Research
We publish what we measure, including the results that do not flatter us. The preprints below are negative results: the validation signal our own pipeline selects on does not predict forward return, neither does the obvious replacement for it, and nor does holding each pick for six months instead of one week. The portfolio chronicles describe what each paper-trading portfolio actually did, trade by trade, and why. These are research documents, not advice, and none of them is a reason to trade.
Research portal
What this portal holds
Two kinds of document, both descriptive and neither an argument for trading. The preprints test how strategies are selected on data their parameters never saw. The portfolio chronicles follow each paper-trading portfolio week by week: what it bought and sold, why each rule fired, why each trade ended, what price did afterward, what the system was doing at the time, and how the result compares with simply buying and holding.