Thinking in Extremes Is The Easy Part

It's easy to figure out whether or not something matters, but it's often very difficult to figure out how much it matters... I leave the second part largely up to you

Note from the author: For some reason I was having difficulty publishing this article, so I apologize if you received 5 copies of it in the last 24 hours! This is the final version you will receive even if this final attempt doesn’t work.

I hope I am not alone when I often ask myself the question "does this matter?"

I ask this question when considering how different athletes affect the outcomes of a professional sports team, whether or not a certain feature might move the sales needle for a new product, if I should weigh different factors when thinking about a move to a new city or apartment, and so on. Really, you can ask this question when considering the characteristics impacting any decision or assessment. More often than not, almost everything matters, and thinking about the extreme ends of the spectrum for whatever facet, component, or quality you are assessing— the best and worst version of your variable— and how it affects the outcome usually makes this very clear. The hard part is figuring out how much something matters, which, for some situations, can require more than a lifetime to properly asses, but I’ll touch on this briefly later in the article.

My current favorite pet example for deploying extreme thinking in the world of sports is the running back position in the NFL. If you tune into most sports news and pundit shows, you’ve probably heard something along the lines of “the running back position doesn’t matter,” and you’d also see that feeling reflected in the running back market in the NFL, where the general consensus says the top RBs are underpaid relative to performance. Most NFL media personalities and teams themselves believe a good offensive line and quarterback make the strength of the running back irrelevant (to winning), but I’d suggest that at least on some level it does matter.

Thinking in the extreme, we can compare my ability to perform as an NFL running back vs. one of the GOATs, Barry Sanders. I’m 6’0”, 165 lbs (mostly bones), and Sanders stood 5’8” with 200+ lbs of muscle. I can’t even pretend to think I’d be able to gain yards in the NFL— this comparison makes it pretty clear in some form that who your running back is affects the outcomes of individual plays and overall games. Of course, this example is maybe too extreme, but even if you just carve out the population of professional running backs— NFL game outcomes are likely to be very different with Barry Sanders on your team as opposed to a perennial practice-squad back. This line of thinking then prompts questions about how much more impact Barry Sanders brings, which I’ll touch on a bit later.

I don’t know why, but when I was thinking about the example above, all I could picture was me as the cinnamon stick guy and Barry Sanders as the apple from Apple Jacks commercials. The body types seem too fitting.

In the world of sports tech and product design, the question of how much time to invest in the aesthetics or form of a product always arises. The amount of abuse by the environment and users in sports-centric settings often drives an inclination to discount the important of how a product looks. The Guardian Cap (which I’ve covered multiple times) presented a concrete example of this tendency. After a pretty long battle to achieve adoption, many NFL players wear Guardian Caps in games now, but every player I’ve seen wear one also uses the cloth cover that makes them look like normal helmet. I think if the company had considered the two extreme aesthetic ends of “make the Guardian Cap look like a normal helmet” or “make it look really obvious you have padding all over your helmet” (like they originally did). I think they would have chosen the latter option from the jump.

A non-sports-related (sacrilege, I know), but highly illustrative example comes from the world of construction. If you scoured the globe, you likely wouldn’t see many, or any, large construction machinery with bright pink paint and finish. Even though large construction and farming equipment should be easily visible (i.e. be brightly colored), you can immediately think about the users of the equipment and the wear on the machinery and confidently assert that magenta equipment wouldn’t sell well. This extreme end of the color spectrum reveals that when most people say “oh I’ll take [insert product here] in any color,” they don’t actually mean it— even simple aesthetic choices do matter, potentially a lot! Just ask Caterpillar, who made their branded yellow so effective for sales and longevity during use, that the term for most large construction machinery is “yellow iron” or “yellow goods.”

Continuing outside the world of sports (further into the sacrilege and my own personal hell of trying to weigh the pros and cons of moving to a new apartment next year) you can think about how commute time would affect willingness to live in different areas around a city. While I’m in a city, I’ve previously thought I’d just want to have a 1-way daily commute under 20 minutes. But thinking again to opposite ends of the travel ease and safety spectrum, my personal happiness would definitely be different if my commute was a 20 minute bike ride through well-partitioned and plowed bike lanes versus a busy bus/subway trip with transfers— so much so that this exact thinking is severely limiting the apartment search for next year in the Boston area, driven predominantly by our “interesting” public transit network (a long-form video for any urban planning nerds out there).

Hopefully, the examples above help demonstrate thinking in extremes as a tool to assess the high-level importance of an individual decision variable, and spark ideas about some applications in your own life and work!

However, as the title of this article states, the process of thinking in extremes is the easy part. The more difficult part is figuring out what the curve between both ends of the spectrum actually looks like when you plot your independent and dependent variables (e.g. running back ability and team wins, or helmet aesthetic quality and number of users). Plotting these with high levels of accuracy can be very difficult or sometimes impossible. Again, though, I think there is a quick-and-dirty exercise you can use to gain at least some insight— simply estimate what you think the curve will look like based on your intuition or any rough data you already have! I think you’ll find you typically end up with one of a few estimated curve types, plotted below:

Importantly, there are certain scenarios where different points along the spectrum are nearly indiscernible in their effect for many of these curves/regimes. Maybe running back impact is a log curve!

Within each of these functions (linear, log, step, power) are infinite variations of order of magnitude, longitudinal scale, number of steps, etc. that require understanding to see the full precise picture, but that aren’t necessary for rough estimates of impact. In some cases, your estimated curve might just contain one half of the spectrum (e.g. just a positive power function), or it may take on a combination of functions. Perhaps running back impact does follow a log curve and the differences at the most elite level don’t matter much, or alternatively, like my analysis showed for the NBA, having a superstar on your team is critical to winning championships! It likely wouldn’t take much data to figure out these insights, and in the absence of data, don’t overthink it, especially for this level of rough assessment.

So beyond the ability to clearly show that most characteristics do matter at the extremes of a spectrum, I hope you see value in being able to inform slightly more nuanced decision making quickly by predicting the curve or function that actually composes the impact of the variable you are considering. I am absolutely not advocating for using this 2-step technique to make really high-impact, important decisions without some sort of quantitative data to back it up, but I really do believe asking the question of “Do the extremes show this matters?” and using our intuition to answer “What impact curve do I think this follows?” can reveal many insights about the scenario and also about our own thinking.

And of course, for the sake of brevity and to make my flight home from vacation, I’ll leave the quantitative analysis about how paint color affects back hoe sales in your capable hands! See you back in the States!

I had to include an image of a trail run from last weekend in Zermatt! I think we may have been the last people to be able to complete this trail without spikes or snow shoes for this year… When my views and activities look less like this I’ll have more time to write!