Some outcomes “feel” more successful than others. The underdog beating the favorite. Meeting a tight deadline despite multiple obstacles. Getting a good grade on a test that most people struggled with. In my personal life there are a set of accomplishments I look back on with pride, even though they were relatively minor. While there is a separate set that I feel ambivalent about despite them being considered more traditionally successful. So what distinguishes which accomplishment falls in which set?

I used to frame accomplishments in terms of expectations. Those that felt the most/least successful were because they had a large distance between expectations and outcome. But framing in terms of expectations doesn’t translate very well to communicating with others. It’s difficult to agree on the set of expectations and people usually have a very personal response if their expectations are much higher/lower than yours. This comes into play in all sorts of interpersonal relationships. Parent-child. Teacher-student. Peer-peer. Boss-employee.

Is there a better mental model we can use for communicating and understanding what outcomes should be considered successful? And I’m not talking about quantifying with metrics. Metrics don’t address the qualitative nature of success, which is equally as important as the quantitative.

There is a mental model called the Monte Carlo generator that I picked up by reading Nassim Nicholas Taleb’s Fooled by Randomness that I believe is extremely applicable for classifying success. I won’t waste space elaborating on Monte Carlo methods because other resources are much better at explaining it than I could. I usually associate Monte Carlo methods with being forward-looking. While a Monte Carlo generator is backwards-looking. Given an outcome, we can “generate” the Monte Carlo simulation that led to that outcome.

By being able to “generate” a fictitious Monte Carlo simulation for any observed outcome, we attach a hypothetical probability distribution to that observed result. This lets us frame the result in terms of where it fits in an underlying probability distribution. Probability distribution may not be the best term here, because it’s usually not a quantifiable experiment we can run. A better term is “alternate timelines”. We are framing the observed outcome in terms of the range of possible outcomes in alternate timelines.

That’s kind of a lot of heavy concepts, so I’ll elaborate with a personal example. For me personally, graduating high school did not feel like a laudable accomplishment. You can think through that in terms of the expectations vs. outcome lens. I always expected to graduate high school and I did, so there was no distance between expectations and outcome. Or you can think through the Monte Carlo generator lens. Given that I graduated high school, I can generate a Monte Carlo simulation with the same initial conditions (stable two-parent household, middle-class, white, male, midwestern values and influences) and most of the same actors and factors (good friends, good study habits, good mentors, etc). In those simulations, I graduate in 95/100 simulations. The outcome of graduating high school happens in 95% of my alternate timelines. I wouldn’t really consider myself lucky or accomplished if I am achieving what is essentially a foregone conclusion.

I obviously don’t want to detract from anyone who did feel accomplished graduating high school. It certainly can be a noteworthy accomplishment. But for others to understand that, it helps them to understand where you think you stand in your alternate timeline distribution.

Taleb elaborates in his book on how this type of framing helps distinguish signal from noise . It helps determine if someone is actually skilled, or if their success is attributable to randomness. People who aren’t able to utilize the Monte Carlo generator model are likely to be deluded and deceived in their everyday life. Taleb lists a ton of good examples in the book. I’ll link the psychic sports picker scam here as one of the most straightforward and illustrative examples. Although don’t think this delusion is exclusive to scams and con-artists. There are tons of examples in everyday life and once you start viewing the world with this lens, its impossible not to see them everywhere.

A couple closing thoughts I think are valuable arising from the Monte Carlo generator model. When you’re evaluating success with this framing, you should really only care about two things. Everything else is randomness.

  1. Measure things that indicate your position relative to the alternate timelines. These are the most valuable metrics you could have. Find and measure them if possible.
  2. Care about things that increase your chances you’ll end up in a good timeline.

Grades are not important. Study habits are. Weight is not important. Fitness habits are. Revenue is not important. Revenue growth is. Shipping new features is not important. Short feedback cycles with customers is.

Just like with any mental model, don’t stretch the Monte Carlo generator too far. It’s not the only lens to view and explain the world. Every analogy breaks down at the margins. But paying too much attention and care to things that don’t frame well through this lens may be an indicator that you’re getting fooled by randomness.

Thanks to Ryan and Sarah for reading drafts of this post.