What Émile Is Not
- Émile does not predict price. He estimates the probability that a token which already reached $10K will reach $30K.
- Émile does not have an edge that guarantees returns. Four features cannot forecast a market.
- The jar is not decorative. If the model is bad, the jar stays empty and the site says so.
1. Borel's Infinite Monkeys
Émile is named after Émile Borel (1871–1956), the French mathematician who formulated the infinite monkey theorem: a monkey hitting keys at random for long enough will, almost surely, type the works of Shakespeare. "Almost surely" is Borel's technical term — an event of probability 1.
2. Why it fits Robinhood Chain
Robinhood Chain is the room full of typewriters. Thousands of people throwing random token names at the DEX pools every day, and once in a while something sticks. Émile is the one monkey who decided to sit down and write the results in a notebook.
3. What Émile Actually Does
Every token launched on Robinhood Chain DEX pools that reaches $10,000 peak market cap enters the study population. Émile observes whether it goes on to reach $30,000 peak market cap, using four feature families (launch hour sin/cos, day of week, 48h holder count, and lore text embeddings).
4. Why the Jar Fills Slowly
A model can look good by luck on a small sample. Vapnik–Chervonenkis theory and bootstrap resamples compute a proven performance floor (proven_floor = min(floor_vc, floor_boot)). The jar level is driven by this floor, not raw score. We explicitly admit that capacity parameter d (28) is a chosen capacity parameter, not a derived VC dimension.
6. Check His Work (Open Data)
Anyone can verify the model, reproduce the AUC score, and check whether the jar is honest.