Return to Website

Number Watch Web Forum

This forum is about wrong numbers in science, politics and the media. It respects good science and good English.

Number Watch Web Forum
Start a New Topic 
Author
Comment
Modern computer modelling

In the "The Right Climate Stuff" thread I mentioned that there is a difference between modern computer modelling and the way it was done in the past before desktop computers replaced mainframe computers. Another strange difference is the tendency to use computer models in the present day for decision-making.

An example of this is given in this 2016 article in the Daily Telegraph about a modeller called Nira Chamberlain, apparently listed by the Science Council as one of the UK’s top practising scientists:

Nira Chamberlain


In the article, Chamberlain briefly describes two projects that he is most proud of:

"The first, a Babcock project that involved the creation of a mathematical cost capability trade-off model for the HMS Queen Elizabeth at a time when the £6.2 billion project was still at the computer design stage and the first sheet of steel had yet to be cut.

His model, which convinced the client that this prestigious aircraft carrier should indeed be built, not only earned plaudits for Babcock but saw Dr Chamberlain himself cited in the Encyclopaedia of Mathematics & Society – making him one of only a handful of British mathematicians to receive such an accolade."


Basically he seems to have devised a computer model which 'proves' that building the Queen Elizabeth aircraft carrier is a great idea, and the model has convinced the client to go ahead with the project, the 'client' presumably being the Ministry of Defence.

Some background information on these new aircraft carriers, the largest warships to be built in the UK, is given in this Wikipedia article. The Queen Elizabeth aircraft carrier was actually launched yesterday.

Queen Elizabeth aircraft carrier

The idea of building ships as big as this doesn't look very sensible to me. Decades ago, many countries had aircraft carriers, but the general cost escalation of military equipment over the years has translated into such ships now being regarded, except in the USA, as a luxury item. Another specific complication for aircraft carriers is that ships built since the 1970s are mainly gas turbine powered, which means that there is no immediately available steam supply for the steam catapults that would be needed to launch normal military jets, and the idea of providing an alternative source of steam is unfavoured for some reason. In the UK we switched in the 1970s to the idea of building smaller size aircraft carriers with specialised short take-off and vertical landing (STOVL) planes (Harrier jets). These new Queen Elizabeth class aircraft carriers are still using the short take-off plane technology instead of normal military jets, but the ship is now three times as big as the previous smaller size aircraft carriers. I'm sure the model explains all this away somehow.

Re: Modern computer modelling

Aircraft Carriers are scary things. They show up off your coast and you are likely to be just a little bit worried. The economics of aircraft carriers is therefor murky. If it gets someone to the conference table, it can pay for itself without the benefit showing up in any model.

How do we inspect that difference?

The HPV vaccine costs $750 for a person under 26. see If you are over 26, in the state of Washington, this vaccination can cost more than $1800 (and is not covered by insurance, people under 26 may get most of it covered). How did they get to this price? By the amount saved the healthcare system in the future.

How much should the price of MMR, Polio, Pertussis, Varicella, etc be if they were based on how much money was saved? MMR, Polio, and Pertussis are all first order vaccinations. The consequences of getting those diseases can result in death in the short term. If you get the disease, you are tap dancing with death.

Varicella is really a 2nd order vaccination. In general, people who get chicken pox survive. Adults who get chicken pox have some complications (Shingles), but few die. There is a danger of pocking though.

HPV is, in my estimation, at least a 3rd order disease. It turns out that just about everyone has had it. Almost no one knew that they had it. The disease went away without them ever knowing they had it. Some people who get it may get cancer 10 to 30 to 50 years from now. (I am highly certain that HPV can be passed through non sexual means).

I am not opposed to the vaccine. Since it doesn't hit my bottom line directly (a 2nd order feedback response problem), my eldest has been given the HPV vaccine. My youngest will get it also. A sexually transmitted disease discussion still has to be had because it doesn't defend against most of them.

Somehow a model was used to make defending against a 3rd order disease more rational than defending against 1st order diseases. Once again, people with science degrees are fixated on the 5th decimal place and sort of losing sight of the 2nd.


Squinting at the walgreens chart, I get about $2000 for a full series of vaccinations for kids. Of this $750 is HPV.

I hope someone here can laugh at my inability to get mad at anyone who doesn't want to pay to defend against 6 of the 30+ strains of HPV..

