Beware of the Other Santa Clause

Reality doesn't often dance to the Extrapolation Tune

Note: this post has been updated on Jan 4th 2023 to better define the clause

You might be thinking what does this post have to do with satellite conjunctions?

Well, the point as you’ll see is that we always have to be careful when seeing relationships that may not be there in the first place. That may be an artefact of the argument framework we have used rather than something repeatible and reliable.

We have to be careful not to extrapolate beyond the bounds of our assumptions. Or if we do we must be very careful that the context of this extrapolation doesn’t lead to all sorts of nonsense.

And damage in the real world.

The Other Santa Clause is a way of talking about this and it’s an example I came up with that I’ve used a few times with non-scientific people. They actually get the meaning behind it too.

Tim Allen sets the scene

The set-up of the original movie (The Santa Clause) is actually a great set-up for the argument I’m making in this article.

The movie, made in 1994 is all about how a man ends up seemingly killing Santa Clause on Christmas Eve and replaces him to become the new Santa Clause. Because Scott (Tim Allen) put on the suit to finish off Christmas deliveries he is now bound by the “Santa Clause” to be Santa. The other subplot deals with how Scott doesn’t believe in Santa Clause anymore but wants his kid to believe in it. His ex-wife and husband want the kid to ‘grow up”.

The subplot is also relevant in a way as we’ll talk about later.

The Other Santa Clause is set up like this:

A major city decides early one Spring that all residents in that city must replace their roof tiles with special non-slip tiles that cost at least the equivalent of $1000 per square foot in whatever currency is relevant to that city/country.

All these renovations must be done before Christmas Eve.

Renovations are being done to prevent a Health & Safety incident and the associated liability if Santa Claus falls off a roof during his duties.

Now the last bit usually pulls a laugh. How stupid does that sound?

Okay, context set up, here’s the Other Santa Clause. :

You can see the link with the original Santa Claus: he has contingency when he falls off a roof. In the Other Santa Clause we add prevention as risk mitigation.

What this clause does is set the frame for further action predicated off of assumptions about Santa Claus. But it leaves open the scope of those measures. However the act of accepting the clause means you also accept the premise.

So is taking action reasonable?

This is where the meat and potatoes of the argument is. And it’s usually where I have to discuss what the Scientific Method actually allows.

It’s a wide spectrum. Much wider than those who quote Richard Feynman:

“First you guess. Don't laugh, this is the most important step. Then you compute the consequences. Compare the consequences to experience. If it disagrees with experience, the guess is wrong. In that simple statement is the key to science. It doesn't matter how beautiful your guess is or how smart you are or what your name is. If it disagrees with experience, it's wrong. That's all there is to it.”

What he leaves out in this particular quote, is that a hypothesis can exist in the body of scientific literature without ever being tested, or tested only decades later. And that this hypothesis can be quoted and referenced and create a whole field of inquiry.

All of this new field hangs on the original assumptions in that original hypothesis.

Where this falls apart is when you have to measure things pretty quickly, such as dictated by your hypothesis. Then Feynman’s quote comes through.

So you might be reading this as a scientist yourself and think no that’s wrong, science is always about testing reality to hypothesis.

But I know as someone who has worked in different physics subjects and as an engineer that you also know that it’s not quite as black and white as that. There are lots of times when general assumptions are made that when you analyse the application you realise that you’ve just invoked the Other Santa Clause.

Let me explain some more:

All Things Being Equal

The most common assumption is All Things Being Equal (ceteris paribus). This runs through many scientific works and even is used in lots of validation and verification activities in engineering.

It is often used when we assume that the situation we are investigating isn’t particulary different than other situations in which said phenomena occurs. That in most if not all cases there aren’t other external factors that play a significant part in what we are asserting over the range of tests or scenarios we present.

So all things being equal the output of process will not change within certain limits during an experiment such that it will then affect the other thing we are trying to measure.

So if you were measuring the electrical output from a plasma and that plasma has been running in a steady-state (according to some definition) and you then vary the inputs slowly, all things being equal there should be the following variation if A, B or C etc is happening.

All things being equal, if a person eats a certain amount of food per day proportioned in certain macro quantities, and performs a certain type and amount of exercise, they are likely to be this certain composition after some months.

This example is a bit more wooly but hopefully you can see that the application of All Things Being Equal may allow for other assumptions to slip in. Assumptions that may then create artefacts in your measurements/observations.

Here’s one such All Things Being Equal clause that might affect the Overhead app output:

Errors / Uncertainties are random ( follow a Gaussian distribution )

A lot of times statisticians will define uncertainty within a certain model and call “errors” what experimenters or anyone working in real measurements calls “offsets”. Most if not all those who deal in the real world blanket the whole error/uncertainty field as just “errors”. That’s what I will do here as well.

In many experimental set-ups, errors are a dark art and cover concepts such as systematic uncertainty and irreducible uncertainty as well as how you minimise them. And many times that involves flicking things, beating things and clanging things together in a rather caveman brutish way. Yet it works.

But sometimes if you have assumed All Things Being Equal The Errors In This Particular Measurement Are Random you may have invoked the Other Santa Clause.

