I opened a demo account with www.virtsupport.com, which seems to be a very good software package. I put in a couple of trades and wiped the $3000 account within a month. I am so relieved that it was not really my money traded. A word of advice anybody please?

What was the spread? In paper currency trading, your average loss per trade should be the spread. So if there was a $1 spread, and you did 3,000 trades, then you broke even.If, however, you consistently lose more than the spread per trade, then you are on to something. In this case, you know that somehow the stimuli you are exposed to are predictors of future currency prices, only that you taking them 180 degrees out of phase. If you are losing more than the spread per trade, you need to 1) reverse your buys and sells, and 2) zero in on the subset of stimuli, and attempt to classify a subset of trades where you are making more money than the spread.If, however, you lost $3000 in just a few trades, then that is not enough trades to discover an average loss per trade. Your average profit per trade will be a reflection of short-run randomness, rather than long-run expectation. It is possible that you are executing a great strategy, which would yield a profit given a long enough time and enough trades for the averages to catch up.In this case, the advice has two parts; 1) recognize that every trade will have a large random component, even if you expect it to be profitable, so you need a big cushion to carry you through the bad-luck stretches and 2) understand that the market is a moving target, so that only the simplest patterns are likely to last in this long run. In other words, patterns more complicated than "buy the breakout" are likely to stop working before the averages ever catch up with you. Patterns more complicated than trends are not likely to keep occurring long enough for your average profit per trade to reveal your expected profit per trade.The best patterns are zero-frequency patterns. An example of this might be if you saw a swimming pool filled with boiling water. You have never seen such a swimming pool before. But you have experience with two simpler phenomena, 1) stepping into a swimming pool, and 2) touching boiling water. You know that if you step into a pool, you get water all over you. And you know that if you touch boiling water, it hurts. So you can guess that if you step into a pool of boiling water, it will hurt all over your body.A market example of a zero-frequency pattern might be selling your stock before the baby-boomers retire. I think such low-frequency trading is often referred to as "event-driven" trading. You will see that such trading often involves converting and utilizing news inputs or linguistic symbols to describe physical events, rather than using patterns comprised of multiple time serieses. But you don't need to be a fundamental trader to do it, you can lower your frequency with charts if you want.Another good type of specific pattern, with a frequency of one... I'll save that for another day.

I put together something for you to play with. According to this theory of recurring phenomena and trader learning as I laid it out, you could jump to a couple of hypotheses:1) Simple, high-frequency patterns - the things that happen every minute like higher highs - shouldn't work, because they will be so obvious that everyone quickly figures them out.2) Complicated, high-frequency patterns - such as an intraday consolidation coincident with a runup and selloff in other instruments - shouldn't work, because these oddities won't keep happening or last very long.3) Simple, low-frequency patterns - things that only happen once a year - should work because they are simple enough to keep happening, but because it will also take a while for people to notice them.So I put together a crude test to see if simple patterns, such as breakouts, 1) recurred, and 2) recurred for longer the lower their frequency, meaning the longer it took for people to get used to them and learn to trade them. I applied a strategy of buying a new high close, and selling a new low close, to a sloppy gold chart going back to the 1970's. I tried buying or selling the close when it was higher or lower than 1) yesterday's close, 2) the highest or lowest close of the last 2 days, 3) the highest or lowest close of the last 4 days, 4) the highest or lowest close of the last 8 days, and 5) the highest or lowest close of the last 16 days. Here is a rough plot of the profits to the same strategy at different frequencies, over the last 30 years:What we are hoping to see is a starting point, or some time period where the strategy works at all five frequencies. Of course this period could be to the left of our chart, so that the experiment begins after they all stop working. But in general, the highest-frequency strategy could only make money for the shortest periods, whereas the lowest-frequency strategy could make money for longer periods, since people have the fewest examples, and the fewest recent examples, of it actually making money.What we do see, is that I got lucky. The lowest-frequency 16-day strategy kept making money, or at least not losing money, the longest. And the highest frequency 1-day strategy, lost the most money. The only strategy really out of order was the 2-day strategy, which made more money than the 4-day strategy or the 8-day strategy.You could say, well, this order of profitability is similar to the one predicted, but it also could have been produced at random, and your methods were crude. But something to consider, is that the highest-frequency 1-day strategy should actually be expected to make more on paper, since it will have to make more money on paper before someone can make money in real life after commissions and slippage. In other words, a strategy which requires me to buy or sell a 100 million dollars worth of some commodity every day, will have to make more money on paper than one that lets me spread out my orders, to make money in practice.So if a low-frequency strategy works on paper, people will trade it until it stops working. If a high-frequency strategy works on paper, people might not be able to trade it since the commissions are too high. So the high-frequency strategy which nobody can actually trade - since no one can guarantee himself the closing price every day - should continue to work on paper. For example, consider that S&P futures trade back and forth at a .25-point increment, very predictable. On paper, you could effectively buy at the bid and sell at the offer, by buying at close of every minute that was a downtick, and selling at the close of every minute that was an uptick. You would make a zillion dollars a day, on paper. But if you were limited to buying or selling only once every day and holding all day, these imaginary profits would disappear.So my chart above, in theory, should show more imaginary profits for the high-frequency strategy which we couldn't actually trade, than for the more realistic buy-and-hold strategy. Paper profits should equal commissions, and should therefore be larger at higher frequencies. But instead, it shows profits consistent with my "theory of trader learning" completely canceling out and reversing this effect. To test whether this result is just sloppy random nonsense, you could run the same experiment on copper, deutschmarks, yen, and other instruments than just gold.

QuoteOriginally posted by: VickygI opened a demo account with www.virtsupport.com, which seems to be a very good software package. I put in a couple of trades and wiped the $3000 account within a month. I am so relieved that it was not really my money traded. A word of advice anybody please?Hi Vickyg-Together with technical analysis, I would recommend you immerse yourself deeply in the fundamentals as well. Bring yourself to a place where you can discuss every word of Alan Greenspan's testimony intelligently. Pull up a speech of his from online and study it profoundly. He does not control markets by any means, but he can move them.This is my recommendation. The central banks are important players in the currency markets.Also, familiarize yourself with the concept of "managed float".Matthew

If you are serious about trading FX prop, I'd recommend getting familiar with the microstructure models about which there are several good books (I have the one by Rich Lyons).From what I understand about the volume of institutional speculation and central bank intervention in the FX markets, they seem to be one of the most efficient markets in the world, especially in major currency pairs, so I'm not convinced that technical or fundamental analysis would help you much. Understanding Fed speak will help you understand what just happened, but the market will almost certainly beat you to any profit opportunities available from those words.On another side, the low trading costs and elegance of triangulation-implied correlations may give you good opportunities in trading FX vol.Good luck...

Vickyg, aside from all the technical aspects of trading, market microstructure, technical analysis, my personal philosophy on trading is the ff: (1) Always have a reason for entering into a trade/position. (2) always stick to your s/l or t/p levels. (3) if upon entering the trade, markets move against you and the premise of you entering the trade is broke, cut your position.You may be surprised to see that you may always have a right view on the market, but if you don't stick to your s/l and t/p levels you'd lose no matter how good your view is.