forex trading for a living

DallasSteve said:
........
"you seem to be suggesting that it's impossible to make trading pay" - Not exactly. Even if the market is a random walk (or close to it), if you have enough participants the law of averages says that some of them will make some significant profits........

Steve
If trading(as opposed to entry technique) were truly a "random walk" would not the law of averages over time ensure than everyone would come out even or close to it?

The law of probability on the other hand would say that some would indeed make significant profits simply by benefiting from the unlikeliest run of good luck. However this "few" would be so infintesimally small as to be insignificant for any reality based discussion.

Now, as the reality is clearly different to these two outcomes the evidence would suggest that trading(again, as opposed to entry technique, which is only one part of the trading process) is not a "random walk" regardless of what a backtesting programme states.

DallasSteve said:
I'm skeptical that any human can trade better than a computer, but since it can't be proved I won't try to debate that point.
What's to debate? For whatever reason, the reality clearly demonstrates that there are many humans that can trade better than a computer. As Charlton points out, the debate is about why this is. It could be experience, judgement, "the myth of irrationality" or esp etc.

Cheers,
PKFFW
 
Each day, I narrow 10,000 U.S. Stocks down to a select few that I choose to trade. The process of narrowing down 10,000 stocks is 80% mechanical and 20% human judgement,

You narrow down 10,000 stocks....every day? Jeez - Thats a busy day ahead of you! I would have thought it would be better to stick to fewer stocks or particular sectors to find your edge?
 
Hi Ewan

I narrow down 10,000 stocks to about 20 using a screening process run on computer - I don't do it manually! The remaining 20 I then manually screen.


To DallasSteve,

You mentioned that the position size you use for testing is a fixed dollar amount. I think this could be a possible cause of your problem.

To bring up Van Tharp's test in more detail, he ran his breakout system starting with $1m of trading capital and traded the DOW 30 Stocks.

For one of the position sizing strategies, he purchased 100 shares for each position traded. At the end of the test this strategy showed a profit of $32k. For another position sizing strategy he risked 1% of total equity on each position traded. At the end of the test, this strategy showed a profit of nearly $2m.

The point here is these results were both from the SAME TRADING SIGNALS. The only thing that changed was the position sizing strategy.

Maybe you have a few successful systems in your computer, but they aren't showing much profit because you aren't using a sound position sizing method.


Thanks

Damian
 
Position Sizes

Damian

I'm still not clear on what's different about Van Tharp's sizing of positions. First you say that "he purchased 100 shares for each position traded". I don't do that. If you purchase 100 shares for a stock trading at $80 it's a much larger position than a position of 100 shares for a stock trading at $8. I use a constant total dollar size for all positions tested and that sounds like the second example you gave for Van Tharp. Anyway, if Van Tharp has an interesting book please recommend it and I may look at it.

Steve
 
Hi Steve,

I was just trying to give you an example of how radically position sizing techniques can effect a trading system's results - I wasn't trying to give you an exact formula to put straight into your computer.

So if you're using a fixed dollar amount per position, can you tell me two things: Firstly, when you run your test, at the beginning of the test what amount of trading capital are you starting with? Secondly, what is the fixed dollar amount you are risking per position?

I'm not asking this to challenge you - I'm just genuinely interested in these backtested systems you're running and why none of them are showing a decent profit.

PS - the book I am referring to by Van Tharp is "Trade Your Way To Financial Freedom", but before you run out and buy it, be aware that it is a book about how to develop a trading system, rather than being a book that teaches any particular method.


Thanks

Damian
 
Damian

I don't backtest in dollar amounts. I backtest in percent returns. Then I can extrapolate that to whatever size portfolio I want and calculate the appropriate impact of commissions and spread. I see no reason to work with a specific size portfolio during the testing stage.

Steve
 
Hi Steve,

Sorry - I don't follow you.

Can you give me an example, because it doesn't make sense to me that a test would be started without knowing your starting balance and your risk amount per trade.

What exactly do you mean by: "I backtest in percent returns." ??


