Clinical trials tell you what caffeine does to the average person. You aren't average.

A 2022 meta-analysis of cognitive performance found day-to-day variation within individuals often exceeds differences between people. Population data hides whether a supplement works for you specifically.

The Problem With Self-Report

Most supplement trackers ask: "How do you feel today?" This fails for three reasons.

Placebo is real. Expecting improvement creates perceived improvement. Double-blind trials exist for a reason.

Memory is unreliable. You remember the good days. Regression to the mean looks like "it stopped working."

Subjective scales shift. A "7/10 focus" today isn't the same as last month's 7.

What Literature Priors Give You

Thalen starts with evidence-based expectations. Caffeine primarily affects attention. Bacopa targets memory. L-theanine improves executive function.

These priors come from systematic review of RCTs mapped to our specific test battery:

Attention: PVT lapses, sustained vigilance
Memory: N-Back d-prime, working memory accuracy
Speed: SRT reaction time, DSST throughput
Executive: Stroop interference, inhibitory control

With limited data, you see what clinical trials predict. With more sessions, you see what actually happens to you.

The Bayesian Update

Each session updates the estimate. The formula:

Final estimate = (prior weight x literature estimate + sessions x your measured effect) / (prior weight + sessions)

Prior weights depend on evidence quality. Caffeine's effect on PVT has strong support (prior weight: 20). Lion's mane has weaker data (prior weight: 5). Your measurements override weak priors faster.

What This Looks Like

Day 7 with caffeine: System shows mostly literature distribution. 50% attention, 30% speed, 15% executive, 5% memory.

Day 30 with caffeine: Your actual PVT and Stroop data start shifting the estimate. Maybe your reaction time improves more than typical.

Day 60 with caffeine: Your personal response dominates. The literature prior is just a starting point now.

Test It Yourself

Log a new supplement in your stack
Complete sessions as normal
Check the Supplements tab after 10+ days
Compare your domain breakdown to the population prior

If your measured response diverges from literature, that's signal. The system adapts to you.

Bottom Line

Population averages are starting points. Your brain chemistry is yours. Thalen combines what science knows with what you measure.