Algo Designer
|
You have asked one of those fundamental questions that bothers many of us!
A technical analysis purist could recommend you to look at MACD, MAs, RSI, OBV, ATR. As people learn that the classical averages suffer from a time lag, they look into fancier formulae/indicators such as StochRSI, TEMA, DEMA, IE/2, adaptive MAs, triangular MAs. The list goes on. With a passage of time, many realise that static symbolic models, even when properly constructed and back-tested, barely make money if at all and "unexpectedly" break down as the market conditions change.
This is where the real fun begins. People try linear regression models and soon discover that even those seeming inefficiencies that looked good on paper are removed by the other market players. The next logical step is non-linear models that cost you lots of sleepless nights, and if you are lucky and persistent enough, let you see some interesting properties that can be exploited in the context of a prop trading model or as a bias adjustment function in various agency strategies.
Sorry for giving you such an open ended answer. There is no magic formula here. If you get your "momentum" function right, your trading model is 60% ready, the rest is the risk management which is... a bit easier.
|