HPV vaccine

In reply to Brad, it looks like the high cost of the HPV vaccine is, as you say, based on avoiding future healthcare costs associated with cervical cancer treatment according to this Discover magazine article from 2007:

vaccine cost

The high price, $360 in 2007, was approved by CDC (Centers for Disease Control and Prevention). The age limit of 26 corresponds to the age range for taking the vaccine recommended by CDC. The even higher price for the vaccine charged to people over 26 might be associated with increased legal liability by the vaccine manufacturer as they are outside CDC's recommended age range.

I suppose the calculation of the price which considers future healthcare costs might come from a computer model as you are suggesting.

One correction I should make to my earlier post is that I said that the new Queen Elizabeth aircraft carrier was launched on 26th June, whereas it would be more accurate to say that it set off on its maiden voyage (for sea trials) on that date. The launch ceremony for the ship apparently took place three years ago in July 2014.

Re: HPV vaccine

The vaccine only affects 4 of the strains.

This article looks at the efficacy of the vaccine. My conclusion from reading it is ..

If your child is less than 16, it appears to be useful. > 16 the tap dance is already underway. Below 16, the RR is 0.25. As the age increases, the effectiveness decreases. This may be because the older people were already infected and the vaccine seems to have no effect on infections already in place.

I talked to a Pathologist about CIN, CIN2, CIN3. These are rankings of the abnormal cell in the cervix biopsy. Apparently 90% of CIN disappear on their own. CIN2 and CIN3 are more likely to develop into full cancer.

In discussions elsewhere on the subject, I have difficulty trying to explain my discontent with the HPV Vaccine.

My son is about to learn to drive. Step 1 is driving school. The school is outside of the main eduction system. It costs $500. He will go to this school. I will pay for it.

My nervous tick, I attempt to express is as this...

"Given $500, what is more likely to save my son's life***.... the HPV Vaccine or Driving School"

How many $500 questions like this do we have to answer before, HPV Vaccine wins?

As an individual, I have to ask that question. It is just as valid a question for any level of management. The only way to make the HPV Vaccine win is to make the buckets used in the analysis change sizes.

One way to is say "Vaccinated people are 4 times less likely to have CIN2 and CIN3". It looks like people who get the vaccine earlier are less likely to have issues with CIN2 and 3 early on. Presented this way, it looks like a no brainer to do the vaccination. If I exclude all other avenues of life saving and cost saving and fixate solely on the expense of one thing, HPV Vaccine it is.

This is how engineers solve problems. Isolate the problem into manageable chunks and evaluate.

There is a cost $X where it becomes not worth worrying about expending the resource.


*** There is a fallacy here. There is some chance it may save expense for women. For boys, the most likely benefit is from not passing it on to women. But now we are in a convoluted discussion of sexual politics.

Re: HPV vaccine

One of the companies I worked for tried to use the HPV Vaccine pricing model. We analyzed the efficiency of power plants and pointed to the inefficiencies that could be reduced. We tried to figure out how to sell the computer model based on 'savings'. We could save the plant $300k per year. You just pay us 10% of that.

A Nuclear Risk Analyst pointed out the conundrum. A typical plant we worked with was between 350MW and 750MW. If a 350 MW facility went down unexpectedly for 1 day. 350MW * $30/MWh * 24hr = $250,000.

1 day was worth the value of our yearly savings.

Where do you focus your efforts? Planned Maintenance? or Optimized Efficiency?

Why not both? Both is fine. Using the example, my linear brain (I will do everything I can to keep a business out of debt by giving you 90% of what you want for 5% of the cost) says 1 Maintenance Engineer gets to devote 10% of his time to optimization.

Any individual in a healthcare system is not likely to be motivated to take the HPV Vaccine. The system itself will quite likely never see the savings. There will always be money in the healthcare system at the beginning of the year. There will always be almost no money at the end. To find the savings will require advanced and very sophisticated analysis, because at no point will a "surplus" magically appear because of this vaccine. In the case of Polio, sanitariums magically disappeared. The money that went to them magically got sucked into a different organization.

I admire the salespeople that managed to get the elevated price. Those salespeople are worth their weight in bitcoin...

Re: Modern computer modelling

Dave,

I too have my doubts about the modelling here. The only comment I'd make though is that carriers are enjoying a moment.

Chinese send their carrier on manoeuvres

Indians building their own aircraft carrier

So clearly there are people out there who see plenty of value in a carrier, and they aren't going anywhere soon.

Re: Modern computer modelling

Another strange feature of modern computer modelling is its very lax attitude towards modelling work actually being valid. Here's an example I noticed from a couple of months ago where modelling of buildings for energy efficiency is reported as not necessarily bearing much relation to reality:

energy efficiency modelling

Extract from the article:

"The difference between how much energy a building is predicted to use and how much it uses in reality has been known in the industry for decades, and is dubbed the ‘performance gap.’ But architects and engineers have traditionally blamed the problem on faulty construction, or unexpected use after completion - such as owners leaving too many lights on.

