An Analysis of Human Decision Making

Everything in life can be boiled down to a single equation. In a sense, humans are nothing more than more complex (or less complex?) computers. Originally, the word "computer" was actually used to describe humans - "human computers" carried out calculations and computations. Alan Turing said in 1950 - "The human computer is supposed to be following fixed rules; he has no authority to deviate from them in any detail."

What if we could come up with a way to make all of our decisions like "human computers". I will argue that it is in everyone's best interest to make every decision in their life based on the effect it will have on their happiness. That is the "fixed rule" of life. Everyone lives and makes decisions for one single reason: to increase their own personal happiness such that the pros outweigh the cons. You decide to eat a sandwich when you are hungry? You're doing this because removing the discomfort of hunger will cause a net increase on your happiness. You decide to follow Christianity? You believe that, by your own assessment of probability, it is possible God exists and spending eternity in Heaven will cause eternal happiness rather than eternal suffering in Hell, which outweighs all negative repercussions of being a Christian. You decide to donate money to a disaster relief charity? You believe the personal satisfaction of donation outweighs the satisfaction that that money could give you otherwise. There is no good or evil (in the definitive sense of the world), there is only a bunch of computers walking around, attempting to maximize an internal variable representing happiness. Every decision you and I make is meant to maximize our net happiness, something I will, from here on out, refer to with the variable NH. This is true for every single decision. Big or small. Any given second of your day, you are making a variety of decisions, and whether you know it or not, every single one is made (in your assessment) to maximize NH. Please allow me give some examples:

  • You are sitting at your desk at work and your forearm itches. Should you itch it or not?
    • A few pros of itching:
      • Relief of discomfort.
    • A few cons of itching:
      • Itching will take a small amount of physical exertion.
      • It will lead to a short period of productivity loss. There is a very small chance it leads to distraction causing even more productivity loss.
      • Sometimes itching causes an itch to increase in the future.
  • You are asked by your sister to babysit her daughter. Should you babysit or not?
    • A few pros of babysitting:
      • Taking care of the child means she would likely recognise your kindness and be more willing to do something for you in the future.
      • Denial may come off as a slight which could mean negative feelings that could impact your relationship with her in the future.
      • Babysitting would cause some contentment based on the fact that you helped someone close to you with a problem they had.
    • A few cons of babysitting:
      • Relaxing would feel better at the moment.
      • Could set a precedent that you are willing (and maybe even eager?) to help out with things like this in the future.
A few notes on the previous examples: Weighing pros and cons can get really, really complicated really, really fast. While it's pretty easy to come up with almost all notable repercussions of a small decision like whether or not you should itch your forearm, when it gets any more complicated it seems almost impossible to go through even a fraction of the possible repercussions. How do we get around this in decision making? I think the best way to go about this is to narrow it down to the most important repercussions, where a "repercussion" is an outcome, resulting from the decion, that will alter your NH. Say there are "n" repercussions to a decision. Most likely "n" is huge - it could be millions, but one could argue the vast majority of repercussions don't even matter; you only need to come up with and analyze the most impactful repercussions such that successful analysis of those few would lead to a decision that would, beyond a reasonable doubt, be unable to be flipped from negative to positive or positive to negative based on the net change resulting from all other repercussions.

Additionally, why do we never consciously analyze decisions like the first example? Most decisions we make have a set of possible outcomes that present such a small impact on our overall NH that it is more damaging to our NH to spend time thinking about which path to take than it is to just make a quick decision. This is the case with subconscious decisions. We don't think about if we want to scratch an itch on our arm, our internal computer decides, before we even need to waste time thinking about it, that it is more worth it to just carry out one of the options of safe decisions.

Equation analysis: In my notation, NH is a spectrum. The negative end of the spectrum goes from slight discomfort to unignorable, unfathomable pain, while the positive end goes from mild satisfaction to absolute euphoria, while zero on the scale signifies complete indifference, a total lack of suffering as well as lack of joy.
I think every decision can be made based on the evaluation of the following equation:

NH = IL + p · ER
  • NH: "Net Happiness" - the variable discussed in detail earlier. If the calculated NH value is negative, the decision should not be made OR should be weighed with the indecision.
  • IL: Lump sum of net happiness loss as a result of the decision. This can only be a negative number. It represents a numerical value for the "cons" of the decision. This number must be scaled by time. E.g. If I lose 5 happiness points per day, and the payout is not for 30 days, then IL = -150 total for this decision. Note that for any even somewhat complicated decision, this will actually be the sum of all detrimental repercussions of the decision, not just a single number.
  • p: Probability of receiving the full expected payout down the road (must be 0-1). This is just 1 if the payout is immediate and certain, but if the decision makes use of delayed gratification this is important to take into account.
  • ER: The expected, eventual lump sum happiness payout of the original decision. Be sure to scale this by time. E.g. If the initial decision’s full payout will be an increase in happiness of 5 points/day for 7 days, then ER = 35. Note that for any even somewhat complicated decision, this will actually be the sum of all net positive repercussions of the decision, not just a single number.

This all might seem dumb or obvious; I am saying we make decisions to be happy? Duh! At the very least, though, I think this is actually a great way to analyze difficult decisions. On a grand scale, however, imagine if we could create a way to accurately calculate and make decisions based on this. A computer could make every single decision for you in your life. This would allow you to go through life in an optimal fashion.

Notes on using this equation:
  • "Machine learning" (learning from experience) should be utilized. Learning from the outcome of one decision will impact decisions in the future. Due to this, new decisions should likely be scaled higher, as they provide more data for your machine learning.
  • Likely more important than any physical decisions for NH are intangible decisions, such as what to spend time thinking about.

  • Are the NH values of others important whatsoever beyond their influence on my own NH (in the fact that others' happiness increases my own happiness)? If so, doesn't that contradict the assertion that one should not be relying on others for my own happiness? If there is a middle ground, how do I define that spectrum in relation to my current idea?
  • Is it feasible to think this equation could ever be implemented beyond just a thought exercise or a method to guide decision making?
  • Could I come up with real units for NH? Four major chemicals in our brain affect happiness: dopamine, oxytocin, serotonin, and endorphins. If there was some way to measure the actual chemical effect of decisions we could make this more tangible useful.

Try it yourself:

Variable Value