I have a serious problem with perfectionism
I was just smacked in the face with an important lesson today. To summarize, perfectionism can be just as evil as full-blown procrastination if you’re not careful.
Why do I say this? Well, I think this is especially important for engineers because based on personal experience, people in this field are prone to such problems.
Let me give you an example.
Every Tuesday I host a movie night. Each movie night I cook food for my guests and I get them to vote on a movie in our local discord. Seems great, it worked great, people showed up, no major complaints.
Every week has a decent turn-out, but one day it occurred to me that our voting system is mathematically flawed, and therefore, can become better.
But it’s not just “can” become better for your inhouse perfectionist. It’s “should” or “will” become better. The question is not if it could be done the question is how soon will it be done.
So how did I set out to improve my voting system that was working perfectly?
Well, let’s start by pointing out how single-elimination voting systems tend to devolve into binary outputs. (just look at the two-party system in America)
People know that outliers have no chance of beating the major nominations, so in an attempt to make their vote matter they only vote on the major nominations. Then they campaign to get even more people to vote only on the major contenders.
Any vote that goes into an outlier nomination is a waste of time because they have no chance of beating a major contender. During the only elimination phase of the vote, outliers will most certainly lose and the votes they received would be lost. This is called first-past-the-post voting.
The unfortunate result from a first-past-the-post system is that the winner tends to be one of only two major players and it tends to be the side that can afford a greater campaign in order to get the largest push for support at the last second.
Take Trump, Clinton, Sanders for example. If all three candidates made it to the final vote. Trump goes in with 39%, Clinton with 36%, and Sanders with 25%.
Trump wins right? But he only won with 39% of the total vote, and most of Sander’s voters would have voted for Clinton as a second choice. The result is the vast majority of the voter pool would not have voted for Trump, and in fact, actively did not want Trump to be president.
So that’s the problem with first-past-the-post. The way we fix it is by implementing roll-over votes. Every voter votes on each candidate by putting them in positions of preference, when a candidate gets eliminated the voter’s votes roll over to the next person of their preference.
Why this was a bad idea
I did it! I decided hey let’s just implement the mathematically superior voting system. I asked my friends to put all of the nominated movies in order of preference. What was the result? Nobody voted or showed up to the movie nights at all until I decided to discard the new voting system.
What did I learn?
For starters, what the hell am I doing spending several hours trying to optimize a simple voting system for movie night? It’s not worth my time.
For second, the extra effort needed in order to vote on such a trivial thing was just enough to de-incentivize my friends’ participation.
This might be important to implement if we were voting on which kid we needed to starve in order to supply food for the other children, but for the purpose of a movie night, I was better off just picking a movie like some kind of movie night dictator.
What did I do to improve the situation?
More food. It’s crazy how effective food is. If I cook more food and advertise the food, and practice my cooking skills, more people show up. This has remained true for months. The other details are comparatively negligible.
In short, Save yourself some time and effort and start appealing to people’s basic needs. Leave the efficiency for your astronomical goals.
