Hi Guys
I would like some help from he more experienced traders. Is this enough to say that is a good system?
W/L is 50% and r.r is around 1.3
Thanks!
No. Without understanding the parameters of the system, you can’t evaluate it. This is a performance record. A performance record tells you nothing about how good the system is.
Suppose for example that I traded a system and its opposite in two different accounts respectively. For a while both systems break even so I quit trading them in those accounts and start trading them in new accounts. I keep switching to new accounts until I get a period where one loses a lot and the other wins a lot (an unusual period in which the opposing systems will respectively perform very well and very badly). Then I go around showing people the one that did very well. Because people cannot see the true performance of the system long term (that it is actually a break-even system) they may think it is a good system. Then when they try to run that same system they will lose money over time with the break-even results which are still negative because of costs. They were able to be deceived by the results I showed them because they thought results are the way to evaluate a system.
What was my system? I was betting on a coin flip. In one account it was heads I win what I bet, tails I lose what I bet. In the other it was tails I win what I bet, heads I lose what I bet. But there is a transaction fee of .01% every flip. If I had plainly told most people what the system parameters were from the beginning they would have ignored my results and said: “You are just going to lose .01% per flip on average.” And they would be right. A coin flip game can go through a series wherein the distribution of heads or tails is skewed heavily and as long as I show people results from a period like that, those that believe such results are an indication of a good system will be tricked.
To evaluate a system, we need to know what it is. What are the parameters of this system?
-Adrian
Not really.
It’s curious that over 700 trades, it won 54% of its short positions but 59% of its long positions. That’s a statistically significant difference, and I’d be wanting to see why.
Taking the 54% figure (which is what I would do) as a win-rate, with an average win of 45 units and an average loss of 50 units (as stated, with a tiny bit of rounding) gives a long-term profit factor of only 1.05.
It’s not [I]losing[/I] money, granted.
But … please excuse my saying that I wouldn’t be wanting to trade it myself, with a PF of only about 1.05. I might well be looking for ways to improve/refine it, rather than discarding it, though.
(I agree with Adrian’s post above, too.)
hey thanks for your answers
Lexys, how did you calculated the proft factor, and what would u say is a good number?
good return, and seems to be a solid system with good risk mangment.
is it manual or AE?
After doing some research, the profict factor is Gross profit/Gross losses. So profit factor of my system is 1.17
(0.54 x 45) / [(1.0 - 0.54) x 50] = 1.056
It’s just total profits divided by total losses (in this case 54% of trades win 45 pips and the other 46% lose 50 pips … I’m taking the lower of the two slightly different win-rate percentages, to be conservative, which minimises the risk that the first 700 trades might turn out to be a fraction better than the next few thousand.)
Anything above 1.0 makes overall profits rather than overall losses (given all the assumptions stated above).
I’ll trade anything I realistically can that comes out at 1.5+, myself, as long as it has a realistic win-rate to start with (I don’t want to trade methods with [U][I]very[/I][/U] low win-rates, even with a good PF, because the losing runs can be so frightening that with those you never quite know whether or not the system’s “stopped working”, and by the time you work it out, you can lose half your account!).
You [B]could[/B] make a living from a system that has a PF of 1.1, or even 1.05, in [U]theory[/U],[B] if [/B]you use appropriate position-sizing for it, [B]if[/B] you have enough starting capital to make it worthwhile, and [B]if[/B] it trades often enough (a system which trades only 5 times a year probably isn’t much use?). But in reality, the systems I’m trading every day all have PF’s above 1.5, so I’m not really going to “incorporate into my repertoire” anything that’s much less than that.
A profit factor as high as 2.0 is generally considered “[I]really[/I] good”, I think.
Hey thanks man. Is an EA developed by myself. The results are only from backtesting. I guess I will demotrade it until the end of the year, to see if it can make money.
This wrong.
Profit factor is = gross profits/gross losses. In the first chart profit factor is 1.32 and in the second is 1.17
IS not the best system, but at least is making money, altough a sample of 5,000 would be better. My advice: demotrade till the end of the year.
good luck
As I already explained a couple of times above, I’m taking the [B]lower[/B] of the two slightly different win-rate percentages (from long and short trades), in order to be conservative, as one should when calculating a PF - this minimises the risk that the first 700 trades might turn out to be a fraction better than the next few thousand. There’s a slight “unresolved red flag issue” in the figures quoted above, in that [U]there’s been no explanation for the statistically significant discrepancy between the short and long trade outcomes[/U], and long experience of doing this professionally leads me to believe that that’s the way to work it out, in those circumstances. If one is offered, then I’d probably revise my figures accordingly.
Not in my opinion, but thanks for your input.
good enough, the back testing would give bad result, because it would trade everything including news.
you need to run it on demo and avoid news release and low liquidity periods.