Thanks

Damian
 
Charlton said:
Well a newbie's simple question has prompted a lot of sound advice from everyone and I hope it has not discouraged Coolshades, but at least made him step back and consider the advice given.

The discussion about systems, backtesting and human intervention is interesting and has appeared from time to time on this board. I would tend to agree with you Damian that, at the moment , the mechanical aspect of trading is not 100% of the activity, but that human intervention occurs. Sometimes this is detrimental to the process e.g. when emotions cause us to turn away from our prescribed strategy. On the other hand there is a certain something that is currently not captured in mechanical systems, which advanced traders might call intuition. The question that has often been asked is whether this is really just experience and advanced skills that allow the proficient trader to see the signs that other's miss and which have not yet been programmed into a system OR whether this is something else in play, which cannot be deconstructed into logical activity and programmed systems.
I can understand that, at the lower end of position sizing, transaction costs will comprise a larger % of a potential profit or loss and therefore would have to be taken into account when backtesting a system - the scalping versus swing/position trading characteristics for example. Was this the point that you were making ?

I would also like to ask Steve, if he has compared the results of his backtesting with some measure of the level of discretion that is required with the system and also whether it indicates a higher or lower degree of profitability when discretion is applied

Charlton



nope, am not discouraged.. i have mentioned before.. i have stepped into this with very realistic expectations. i want to make this work and am determined to make it work but am also aware that i will lose money and am prepared for it. i have a very successful day job and the budget i have allocated to this is something i can afford to lose....

now to the bigger question of negativity in the threads.. i keep wondering, if people are not able to make money here, why do it?
 
Hi Coolshades,

As regards negativity on the threads, you have to bear in mind that more people are unsuccessful at trading than are successful. Therefore, it stands to reason that there will be more negative threads than positive threads.

What keeps everyone going is the relentless desire to solve the trading puzzle. When you know that there are other people out there who are successful at trading, it fuels your need to figure it all out, knowing that it IS possible and you're not just working towards an unrealistic goal.


Thanks

Damian
 
The Thread That Won't Die

damianoakley said:
Hi Steve,

Sorry - I don't follow you.

Can you give me an example, because it doesn't make sense to me that a test would be started without knowing your starting balance and your risk amount per trade.

What exactly do you mean by: "I backtest in percent returns." ??


Thanks

Damian

Damian

I'm probably not going to work this thread much more because I'm not learning anything from it. If you don't understand, I don't understand why you don't understand. We are dealing with paper trades. It's all hypothetical. You don't have to have a starting balance. You don't have to have an acccount. You don't even have to have a shilling. It all happens in cyber-fantasyland.

I'll try this one more time: When I run a test each trade is a long or short. It has an entry price and an exit price. You take the price sold minus the price bought and divide that difference by the entry price. Voila! Your percent return. That's all I care about. If that percent return on average is enough to beat commissions and spread I've got a viable system. If it's not, I try again.

Steve
 
coolshades said:
i keep wondering, if people are not able to make money here, why do it?

CoolShades

That is a question of human psychology. People gamble because they might get rich quick. Their greed overtakes their reason. Trading is driven by two emotions: greed and fear. This has been covered by many others before me and I can't do it justice. I've made money trading and I've lost money trading. It's one of the greatest thrills when you win and one of the most infuriating experiences when you lose.

Steve
 
Hi Steve,

Now I understand - thank you for clarifying.

Just looking at the percent return will NOT get you a winning system......and that is your problem.

You should be looking at what your return per trade is relative to your risk per trade.

Sort that out and you'll find your trading system.

The reason just relying on percent return won't work is because lower priced stocks will generate higher percentage gains and losses than higher priced stocks, so your risk per trade is not equalised. There is your problem.


Thanks

Damian
 
Damian

That's an interesting concept that you present and I see some validity in it, however, I also see some contrasting realities to consider. Generally lower priced stocks will have a higher beta (more volatility). And as you point out both their profits and losses will be greater as a result. Hence, won't they tend to offset and converge back to the same rate of return that higher priced stocks generate using the same system? Or put another way, when your risk is greater your reward is greater. No pain no gain.

