# Algorithms for better decision making

Well, good news for any indecisive people out there, there’s an algorithm for it!Ideally, whenever you make a decision you’d want the best outcome with the least amount of regret.

There are two ways you could approach this, one involves calculations, which isn’t ideal when picking the best bag of chips at the grocery store, but can be useful when you have time to measure out which option would be best.

These options are optimal stopping and Bayesian Bandits, respectively.

Though both have the same idea of exploring and exploiting, though exploiting in this case means reap the full rewards from, rather than selfishly manipulate someone.

the stormy world of algorithmsHow can you hope to make an informed decision without a baseline set for comparison?.Well, you can’t, not if you want the highest chances of the best outcome.

Using a little ditty called optimal stopping, you can figure out when to take a specific action for the best outcome.

So, if you have a set amount of something to choose from coming at you one at a time, which one should I choose so that I have the highest chance of choosing the best option?.This can be used for the hiring process, choosing a house/apartment, or even finding your soulmate.

It’s divided into 2 phases, the exploring phase, and the choosing phase.

You should explore for a set amount of options to establish your baseline, then chose the first option that surpasses your baseline after it’s defined.

But wait, how many options should I go through to set my baseline?.How can I prevent looking too much or too little?.Well, optimal stopping says you should go through 37% of your options with no commitments, just to establish your baseline, and this maximizes your probability of picking the best out your pool.

And what’s your probability of picking the best out of your pool with this method?.You guessed it, 37%, now that may not seem high but with our tendency as humans to look for too long or too little, that resulting probability would be much lower at around 11%.

Now we can get infinitely more complex trying to consider every single unpredictable aspect of real-life, but that’s impractical.

So once our baseline is set and we’ve been through 37% of our options, the first option that exceeds whatever we’ve already seen is what we’ll chose to have our 37% chance of choosing the best possible option.

But Romy, how can I choose a soulmate out of billions?.You don’t, you set the pool amount yourself because if you don’t you’ll be dating until the heat death of the universe.

Same thing with house hunting, obviously you’re not going to check every house in the nation, but you’ll select a set amount then explore those options.

The other option, Bayesian Bandits, is used more by companies and business when exploring implementations of their product.

It stems from a problem called the multi-armed bandit problem in which you face several slot machines with different payouts, and the question is how can you find the one with the best payout rate while maximizing winnings?.You can start to see why businesses use this often because you can have many iterations of a product, or many presentations, and as a business you’d want to use the one that maximizes profits.

It’s also used within the medical industry when performing clinical trials.

A simple version would be an A/B test, you have two options, the drug and the placebo in the case of clinical trials or 2 different website layouts to test.

Then using the collected data, you’ll be able to calculate the probability which option is better in the long run.

It becomes more complex with more arms to pull and more factors considered, hence the practicality of using this for day-to-day life is not too beneficial for the average joe, they’d benefit more from the optimal stopping as it’s already calculated for you.

I can write another article explaining the complexities of implementing the Bayesian Bandits and the mathematical equations of it, but the general idea is testing the probability of each arm and going with the one with the highest success probability.

source: https://giphy.

com/Algorithms can be extremely complex and have some high level mathematics, but they can be useful in day-to-day life, though most of the time it might not be worth it to compute complex equations for choosing a shampoo brand.

Algorithms can simplify the complexities of life as well, because life just has so many moving parts and variables, it can be humbling to think about things in terms of concrete answers rather than in hypotheticals and infinite possibilities.