It is less
important to be right than to be protected when being wrong.
It is better to diversify a portfolio in several strategies based on different logics than in a specific number of stocks and sectors. The home page gives examples of strategies with very different logics.
Robust Strategy Design
We elaborate strategies starting from fundamental and technical analysis research and avoid super-optimization.
We use professional data
and an outsourced simulation server certified by Investorside
Association. The mostly used period of
simulation is from January 1999 to December
2013. These 15 years cover all market conditions: two crashes, good years and flat years. We select strategies with
a good behavior not only on the whole period, but also every single year.
The Luck Factor
Luck does exist. With the same probability to win (p) and the same average win/average loss ratio (w/l), your account may experience very different possible futures depending on the sequence of gains and losses. Example with w/l=1,4 and p=50%:
That is why we select only strategies that have a clearly oriented and focused beam of possible paths, like this one (see the technical summary) :
Going further in risk
evaluation, we use probabilistic indicators taking into account the
data sample size:
p95: probability with a 5% error (real probability has a 95% chance to be higher).
K95: Kelly criterion for p95.
Every model is an approximate representation of a real phenomenon. It is impossible to reduce the risk to zero or to predict the future, but it is possible to detect when a strategy goes out of its normal limit.
Even a very good strategy may have a 30% drawdown and stay in negative territory during months. Easy to imagine, harder to live. It is better to invest less money than giving up under pressure with a good strategy. Without a long term vision, the short term makes no sense.
2012. All rights reserved.