Most retail investors lose money to feelings, not facts. They buy when things feel safe and sell when things feel scary. The institutions on the other side are using probability.
Dipsern turns 50+ years of historical drawdowns into honest, probabilistic forecasts. One question, asked the same way every morning: at this drawdown level, what has historically happened next?
Started in a Jupyter notebook in 2023. Shipped when it became clear other people wanted the same answer.

Dipsern is a one-person project. Code, infrastructure, math, and this paragraph were all written by José Andrés Ruiz Elizondo. A civil engineer by training, algorithmic trader by practice.
Before Dipsern, José was a partner at Digital Hedge Capital, where he built the Orchid System, an algorithmic trading engine that processed over $45M in volume by filtering volatility signals into measurable rules. He also founded Muralia.mx and structured institutional cryptocurrency products.
“I don’t just theorize risk. I engineer reliability.”
The thesis is simple: capital markets reward structural integrity over speculation. Drawdowns are not noise, they are data. Dipsern measures every drawdown an asset has ever taken and tells you, in plain numbers, what happened next.
I read every email. Reach me at ruizjoseandres1@gmail.com.
Every signal is a distribution, not a verdict. You see win rate, prediction error, and sample size. Never a single magic number with a fake confidence boost.
Every signal is a distribution. Win rate, prediction error, sample size. Never a single magic number.
The full algorithm lives in /methodology. About 600 lines of pure NumPy you can audit on a Saturday.
Dipsern does not execute trades or take custody. You stay in control. We have no incentive to make you trade more.
No FOMO timers, no fake testimonials, no "expert says" takes. The only argument is what historically happened.
What Dipsern will never do
For informational purposes only. Not financial advice. Past performance does not guarantee future results.