About a week ago, I sent each of my contacts on Google Hangouts the following message:
"You are competing against all of my other Hangouts contacts. Vote for either avocado or cucumber. You earn 1 point for being in the minority."
But first I should probably explain how I got there.*
📷📷
The Origin Story
As you may know, I took a trip to Washington, D.C. recently; two weeks ago I whined about waiting and took pictures of funny signs in a post about it.
Anyway, I was on an Uber** and, on the back of the passenger seat, there was an iPad. Like, it was attached somehow to the seat so the people in the backseat, such as me, could see it. Anyway, the point is, the iPad had some sort of trivia game on it. It took me a while to figure out what the objective of the game was. At first, I thought it was something else, and I will explain what I thought, which became my idea.
📷
I imagined a game where you voted between two things (say, avocado and cucumber) and you earn 1 point if you are in the majority. As a concrete example, you vote for avocado. It turns out that 11 people, including you, chose avocado. 6 people voted for cucumber. Since 11 is more than 6, you earn 1 point. So, the idea is you're voting to see what most people think, but the thing is, it includes you. I was very excited about this game idea.***
In The Minority
I also pondered the opposite version–where you voted to be in the minority. Then, in the previous example, you lose if avocado is ahead 11-6. Your goal is to vote for what you think fewer people would vote for. I wasn't sure which version was better, but I liked where this was going.
Then I wondered what would happen if people formed teams. In the majority version, if over 50% of the people went avocado, the game becomes boring, and everyone eventually goes for the avocado. But in the minority version, the best alliances are where people split their votes. If you vote avocado, you want to convince other people to vote cucumber. So not only is it in their best interest, it's in the game's best interest too.
I was convinced the minority game was the way to go.
📷
Potential Problems
I was pretty confident so far, but there were some issues I had in my head. For example, what if it was a tie right before someone votes? If they voted last, then no matter which way they vote they lose. However, this problem seemed like it could be masked by simply keeping the votes secret until the votes were counted.
Also, what would happen if the round ended in a tie? That had an easy fix. No one gets any points! It's not that fun, though.
Anyway, the rest of the real issues remained to be seen until I playtested, when I also found fixes for the previous two problems.
A Few Cosmetic Things
A minor problem with this game was that it needed a name. All good games have names. I'm currently referring to it as Avocado vs. Cucumber.**** But it's gone through several names already, such as The Voting Game, The Minority Game, and Be in the Minority. If any of you have a good name, please contact me.
Also, I am having fun thinking of round topics. Dogs vs. cats, bath vs. shower, piano vs. guitar, baseball vs. football, purple vs. yellow, magic vs. technology, flying vs. invisibility, shirt vs. pants, floor vs. ceiling, non sequiturs vs. turnips, you get the idea. See how many you can think of.
📷
Playtesting
I realized eventually that I needed to just playtest the game, because otherwise it would have no publicity, and I wouldn't really understand it. I tried to mock up a game on Scratch,***** but that kind of failed because it needed a multiplayer interface. I wasn't sure the best way to do this, but I eventually decided I could run it over Hangouts. The key was to message everyone privately. So I sent everyone the message from the start of this post. And that's where the story started.
Thanks for reading what I wrote. Stop by for Part 2 to learn what happened when I tried out the game.
–beautifulthorns
Originally aired April 24, 2019
*And to do that, I will use section headers, which I'm testing out in this post.
**I don't recall where it was headed.
***Unfortunately, this game sorta kinda already exists. (https://en.wikipedia.org/wiki/El_Farol_Bar_problem#Minority_game.) I hate echoes. But I'm adding improvements, and discovered it independently anyhow; you'll see how in Part 2.
****A side effect was that I was going to title this post "Avocado vs. Cucumber: The Game," but then when I realized it was a two-parter, I had to change it since I think that "Avocado vs. Cucumber: The Game: Part 1" looks weird.
*****I don't know how to program anywhere besides Scratch. (Depending on who you ask, that means I don't know how to program.)
----
It's time to revisit the battle for the vegetables.
We've got a lot to get to, so let's jump in.
The Abstain Option
I mentioned in Part 1 how I had foreseen a problem with votes being public, where it would feel bad to be the last person to vote, as you had no way of winning. What I failed to realize was that it felt bad even if you only found out later. For example, if you vote avocado, and avocado is ahead by a 7-6 margin, the cucumber voters earn a point. You lose. But if you switched to cucumber, then that swings the round to 6-7. Then the avocado people win. So you lose either way. You get to pick who wins, but you lose.
📷
Generally, games should be as simple as possible. But I realized that this was a big issue with Avocado vs. Cucumber, so I needed to think of a solution. After considering this for a while, I thought of a solution, in the form of a third option: abstaining.* When you abstain, you earn 1 point if and only if the margin is 0 or 1. For example, avocado wins 4-3. The margin is 1. You get a point. If cucumber wins 6-5, the margin is also 1. You get a point. However, if avocado wins 7-4, the margin is 3, and you lose.
