make the seed random. This is not even as sophisticated as SimCity, which is really questionable. This is where the Imperial College claims that the errors will vanish if you run it on an old system in the single-threaded mode as if you were using a 1980s XT.
In programming, you run what is known as a regression-test, which is re-running a functional and non-functional test to ensure that previously developed and tested software still performs after a change. In market terminology, its called back-testing. In the most unprofessional manner imaginable, the Imperial College code does not even have a regression-test structure. They apparently attempted to but the extent of the random behavior caused by bugs in the code to prevent that check? On April 4th, 2020, Imperial College noted:
However, we havent had the time to work out a scalable and maintainable way of running the regression test in a way that allows a small amount of variation, but doesnt let the figures drift over time.
This Ferguson Model is such a joke it is either an outright fraud, or it is the most inept piece of programming I may have ever seen in my life. There is no valid test to warrant any funding of Imperial College for providing ANY forecast based upon this model. This is the most UNPROFESSIONAL operation perhaps in computer science. The entire team should be disbanded and an independent team put in place to review the world of Neil Ferguson and he should NOT be allowed to oversee any review of this model.
The only REASONABLE conclusion I can reach is that this has been deliberately used to justify bogus forecasts intent for political activism, or I must accept that these academics are totally incapable of even creating a theoretical model no less coding it as a programmer. There seems to have been no independent review of Fergusons work which is unimaginable!
A 15,000 line program is nothing. I will be glad to write a model like this in two weeks and will only charge $1 million instead of $79 million. If you really want one to work globally, no problem. It will take a bit more time and the price will be at a discount only $50 million on sale refunds not accepted as is the deal with Imperial College.
as we from the 70’s used to joke, “this sounds like a big Kludge.”
Gates indeed. some friends of mine used to joke back in the day that they needed to keep “Bill” away from the code.
however, no joking here, unlike Gates and DOS/Windows, this software is killing people.
Seriously. The way I read this, the code is designed to produce different results each time it is run.
Well, okay, that might make sense. The inputs have uncertainty, so the result will necessarily have uncertainty. You'd think the code would then loop and "do the uncertainty math" on the conclusion. THAT result should be repeatable withing a defined uncertainty / Z-value, whatever you choose for your distribution.
At any rate, instead of GIGO, we have garbage in, garbage algorithm, garbage out ... GIGAGO
This software WRECKED a 20 trillion dollar economy just in this country, not accounting for the rest of the world. No one should be defending it on any level.
Follow the Science, they said
Trump should let the science experts manage the national response, they said
Fauci and Birx are the smartest scientists in the US, they said
Funny, they are still saying this
You are talking about the Fearpers, right?
Nope
Ridiculous. Applying past experience with other diseases to new ones is the essence of epidemiology.
To think we can't learn anything from earlier virus-spread diseases is silly.
You say modeling based on our past experience is useless "NOTHING to support prediction".
Our leaders need to make decisions. How do you propose they inform themselves?
You can only program in ANY on a computer with an ANY key. Everyone knows this. /s
To properly regression test, there should be a way to circumvent the randomness so as to get a repeatable answer. Otherwise, how do you distinguish between a breaking change and an expected random result?
Again, the hockey stick from global warming was the worst example of this at the time. Now I would have to say that this model and its dire predictions have been worse to the world than the hockey stick ever was.
That’s kind of QWERKEY.
Wickedly fast and accurate. It still runs on your high end graphic cards. Not for business data processing. Algorithms have been tuned over half a century like modern jet engines
Mostly missing miles of unwicked object oriented abstraction layers.
Also no real recursion in the good parts. Work done in registers not virtual ram. This confuses java thinkers.
Yep. In several places, I gather. I suspect randomness is introduced in more than one place/variable in the chain.
The epidemiologists use the very limited data they have about the new disease, make some assumptions based on the likelihood it will perform somewhat like similar pathogens, and caveat the hell out of their predictions.
So again, at the beginning of a pandemic with a novel virus how should people make decisions?
I would say anyone developing in FORTRAN in these days is a retarded.
I always like the phrase some people get 5 - 10 years experience, but some people get 1 year of experience 5 or 10 times.
What similar pathogens did they base all their other assumptions on?
Feel free to continue your argument with someone else. I have little respect for what you refer to as epidemiologist, or any medical or scientific specialty, who would use 'limited data' to wreck entire societies for their own vanity.
It’s very clear we don’t have to understand or even look at the details of the code in order to know the whole thing was a sham.
The answer to your question is FIRST AND FOREMOST they should separate the natural occurrences from the unnatural.
A virus created in a lab is unnatural.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.