Research library
What the evidence actually supports
A decade of investment research, published in full — including the parts that did not work. Everything here is general information, identical for every reader, and written before we know who is reading it.
What we tested and rejected
Twelve candidate strategies that did not survive their own pre-registered tests, with the numbers that killed them. Almost nobody publishes this, which is exactly why it is here first.
How we test things
Register the hypothesis before touching the data. Compute achievable power before setting the bar. One attempt, no rescues. Adversarial review by someone who did not do the work.
The risk ladder
Why the benefit of taking more risk flattens well before the top of the range, and why low volatility is not the same thing as safety once inflation is counted.
Why publish the failures
Most investment research you can read is the survivor of an unreported search. A dozen variants were tried, one looked good, and only that one was written up. The result is a literature that appears far more successful than the underlying process was.
The only real defence is to fix the test in advance and then publish the outcome whichever way it goes. That is what the record here is: not a highlight reel, but the whole ledger, kept because a research programme that reports only its successes is not reporting anything.