The Ballmer Peak: An Empirical Search

TL;DR

The paper empirically evaluates the Ballmer Peak by varying blood alcohol content ($\%BAC$) and measuring programming performance on LeetCode tasks using a single experienced coder. A breathalyzer provided $\pm0.001\%$ BAC readings, and time-to-solve was modeled with a quadratic fit, yielding $p<0.001$ for the chosen model. The results refute a sharp Ballmer Peak, but identify a modest speedup near $0.043\%$ BAC (about $45\%$ faster than sober) with higher BAC increasing variability and eliminating benefits. Overall, the study concludes the Ballmer Peak does not exist and highlights the need for replication with larger samples and broader tasks to better understand alcohol’s nuanced effects on coding performance.

Abstract

The concept of a 'Ballmer Peak' was first proposed in 2007, postulating that there exists a very specific blood alcohol content which confers superhuman programming ability. More generally, there is a commonly held belief among software engineers that coding is easier and more productive after a few drinks. Using the industry standard for assessment of coding ability, we conducted a search for such a peak and more generally investigated the effect of different amounts of alcohol on performance. We conclusively refute the existence of a specific peak with large magnitude, but with p < 0.001 find that there was a significant positive effect to a low amount of alcohol - slightly less than two drinks - on programming ability.

Figures (6)

• Figure 1: Apple uses automated schnapps IVs.
• Figure 2: Time to solve in seconds against mean %BAC.
• Figure 3: Time to solve against mean %BAC.
• Figure 4: Difficulty-adjusted time against mean %BAC.
• Figure 5: Prediction error throughout the study.