After going thru this crap for years, I now disbelieve almost everything I see, and build my own methods, testing them till the cows come home. My batting average for the last 20 trades is 95% wins. Long term average (years) is just below 85%. There ain't no free lunch, just a lot of complex work. [/QUOTE] Sounds too good to be true. Just kidding. Are you talking about scalping or minute trading? And what do you trade? An even more important question seems to me measuring targets - apart from scalping obviously - if there is a way besides getting stopped out via trailing stops. could not figure what was yours and what you were quoting... But winning % is the most useless metric in trading. Newbies focus on winning%, serious traders usually realize that winning% is only a function of money management, usually by holding onto losers and banking small wins. I cannot imagine a seriously profitable, lognterm trader on ET who crows about their winning%. Profit factor, Sharpe/Sortino, max DDs, etc., yes...
Scalping? Almost. I trade only high volatility stuff -- 2x, 3x ETFs and "risky" stocks. Almost invariably it would be classed as "high beta". As far as winning % being a useless metric, you might be interested in studying probability and mathematical expectation.
erm.... Renaissance Technologies ..... Anthony Bolton's Fidelity Special Situations Fund ......? http://www.guardian.co.uk/business/2008/jun/13/investing.banking
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ We are still waiting for your 3rd posting giving some hints how to get started. Thanks
I have been questioning if I should post this but here it is. I dont think that its important to determine what indicators big banks are using to determine there trades, after all you also have to have the timing as well and know when they will act and when they wont. What important I think is to use whats in front of you and available to develop a strategy that identifies that they are acting now and jump in with them when the move is starting. Its also important to note that they have many programmers/experts working on different things that aren't available to the public whether it be neural networks patterns, indicators what ever it may be. One can agrue that Elliot wave is up to the person using it to determine all of the levels, if you apply that theory and try to copy the major moves with whats available you will get very close. I guess what Im trying to say is, ID major market moves and find out everything that surrounds them, then apply your findings forward.
According to the Turtle Trading, they did use TA to trade no matter what TZ says. So it did work in the past to trade large size, not sure what they use now. Also, the main thing was that they followed their rules which is the hard thing when trades don't go your way.
I think maybe a 200 week simple moving average is something that some big time advisers use, not sure if people with big money actually use it at all, that's different than advisers...
I would like to contribute a few of my ideas for the group's consideration. Maybe I can shed some light on a few of the problem areas that arise in investing. By way of introduction, I am an old guy, retired from 40 years of computer programming, now living in the mountains of Virginia. I'm a math nut, even to the extent of taking a part time job a few years ago as a math tutor to youngsters seeking their GED. Math is a common problem. During that period, I began to see that the most important aspect of learning any type of math was the gaining of INSIGHT into what was happening as a result of the use of the mathematical processes. And insight into the math is critical for the trader as well. That is why I pick apart the hallowed indicators. Most of them are bunk. Most will not stand an intensive search for logical substance. Look at a simple indicator, the moving average. Very common. Traders cheer about the 50 day MA crossing above the 200 day line, and the like. Now lets ask some questions about these averages. The SMA (simple moving average) is easy to do -- just add up the entries and divide by the number of entries you have. But when we add the entries, we are IMPLICITLY assuming that each of the entries has the SAME IMPACT on the result. When we find that yesterday's price was 6.50 and the price 200 market days ago was 6.50, we simply add those numbers into the array. But is something that happened 200 days ago just as important as something that happened yesterday? I don't think so. The averages need to be weighted, so that yesterday's price has more impact than last month's or last year's or the closing price on the third Friday in November 1923. How do we weight the entries? Here, the judgement factor enters the picture, and our opinions may differ. I have tried a variety of methods,, but the one that seems most satisfying LOGICALLY uses the sine of an angle that varies from near zero (for the oldest data) to 90 (for the most recent data). Thus, the weights vary from near zero to 1. The higher weights are bunched near the recent end, and the older data shows decreased weight. The divisor in this case is not the count of the entries, but is the sum of the weights applied. Now, suppose we decide that this is the perfect solution to the question of weighting. Are we ready to USE it? Not a chance -- because most of all, we must doubt ourselves. That means that we must test the new average in a large variety of circumstances, bull markets, bear, and sideways, trying to find logical inconsistancies. When you find that extensive testing shows a very high percentage of correct decisions, you're getting close. As I said earlier, study mathematical expectation if you haven't already.
Consider abscribing more humanly motivated waypoints for price moves such as a consistent bottom at a given time interval too coincidental to be left to random taming calcs. MOS seems to have made intermediate bottoms at or near 10 in the morning ET with a modest consistency, that might be skimmable. How about lunch?