Not all of my backtesting involves multiple stocks and not all of it involves stocks at all. Sometimes I backtest Forex and sometimes I only backtest one security/currency like EUR/USD or QQQQ. However, regardless of what I'm testing and regardless of how I test it I still find a strong tendency to generate a return of 0 or very close to zero tested honestly over the long run. Of course, it must be that untestable "human factor" that allows successful traders to make a consistent profit.

Steve
 
If you are only looking at percent return, you are really limiting your systems.

When Van Tharp ran his test, the percent return strategy made the least amount of money: $32k on a $1m account.

Your results seem to match his. Your position sizing technique is definitely wrong in my opinion. You would be better to use a "percent equity" method instead of "percent return", if you know what this means.


Damian
 
It means you equalise the risk you take on every single trade by risking a set percentage of your trading capital on each trade.

For example, taking stocks as an example. Let's say your trading system gives you a signal to buy stock XXX at $30 with a stop loss at $25. Your risk here is $5.

Let's say you have $10,000 in your trading account and you want to risk 1% of it on this trade. Basically, you need to calculate how many shares of stock XXX you'd need to buy so that if your stop is hit, the maximum you will lose is 1% of $10,000 - ie - $100.

Doing it this way means that no matter what market you are trading, or no matter where you place your stop, you can equalise your risk on every single trade you make.

That is a much more sophisticated position sizing strategy than just looking at your percent return.

The only way a percent return method will work is if you limit your loss on every trade to a set percentage, ie - 5% or 10%. For me, if I chose to trade as you do using percent returns, I would reject the above trade because it would mean risking 16% if my stop was hit.


Thanks

Damian
 
Hi Steve,

I have never backtested anything, so I can't talk about building systems based on historical data. I also can't grasp the idea of testing in order to profit from the same patterns in the future. Linda Bradford Raschke says she knows only two people who make money on a back tested purely mechanical system, and this woman has been in the business for 25 years. I also remember reading in a trading megazine an interview in which a guy said he nerver knows anybody who makes money trading mechanically over any meaningful strech of time. I can see why this might be the case, so I understand and agree with what you say.

But it doesn't follow, of course, that you cannot make money trading. You certainly can. I am a discretionary trader and I am profitable almost every month. It is hard to get rich trading if you spend the money on bills rather than compunding it, but you can make a decent living doing it.

About Van Tharp. I really dislike trading books. Van Tharp makes a religion out of developing a trading system and you can tell from the way he writesthat, either he is not a trader or is not a very good one. The point discussed here (position sizing) is one thing he seems to believe will make you successful. The fact is, if you haven't got a profitable method, no 'money management' system is going to make you profitable, which is what you seem to be saying. He also seems to believe he is telling us something very deep when he takes page after page explaining very basic arithmetic: if you risk x% of your bank and recalculate your position size after every trade, you will do well. Well, that is bloody obvious, i would have thought. What you are doing, of course, is compounding as you make profits and reverse-compounding as you make losses. That is a pretty safe way of trading, but that is all it is. The idea that it is the best way to manage your money or will make you profitable is absurd.

Anyway, my suspicion is that you are probably right about trading systems failing over the long term. The trick is to be a good discretionary trader, in my opinion. All the best.
 
Hi fxscalper,

I couldn't agree with you more about money management........no money management or position sizing technique will turn a non-profitable strategy into a profitable one. But - a good position sizing technique like "percent equity" will keep you in the game long enough for your system's edge to shine through and show a profit.

Whilst this technique may seem "bloody obvious" to you, it is not so obvious to many traders. Steve is a prime example - he had never considered it. Many beginner traders that I coach have never considered it either.

I agree with you as well about the human factor in trading. Backtesting assumes that a robot will be trading and this is simply not the case when you are trading every day. I do think that to remove human decision completely from trading is not a good idea.


Thanks

Damian
 
Top