The best will get worse and the worst will get better over time.
Performances of sports teams, business metrics trends or even the states of our own romantic relationships over time — the outcomes we observe in all of these areas depend on two major factors: the real underlying trend and some random fluctuation around it; luck and bad luck.
If we observe a player play a never-before-seen phenomenal season, they will likely play worse next season. If our business numbers are beyond expectations in one year, they will likely not sustain the same level next year. If our relationship is plagued with a series of conflict and arguments, it will likely get somewhat better over time.
Whenever someone performs exceptionally well or poorly, luck has likely played a role. This means that outcomes are not perfectly correlated with ability or skill. As luck itself is not stable over time, and the random fluctuations around the trend tend to equal out over time, extraordinary outcomes will eventually regress back to their average trend.
For example:
I perform extremely well which is partly based on luck. As this luck isn't stable and tends to even out over time, the "random" (lucky) part of my performance will get worse at some point — my performance declines overall. I regress downwards to the mean (= the real trend of my performance ability).
I perform extremely poorly which is partly based on bad luck. As bad luck isn't stable and tends to even out over time, the "random" (unlucky) part of my performance will get better at some point — my performance improves overall. I regress upwards to the mean (= the real trend of my performance ability).
While the science of performance and real-world outcomes is complex and context-dependent, random chance always plays some role in basically everything we do. Regardless of our skills at playing basketball or picking effective business strategies, randomness will influence the outcome to some degree.
The principle of Regression To The Mean applies whenever we encounter strong deviations from averages (or means) or look at extreme groups. Whenever we observe an outcome that is way outside the normal range that we would expect, it’s likely that following outcomes won’t be as strong in that direction anymore.
Understanding Mean Regression can help us contextualize and deal with outcomes better as well as predict more accurately what might happen down the line — without falling prey to erroneous extrapolation.
Resources
A few further resources you might like if you find the above idea interesting:
📚 Jonathan Evans’ VSI: Thinking & Reasoning
📚 Nassim Taleb’s Fooled by Randomness
📚 Daniel Kahnemann’s Thinking Fast and Slow
🎙 Podcast: You Are Not So Smart
🎥 YouTube: How We’re Fooled by Statistics
🎥 YouTube (channel): Crash Course: Statistics
📝 MindVault: Causal Reductionism
📝 MindVault: Chaos Theory