Does your strat give a "strength of signal"?

Discussion in 'Strategy Development' started by SomeYoungGuy, May 13, 2010.

  1. Meaning do you come up with some kind of a number, like "80% buy" or "4:1 short"?

    Or is it simply binary, like once conditions are met, it spits out "buy" or "sell" or "close"?
  2. Unless you are using a Cray for optimization, it has to be binary. Otherwise the optimization is outrageously complicated, and maybe even optimizing randomness. Scaling in as the probability of win increases as a winning trade evolves is a different matter. There you can use statistics.
  3. This is interesting topic. The p-Indicator developed by Michael Harris calculates both the rate of success but also the strenght of the signal. I have not looked into this in detail since I only recently got the software. I worked out some examples for another thread in this section using this software but not the p-Indicator. It sounds like a good concept.
  4. My Strat is pre CBS and there wasn't any gagets on it - just straight thru. I just plug it in and let it rip.

    If my ears are ringing after I play I know the strength of signal was good.
  5. As an off-topic aside, "PIS the value for short positions" is kind of a poor choice of acronyms.

    I prefer the Universal Return Investment Net Evaluator.
  6. Oh trading? Not music?

    Then yes, but it is visual. It's all about measuring energy and if it is working with or against momentum. Simple physics.

    So the strength of the signal is only part of the equation. Without understanding and quantifying momentum your stats will be skewed.

    It's like the effort required to push a car up a small slope, the effort to crest the hilltop and the effort to push it downhill. All the same energy but with tremendously different results.
  7. Let me clarify. Suppose you have developed a system using one or more parameters. It has a positive expectation. Then you discover that yet another parameter affects the expectation. You obtain several different levels of "confidence" based on parametric backtesting. What you have done is refine a startegy, not given it "confidence levels." So you pick the one with the highest expectation and always bet the farm on that version. It is the same with a stable of strategies. Why have a stable if one is always better? You always play your best strategy. Any others are just backups in case the best one craters.