A developer frustrated with Garmin's focus on physical metrics built a custom model to predict his daily chess performance. Using data from Lichess and his watch, he trained a logistic regression algorithm on sleep and exercise signals. The project aimed to bridge the gap between physical health data and actual mental sharpness, a common complaint among biohackers.

Contrary to common assumptions, the model revealed surprising correlations. REM sleep was the strongest positive predictor, while recent exercise showed a negative impact on his chess ELO. Even more unexpected, Garmin's stress metric—often viewed as purely negative—correlated with better performance, aligning with the Yerkes-Dodson law that optimal arousal boosts cognitive output.

These findings directly challenged the author's intuition and Garmin's own health advice. He learned that feeling physically rested doesn't guarantee mental acuity. The data showed that short-term cognitive demands and long-term physical health are often misaligned, explaining why generic wellness scores often fail to predict high-stakes mental performance.

Motivated by these insights, he developed a personal Garmin watch app. This tool calculates a daily probability of his ELO increasing, often contradicting the device's standard morning review. He has since adjusted his lifestyle, prioritizing alcohol reduction to boost REM sleep. Garmin's restrictive API unfortunately prevents a public release of this specific tool.