However David Coley, Professor of Low Carbon Design at the University of Bath, said the real problem stemmed from the practice of building modelling, which is not ‘fit for purpose.’

“It’s a serious scandal,” he said. “It affects all new buildings as well as the refurbishment of older ones.

“When one school in Plymouth was rebuilt, the energy bills for a month ended up costing the same as for an entire year in the old 1950s building.

“The problem is nobody checks that the building is performing as promised. There is very little regulation. They can't be sued. It’s like a surgeon not being bothered about whether their patient lived or died.

“The impact of the inaccuracies of building modelling professionals has severe financial and environmental implications for both the government’s global warming targets as well as building owners who are purchasing homes and other buildings that are sold to be energy efficient but in reality are not.”"


I think the origin of this requirement to model buildings for evaluation of energy efficiency came from the "Energy Performance Certificate" idea that was introduced in some UK legislation about ten years ago, and going further back than that, it comes from some EU directive (and further back than that it will have been the result of lobbying by "Big Green"). This Wikipedia article describes the Energy Performance Certificate idea:

Energy Performance Certificate

I would argue that there is not much point in trying to model the buildings if this work cannot really be done properly. The legislation should not have been introduced until there was sufficient confidence that energy efficiency of buildings could be modelled satisfactorily.

This Energy Performance Certificate idea may have contributed to the recent Grenfell tower block fire event in London. The new cladding (or "siding" as I believe it is called in other countries) for the tower block was probably introduced mainly for cosmetic purposes to help maintain high local property prices. But the Energy Performance Certificate legislation is likely to encourage somebody to think it might be a very good idea to incorporate insulating material into the cladding construction, which unfortunately increased the fire risk for the building. There is a possibility that the Grenfell tower block was in reality not that much more energy efficient than it was before the cladding was installed, even if it got a much improved rating on its Energy Performance Certificate, as the modelling activity is a bit suspect.

Re: Modern computer modelling

I think modeling has a place. It can help me recognize where efficiencies may be gained by pointing out things that might be problems. Can and may aren't will.

I suspect that some folks think that the mays and mights are as good as "will".

I am forced to camp from time to time. I have a sleeping bag rated to 20F. It is rare that I zip it up. Normally it just lays on top of me.

A tent with few vents will keep you ever so much warmer than a building with open walls.

A building with open walls will keep you ever so much cooler than a tent with no vents.

Those are the same statements. In summer, the bottom is better. In winter the top is better. This may sound so obvious as to be foolish to even talk about. I suspect that many of these folks have never integrated this information though.

A piece of 1 mil plastic will turn that frigid Adirondack into a toasty oasis (toasty is relative), but it is the difference between getting completely mummified in my sleeping bag or just lying under it.

Windows are tradeoffs. Security vs light. Vs airflow. Vs. egress. Vs. Comfort. Eventually you get to very marginal returns. If your energy bill is $800/month, there is lots of room to make improvements. If the energy bill is $200, things get a little more uncertain. $100? We now might want to look at things that will use more energy to make our lives a little easier.

Re: Modern computer modelling

In my previous post about modelling buildings for energy efficiency, I effectively made the claim that back in the old days of mainframe computers, people were much more concerned about whether modelling work was valid. I can give some evidence for that claim from my own personal experience.

In the first two years of my working life, at the end of the 1970s, I worked on noise and vibration reduction for nuclear submarines using various computer models. The models we used were taken from methods given in mechanical engineering textbooks, and we used some finite element analysis programs, STRUDL and later on PAFEC, and also a program devised by Imperial College called COUPLE which combined analysis with test data. Whenever we compared our computer models with test data, the comparison was not very impressive. Basically we plodded on doing this work, which was funded by various MoD research establishments to get some sort of onsite analysis capability into a shipyard, and it had no influence on the design of any submarine. The computer models were not regarded as being "ready for primetime". About ten years later I happened to meet somebody who was currently working in the section that did this noise and vibration reduction analysis work, and I discovered that all our computer models from the end of the 1970s had just been thrown away.

But that strikes me as the correct attitude to computer modelling work. If it isn't valid, you can still potentially fund it to some extent, but you shouldn't really make any use of it. There seemed to be a more sceptical attitude towards computer modelling work in the days of mainframe computers. When desktop computers took over, they were portrayed as representing the future of office work, and the attitude to things like computer modelling became less sceptical.