Survivors and false patterns
The data you can see is often the data that survived. Patterns found in it can be an artefact of what went missing. These stops show how the trap works, and the habits that keep you out of it.
-
Pattern Essays
Survivorship Bias
We study what survived and quietly forget everything that did not — and the survivors look smarter than they were.
-
The Toolkit
Survivorship & Look-Ahead Bias
The same trap in a backtest: today's index members, or knowledge you would not have had at the time, make a strategy look better than it could have been.
-
Market Psychology
Hindsight Bias
After the fact, the outcome looks obvious. That feeling is the memory of what happened, not evidence you could have known.
-
The Toolkit
Cross-Validation Done Right
Honest evaluation starts with honest splits — no answers leaking into the data the model learns from.
-
Timeseries Engineering
Backtesting Forecasts
For anything ordered in time, test the way you would have used it: forward, one step at a time.
-
The Toolkit
Learning Curves & Overfitting
The gap between training and validation performance is overfitting made visible.
-
Pattern Essays
The Map and the Territory
Every model is a simplification. The danger is forgetting what was left out — including everything that did not survive into the data.
Ask what is missing from the data, and what you know now that you would not have known then. Most false patterns live in one of those two gaps.
Add up enough small, independent nudges and the same curve appears — in sample averages, in price bands, in risk models. Follow it across four fields, then see where it quietly stops being true.
Extreme results tend to be followed by less extreme ones. It looks like a force pulling things back; it is arithmetic. Watch the same idea turn into trading signals, portfolio rules and model penalties.
Every field has the same problem: real structure mixed with randomness. These stops are the tools for pulling them apart — and for checking you have not invented a pattern that was never there.
When a system's output feeds back into its input, small nudges can snowball. Crowds, prices, volatility and even deployed models run on loops like this — and they tend to stay quiet until they don't.
Recent direction tends to carry on, for a while. The same idea speeds up neural-network training, drives a family of trading indicators, shapes how traders add to winners and fuels the crowd. Follow it, and see where it stops paying.
Many systems change all at once when a line is crossed. A neuron fires, a classifier says yes, a language model drops its unlikely words, a price breaks out, a stop-loss sells. Each stop is a threshold, and each has to answer the same question: where to draw the line.