Methodology
The Vordeon validation battery: nine phases, a frozen protocol and a verdict with a written reason.
with trades
1 · Out of sample
It must work on data not used to design it, also without its 5 best trades.
needs the rule in code
2 · Walk-forward
Re-optimize in each window and test on the next one.
needs the rule in code
3 · Plateau
The chosen parameter must not be an isolated peak among its neighbours.
with trades
4 · Monte Carlo
Luck in the mix of trades: 5,000 resamples with replacement.
with trades
5 · Stress
Double costs and the result without the two best years.
needs the rule in code
6 · Placebo
The same rule on invented prices: if it wins, the future is leaking in.
needs the rule in code
7 · Nulls
It must beat chance with the same exits, and beat “always long”.
with trades
8 · Funded account
How many evaluations it would pass, against the zero-edge barrier.
with trades
9 · Deflated Sharpe
Corrects the Sharpe ratio for the number of variants tested.
Current protocol
Current thresholds
| Loading… |