Quote from nononsense:
Hi JohnnyK,
This may be true, or it may not be true.
In the case of Tripack's roulette example, it won't work as in this process you have no exit option.
If you would bring in stochastic process control theory, something could be said about entry/exit policy merits. However under this theory you would have to:
(1) rigorously define the underlying stochastic process (futile in a market situation);
(2) rigorously define your control laws governing entry and exit (seems easier than it is).
Not pretending to say anything rigorously, it seems sensible to not choose a blind or random entry and from then on apply a policy for exit, but, to make use of all your "knowledge" about the process to determine both entry and exit.
Not gotten very far with mathematical "market" description, I for my part will keep on fiddling with both entry and exit. I'm not smart enough to do with exit only!
Yours and Tripack's attitude on money management is reasonable. You must be a fool to believe that money manegement - often poorly defined and understood - is going to save you from going down in flames with "random entries" (or other crazy techniques).
When I look at these things, I don't babble too much about money management but rather try to asses "risk of ruin" for my strategies.
Be good,
nononsense
Hi nononsense,
My comments stem from some experience running TA (and money management) on random simulated market environments. I did not comment on roullete. I did not include a single roulette bet as part of a debate on entry vs. exit. There has been a comparison of apples and oranges here.
Let's revisit this. If we can agree that there are three parts to any trade - entry, hold, exit - and if this trade/system has a zip edge, the edge can come from any or all parts of it.
There is some unecessary debate as to which part is most important. Give me a 1% edge and I don't care where it comes from! Money management can then grow that 1% into kajillions in short order with safe drawdowns.
It does beg the question however, that if someone has an exit edge and knows it, why not turn it around and make it an entry edge too?! Probably because people are unsure about exactly where thier zip edge is coming from...that is - which of the three components - is delivering their zip edge.
Practically speaking, I would be happy to make money first, then figure out where the zip edge is coming from later. Just find what works first, no matter how bizzare sounding.
dnaJ65000 is running a process of elimination. Rather scientific actually. We know the entry is random. We are trying to determine if the exit has a zip edge...in the Forex market.
Other factors could be added later, such as a trail stop system. This would factor into the *holding* component of the three components. If a trail stop system turned a zero zip edge whole trade into a trade with a zip edge, then we would know where the zip edge is coming from. This is just science.
The expectancy is for zero zip edge on this exit policy. If nothing else, it can help dispell myths, and save us some time by proving what does not work. But the Forex may offer a surprise, who knows?
In this exit policy the profits are not allowed to *run* so to speak. Some people swear by the *let profits run* thing. So let's see. Anticipating a zero zip edge on this test, a couple of us have forwarded *run* test suggestions, for dnaJ65000's next test.
Bobcathy1 suggests short skirting yesterday's close for an entry possibility.
I have suggested a time exit where the other side of the skirt (bracket) becomes a short stop.
This way, price can *run* as far as it wants, or a loss is cut short by the short stop.
Note there is no trail stop policy to muck up the scientificity of this test. (Test one unknown at a time). As I said, my tests show it adds nothing, meaning it deserves a lower priority in order of indroduction to a test.
If the bracket entry/ time exit test has a zip edge, we will not know if it is the entry or the exit because two unknowns were introduced unscientifically at the same time. But it might save time by eliminating two possibilities at the same time.
Roullette? That's another article altogether.
JohnnyK
, it still takes a lot of skill to distinguish a trending from a non-trending market and that's where my learning curve is at now. If anyone knows how please let me know.
