Precisely, well said. It's not a matter of letting profits run and cutting losses short (although if you can consistently do so, you'll be profitable in the long run), but moving with the herd. The best money will be made by learning to gauge the herd, one must stay with it during strong movements and get a "feel" for exhaustions during both chop and trend. In fact, if one is able to understand market/herd psychology as it applies to stock/futures movements, there is no need to distinguish between chop/trend markets, simply to move (and anticipate the moves) of the herd to the best of our abilities. Getting out quickly when we're wrong, and capitilizing when we anticipate correctly. I too am here in the learning curve, very tough step.
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
Hi Johnny, OK about roulette. I didn't dwell on this. Tripacks example is correct, but I don't think he meant this to be immediately applicable to trading. Now about the other points you make, all I can say that in trading you can't rule out anything. However to use this as an argument to support any claim is far from valid either. I definitely don't agree with you anymore when you use "expectation" which supposses a rather stringent level of rigor. Also your "this is just science" and later "unscientifically" are out of place in this discussion. We have not even a sparkle of anything here that could qualify as a "science". As I indicated, I would see the possibility of concluding something "scientifically valid", provided the underlying stochastic process were known and of such a nature that it could be mathematically handled. This is very far removed from the subject we are dealing with. Referring to "edges" is also very loose - entry and exit policies would have to be completely specified. So what remains. Not very much. Only simulation/backtesting might tell us something that could be applied to make money. If you look back at the ET threads on random entry, I can't recall of a single post of somebody presenting a veryfiable example of an even primitive backtest showing profitability. Good luck with your tests. nononsense
Other thoughts below: It appears using TA (or else) can achieve at least two advantages that could not be done by merely flipping a coin for entry. - Waiting for the proper timing for an early entry of a trend. - Scanning some high probability trades among many different markets/stocks.
If the market were deterministic, then, I think flipping a coin to enter would be a valid strategy. But, since I think the market is chaotic, it is better to focus on determining how that chaotic structure works. In that context, coin-flipping is a useful tool for describing market activity i.e. an up day vs. a down day.
No, huh uh. Here is an example of deterministic chaos: http://monet.physik.unibas.ch/~elmer/pendulum/chaos.htm The market behaves in much the same manner, swinging between up days/down days.
Probably the system might be tradable and profitable, but its potential drawdown could first (50%) reach a level beyond expectation just before getting to a desired profitable level. Accordingly, TA would be still a better approach in order to maximise ultilisation of capital and manage risk.
I've discussed this before. That was before I learned it had been discussed long before: Nobody Asked Me, But...Let us recall that the entire subject of statistical properties of stock prices, which led to random walks and efficient markets, started out with the studies of streaks (they're called runs and reversals) by Jones, Cowles and Davis, and that after the always amateurish and misguided work of a certain University of Chicago professor given to stereotypical results and outdated data, the subject has been considered often by others and written about with respect to individual stocks and markets, sometimes with great style on the site that the marqueeist Jubak writes upon, and that the methodologies of streak-based trading are followed by billions and billions of dollars overseen by quasi-scientific second-handers. A preliminary to considering streaks is to consider expectations after runs and breaks of run of varying lengths. And the work of Davis in the Analysis of Economic Time Series was a good start in that direction. As for streaks in racket sports...well, that's another story. --Vic from http://www.dailyspeculations.com/index.htm My contribution is here: http://www.futuresmag.com/industry/downloads/downloads.html Look at the Spreadsheet for April 2002. Hope you are not bored to death, lol.