The Difference in the Real World

Many scientific researchers deal with the real world i.e. trying to make measurements with tools that allow repeatability and reliabilty, and that can be characterised, calibrated and maintained to do this.

Others do not. They deal with a list of assumptions about data that must be put in place before they can investigate and analyse. And in doing so they are limited to the reach of those assumptions. They cannot extrapolate beyond them without explaining this context.

But sadly that doesn’t happen if there is enough pressure to be popular, or to advance a particular cause that there is money in. Or if your whole field has already assumed something to be true and this assumption is so basic and fundamental it doesn’t get mentioned that much.

It is either a sobering reminder that you don’t need to invoke each time…

Or it becomes the Emperor’s New Clothes.

I’m writing this particularly about the Overhead app outputs which are a filtered set of satellite conjunction data based off of the daily satellite track data (from Space Trak ) and the popular propagation software (sgp4). As stated in the app About page, there are a certain number of assumptions made and limitations recognised to use this approach to get an idea of a satellite’s operational baseline. As in just how many conjunctions occur per month or per year?

One aspect of the work is to guess what would happen if a regulator looked at the change in conjunctions and thought: “We need to invoke more oversight?”. Which may mean cost multiplication and the headache of that.

Maybe those close conjunctions didn’t happen at those distances? Maybe the risk is not as large as the current close conjunction would suggest and was a result of risk mitigation decisions made days before i.e. orbital corrections were already made.

Any projections or speculations that can be made from the data must always bear in mind the limitations of the data.

We must be cogniscant of the limitations of the tools used.

So when it comes to applying All Things Being Equal we CANNOT go beyond the limits of the tools used if we wish to stay consistent.

We cannot just assume all things are equal and that errors are random for say a temperature sensor that was scoped to read consistently and reliably (with maintenance and calibration) to +/- 0.5 degrees C. We cannot take measurements made with those sensors and average them to get greater precision (as certain mathematical theories allow) without stating and reminding that this is what we are doing.

And we cannot take this data and use it to determine actions to take in the real world without stating these assumptions. Nor can we proclaim that the “higher precision” data is varying and “things need to be done”

Because if we do we may end up invoking the Other Santa Clause.

And end up wasting gargantuan sums of money over nonsense and falsehoods.

Arguing the Other Santa Clause

What this clause deals with is often called the Logical Hypothesis in the definition of the Scientific Method. There is also the Experimental Hypothesis that requires measurements and is the one most would point to as “science”.

I didn’t know this for many years and had to look it up. It’s not always taught in such stark terms. But I did know that the majority of experiments and investigations in physics take place in highly controlled situations where you try and minimise the amount of assumptions. However, you end up with a mix of Logical and Experimental hypotheses.

So even if we invoke the Other Santa Clause, the degree of assumptions may lead to conclusions not that far-fetched even when we are talking about a magical character with flying reindeer.

We could, by right, have deep scientific or philosophical discussions about the range of tiling measures and costs for certain qualities of Santa and his reindeer. Maybe even quantum mechanics would get invoked to deal with the delivery time conundrum.

But most people, scientific and otherwise would not apply those conclusions to the real world.

Because you know, Santa Claus has certain qualities that preclude that.

But a lot of people WILL be fooled by invocations of the Other Santa Clause in lots of scientific fields popular in the mainstream.

And they’ll not realise that this is happening.

And if they don’t realise it’s happening, they may end up with the long-term bad consquences of that.

And then falling off a roof will be the least of our worries.

I am always aware that the outputs of the Overhead app may just be interesting noise. I don’t think it’s quite as bad as that. However I do think that in the coming years with the increasing amount of satellites being launched for low-earth-orbit constellations, these numbers will increase to such a high number that any bit of intelligence will be useful.

Because right now just from Starlink alone we get something like this increase in the last 3 years:

And what about the Subplot?

In the Santa Clause one of the main subplots is about how Scott’s son should grow up and stop believing in Santa Claus. This is enhanced by making the pressure come from his ex-wife and new husband. So even though Scott doesn’t believe in Santa, he’s willing to let his son do what he wants until it’s time to change. But the ex-wife (representing an antagonist of a personal nature) is trying to dilute his authority as the father.

The ex-wife and new husband also represent evangelising new ideas based off of beliefs that may or may not have a solid background. Yes it is natural for a child to grow up; for a boy to become a man. But how that is achieved and the theoretical basis of “the right way to do it” are not as singularly focused as the antagonists perceive them to be.

And just like these beliefs we have to be careful not to create subplots of our own that we are so adamant about, especially if they have come from a situation where we invoked the Other Santa Clause.

Because when we do, many more opinions and subplots have the same weight, which is not very much at all.

And just because you hold onto one so tightly, it doesn’t mean you are correct.

It’s the things you know, that just ain’t so that get you into trouble, as Mark Twain remarked.

And trouble can mean unnecessary expensive roof tiles for the whole city just because you are so invested in your subplot.

To summarise…

This applies to me and the Overhead app data as much as other fields:

Always be careful that you aren’t inventing reasons to take real world action based off of extrapolations beyond the limits of arguments. You will end up basing decisions on mythical entities.

And then worrying that they may fall off a roof.