Sorry I haven’t posted for a while. We had a sickness go through the home and I just didn’t feel like writing. This post is way more esoteric than the usual “how to save money” article, but it’s something I’ve been thinking about lately. Statistics are 100% true but only under the variables at which they were measured. Any application or inference drawn from a statistic is by definition false, because not all the variables can be accounted for. The application may be useful, but it’s not true like the statistic itself is. This is because chance doesn’t exist.
When a study is performed or survey is conducted, a specific group is sampled. The facts about this group are compiled into statistic. These statistics are true for that specific group at that specific time. The design is usually that the sampled group is representative of the whole and therefore applications drawn from that statistic may be useful for the whole. But since not everyone is sampled, the resulting statistic isn’t true for everyone. It is only true for the sampled group. It is only true for that specific group at that specific time. This means it is not 100% true for you now.
A Coin Flip
Here’s an example of what I’m saying. Statistically a coin flip is 50-50 heads or tails. If someone flipped a coin 100 times and got 50 heads and 50 tails you’d say that makes sense. Someone will then say, based off this statistic, that a coin flip has a 50% chance of landing on heads and 50% chance of landing on tails. But that’s not true.
The reason it’s not true is because chance itself is not a factor for causing a coin to flip or land. The factors include: your thumb, the amount of force you exert upon the coin, the angle at which you flip it, the density of the air it’s falling through, whether there’s a sudden gust, etc. All of those are variables affecting the process of flipping the coin.
We call it chance because we don’t know or account for all the variables (and sometimes we can’t even know them all!), but all those variables are deterministic in nature. Meaning if the variables are controlled for, the outcome is determined. If you are able to flip a coin with the same amount of force, at the same angle through air which has the same density etc. you will get the same result every time. This is because nature is deterministic.
In fact researchers at Stanford actually did this (Source). They built a machine that could flip a coin with repeatable force and angle and found that “With careful adjustment, the coin started heads up always lands heads up – one hundred percent of the time. We conclude that coin-tossing is ‘physics’ not ‘random’.”
So does that mean that you can’t use a coin flip to make a 50-50 choice? No, it works fine. Because humans are good at introducing extra variables into their actions, it’s a good approximation of chance. It’s close enough that it’s useful, but it’s not explicitly true that if you flip a coin it will have a 50% chance of landing on heads or tails. This is because chance does not exist. The world is deterministic.
80% of Students Change their Major
Another example is the statistic that about 80% of college students change their majors at least once sometime before graduation. This statistic is 100% true, for those specific students at that specific time. Roughly 80% of the people surveyed changed their major at least once. But there’s no true application that can be gathered from it. You could say that based on past results a college freshmen is likely to change his major, but you cannot say to that college freshmen that he has an 80% chance of changing his major.
The reason you can’t say that is because chance isn’t a factor of collegiate study decision. The factors include: interest, drive, stick-to-itiveness, finances, input from friends and family, etc. If this freshmen loves his current chosen field, and has the drive and financial abilities, He is likely to stick with it. He does not have an 80% chance of switching majors. Whereas someone who doesn’t have an interest in college and just signed up for whatever program their parents recommended is much more likely to switch majors.
What we call chance is just our inability to account for all the variables.
A quick sum up (so far)
These two examples have two entirely different sets of variables so their outcomes are going to be different. But still their outcomes are based on the variables affecting the mechanisms (factors) that drive results. The issue is that chance doesn’t affect results because chance isn’t a mechanism that drives results. What we call chance is just our inability to account for all the variables.
Statistics may be true, but the applications drawn from them are not. They have varying degrees of usefulness. Consider the application that if you flip a coin, you will get a random 50/50 result. That application is not true, but it is very useful because empirically the answer usually comes out close to 50/50.
The application drawn from the statistic that 80% of college students change their major, that you have an 80% chance of changing your major, is less useful. It’s not entirely un-useful, but the mechanisms that drive collegiate study decisions are much more varied than a coin flip. And there are many more variables that come into play.
Covid-19
Let’s now look at an example that is very applicable to everyone currently: Covid statistics. Why do some people catch Covid and some don’t? Why does Covid affect some people so badly while others are completely asymptomatic? We don’t know a lot of the variables that surround Covid infections, but we do have lots of statistics.
In the US 1.7% of Covid cases resulted in death. What application can you draw from that? If you said, “If you catch Covid 19 you have a 1.7% chance of dying from it” you have missed the point. The only helpful applications we can draw are that Covid is more deadly than other similarly contagious viruses that have lower mortality rates.
Saying that if you catch it you have a 1.7% chance of dying from it just isn’t true. It also isn’t helpful since it ignores all the variables at play. Early in the pandemic the mortality rate was higher, implying that a given person is less likely to die from Covid now than they were in April of 2020. We also know that variables of age and health are big factors. A young healthy individual is much more likely to survive Covid than an old fat one.
But either way neither one has a “chance” of dying of Covid because chance never killed anyone. There are factors behind death and those factors are affected by many variables. Many of those variables are still yet to be pinned down. For example there are healthy young people who have died of Covid and unhealthy old people who have been asymptomatic, and we don’t know why.
Purity of Science
What we do know with 100% certainty (assuming the statistical data collection methods are true) is that 1.7% of Covid cases in America have resulted in death. We also know that the world is deterministic. If the same variables affect the same situations the results are the same. This is why Physics is considered a pure science. There are relatively few variables. We know that if something with a known mass is launched from earth at a known force and angle, it will land at a known location. It’s very reproducible.
Biology and medicine are less pure sciences because they have more variables. If you set up the same experiment with two different people, you can have wildly different results because the human body has millions of mechanisms that are each affected by millions of variables. Social sciences like psychology or sociology make biology look like a pure science because of their massive amount of variables. But like anything else in this world, there is still no chance. The absence of chance is what makes science possible. The fact that the world is deterministic makes experiments repeatable. We may not be able to account for all variables, but we can draw inferences from them. We just need to be clear that the inferences are always untrue. They may be useful, but they are always untrue.
How does this relate to FIRE?
FIRE utilizes a lot of statistics. The 4% rule is based on statistics, the believe that the stock market will produce an average annual gain of 7% is based on statistics. The average retirement age, income, cost of living etc. are all based on statistics. These statistics may be true, but the applications drawn from them are not. They may be useful, and you can use them as guidelines to model your financial independence plans after, but everyone is unique.
According to the Trinity Study, if you have a portfolio of 75% stocks and 25% bonds and you withdraw 4% of your portfolio per year and increase with inflation, you have an 88% chance of your money lasting 50 years. But as I’ve been saying, that conclusion is false because chance doesn’t exist. The Stock Market has its own factors that affect how it produces returns. Also we as investors have our values and variables that affect the mechanisms by which we spend money.
If you are already good at adjusting your spending for when times are tough, a stock market correction early into retirement won’t affect you as much as it would someone who can’t reign in their spending. What I’m saying is you don’t have a “chance” of things turning out a specific way. You have control over it. Obviously there are things outside of you control, but nothing is outside of control. The universe is deterministic because God is sovereign. Control what is in your sphere of influence and leave to God all that isn’t (and all that is as well).
What do you think? Was this post too esoteric? Should I have stayed sick longer? Let us know in the comments below!