Ever since I started working at ArgonDigital, I have been very interested in finding ways to apply the visual models and techniques we learn on the job to other real-world or personal situations. Early on, I learned about all of the visual models we use in RML®. When I first heard about Decision Trees and their counterparts, Decision Tables, I couldn’t help but remember the “Choose Your Own Adventure™” books that I read as a child.
If you’re not familiar with the “Choose Your Own Adventure™” style books, they are usually short children’s books (about 100 pages) where on every other page or so, the reader is given a choice of where to go next in the story (with the decision taking them to two different pages in the book). I used to love these books as kid, getting to choose what ending the book had and being able to “participate” in the book more.
Sometimes, I would make silly decisions and my book would end abruptly; other times, it seemed like I would never finish the book! So when I heard about using Decision Trees and Decision Tables to map a decision making process, I thought it would be an interesting experiment to map out the decisions in one of the “Choose Your Own Adventure™” books.
To do this, I read “The Antimatter Formula” by Jay Leibold. In this story, you are a young child, whose parents have been working in antimatter research. You wake up one morning to the whole world being empty of people except one: a nefarious doctor who stole a secret antimatter device (in the form of a TV) from your parents.
The cover of the book claims 32 different endings, so I went through and followed every line of decision making to reach all of the possible endings. Then, I mapped these decisions and outcomes in both a Decision Tree and a Decision Table to experience firsthand the advantages and disadvantages of each model.
Before we go further, do you want to learn more about Decision Trees or Decision Tables?
If you want to read more about Decision Trees, go to paragraph 1 below.
If you want to read more about Decision Tables, wait for part 2 of this blog post, coming soon…
1. You chose Decision Trees, a most worthy place to start. In fact, it’s where I started when making these models! As I was reading “The Antimatter Formula,” I quickly realized that this Decision Tree would get a little unruly. Partly, this is because of the nature of the decision (every decision is a one-or-the-other-type decision).
With multiple choices at each decision, the Decision Tree could be simplified, but that wouldn’t follow the nature of the book. The Decision Tree mapping every decision in the book to the 32 outcomes is shown below. (Because of the size of the Decision Tree, the image below is used mostly to illustrate how quickly these can branch out. If you would like the full version, please email us.)
A little crazy, no? When this Decision Tree was complete, I saw that there were four basic outcomes to the story, despite there being 32 ways to get there: you could end up back home with your parents, be stuck in a parallel world in a positive way, be stuck in a parallel world in a negative way, or die in a parallel world.
I could see that some outcomes occurred more frequently than other, but couldn’t do a thorough assessment of the outcomes in this form. It was also difficult to choose an outcome and make my decisions in such a way as to effect that outcome.
Below, I’ve made a list of advantages and disadvantages of using just the Decision Tree to model the book.
- Easy to ensure completion: As I was making the Decision Tree, I would add the choices to each decision box as I came across it, and thus it was easy, at a glance, to see where I still needed to go back and fill in the tree.
- Easy to see minimum and maximum number of decisions: After the Decision Tree was completed, I could rather easily see that I could finish the book in as little as 3 decisions or as many as 10 decisions.
- Visually pleasing: The Decision Tree is easier to digest as readers as we can follow each line of decision making and see where each decision will take us.
- Lack of significant analysis of the endings: From the Decision Tree alone, it was difficult to understand how many endings were happy/unhappy, how I could end the story back home, or how likely I was to choose an ending that would have my character dying.
- Lack of backward traceability: With the Decision Tree, it was difficult to choose an ending (say, back home with parents) and then know what decisions to make along the way to ensure that ending.
- Size: The Decision Tree, with all of its branches, got really big really fast and there wasn’t much I could do about it without making it difficult to follow.
Return next time to learn more about Decision Trees and Decision Tables…