The idea behind the abstain option is that, no matter how everyone else votes, there is always precisely one thing you can vote for that will guarantee victory. So far, it's done its job well–to make people feel like the odds weren't stacked against them. The abstain option reminds me of insurance in blackjack,** in the way that it can make you feel like there was a way to win, even if you shouldn't do that most of the time.
The Winner Keeps Winning
In this section, I will discuss the play patterns I noticed. You can look at the raw data here.
First of all, avocado won three out of three rounds against cucumber–and by bigger margins each time. Common sense would dictate that if avocado wins Round 1, people anticipate that it will win again, and switch to cucumber, and that this will cause cucumber to win Round 2. However, many people seemed to think this, and avocado won by more decisive margins in Round 2. It won again in Round 3. This suggests to me that people anticipated that the switch would happen, but it never did.
Additionally, people are stubborn. About 70% of the time***, people kept their vote the same instead of switching from round to round. Maybe this is a product of what I was talking about, or maybe people are just stubborn and hope that eventually, they will be right.
📷
However, I suspect that a good strategy is very much a product of the conventional wisdom people use when deciding to vote. If people decided that it was a good strategy to hold strong, people would switch to the losing side, causing the opposite to occur. So it's very much about trying to figure out what other people think. The mistake many players made was to not realize that other people probably had similar instincts to them, and to choose the opposite.
Therefore, my #1 piece of strategic advice is probably to realize that other players are smart too, and will think similarly to how you think. So my technique at this game might be to figure out my instinct, and then vote the opposite way.
The Losers Keep Losing
Another pattern for Avocado vs. Cucumber is that a lot more people lose a round than win it. While I didn't receive any feedback on this, I glossed over it in Part 1 and want to talk about it.
Out of 5 rounds, the highest score achieved, among 13 participants, was 3, whereas three different players earned 0 points. Theoretically, if you conducted this randomly, each person would have a 50% chance of winning (setting aside abstain and ties for the moment.) This effect might make the game feel like an uphill battle. However, when you consider everyone, more than 50% can't win.
📷
Presentation
It's been shown that presentation can have an effect on how people vote in some contexts. For example, in the image above, the cat looks angry, making the dog look cute in comparison. Or if a poll of the Democratic candidates for president always puts Pete Buttigieg first, many people will simply pick him. So in the last round I experimented a bit. For approximately half of the voters, I told them the round was "cats vs. dogs." For the other half, "dogs vs. cats." This is in no way statistically significant, but I wanted to mention that two dog voters both came from the second camp.
Rock-Paper-Scissors
In this game, the first round with a topic is some form of guessing how popular things are.**** Subsequent rounds involve trying to predict what other people are going to do. In that way, it makes Avocado vs. Cucumber akin to rock-paper-scissors.***** It's more complex, since there are more players, and more nuanced, since you are trying to predict the behavior of a group instead of a person. However, many of the same ideas are involved, like reverse^k psychology.******
📷
I believe that in Avocado vs. Cucumber, just like in rock-paper-scissors, rounds after the first provide a more interesting test of psychological knowledge. Though I've received conflicting feedback on this, if I played again, I would do avocado vs. cucumber for every round.
Final Thoughts
Through this endeavor, I've learned that:
- Like dreams, instead of letting ideas go, you should think about them, and see if you can find a way to make them interesting.
- Playtesting is important to a game, since it catches issues you would never have thought of otherwise
- Listen when players have feedback on your game (e.g. abstain) instead of dismissing it.
- Avocado vs. Cucumber is akin to rock-paper-scissors.
- As far as specific strategies for this game, generally you shouldn't abstain.
- Remember that other players think similarly to you. This makes one strategy to go against whatever your instinct is.
Thanks for reading what I wrote. And finally, it may have won/lost/received more votes three times, but I'll be an avocadoholic******* forever.
–beautifulthorns
Originally aired May 01, 2019
Next: AbstainTrain
*It was originally called the reporter, since reporters like close races, but that turned out to be confusing (people thought they were choosing a different participant to be the reporter) so it got renamed.
**If you aren't familiar with insurance, it's a bet you can make if the dealer shows an ace that's independent of the main game (meaning whether you win or lose is irrelevant.) For each $1 you bet on insurance, you get $2 if the dealer gets blackjack. Insurance is not usually a good option.
***I ignored when people abstained in Round 3 since the option was unavailable in the previous 2 rounds to calculate this statistic. I also ignored Rounds 4 and 5, due to them having a different topic.
****Though really, you should be guessing how popular people think it will be, and then do the opposite. Though not everyone votes that way. Aah, this is horribly meta. I mean amazingly meta!
*****If you aren't familiar with rock-paper-scissors, then you're one of today's lucky 10,000! Go read this. Also, yes, Avocado vs. Cucumber is also similar to rock-paper-scissors-lizard-Spock.
As a sidenote, that probably wasn't a good advertising line. However, rock-paper-scissors can have lots of strategy, while more about psychology than game theory.
******That means reverse psychology or reverse reverse psychology or reverse reverse reverse psychology or... you get the idea.
*******Etymologically, the "hol" in "alcoholic" refers to alcohol, not being addicted to it. So constructs like "workaholic" and "avocadoholic" don't really make sense. Then again, this is English. It doesn't make much sense logically.