In this episode of the fxcm educational series, we’re going to be taking a look at trading strategies now, what I’m going to do is to illustrate some key concepts so on a step-by-step basis as to how do traders deal with and create trading strategies. In the first place, and I’m gonna do it by showing you an example of a trading strategy as well, so I like to separate this up into a couple of steps. So today, what we’re gonna do is we’re gonna start with step number one which is to build our rules, so a rule set is just what is the strategy?
What is the signal that we’re using to indicate whether or not a trader should be buying or selling shorting closing their trade or what-have-you? Now the system that I’m going to illustrate here is actually a very common one. We’re gonna be looking at technical divergences, so I’m gonna walk you through the basic ideas behind a long entry and a short entry as an example now in step number one here, if we’re trying to define our our entry signal, I’m going to use a Bearish divergence and then I’m also going to tackle a bullish divergence as well. So let’s just walk through the rules for a bullish entry now initially – and here what I’ve done is I got the black line here.
That represents basically the price, so the exchange rate itself and as you can see, the price of the exchange rate has been forming higher highs, while at the same time in here, I’ve drawn a diagram of a stochastic oscillator which we’re going to take a much closer. Look at here in just a minute on the charts. Well, at the same time that the price has been forming higher highs, the stochastics oscillator, which ranges it’s a line that ranges between 0 and 100 and when it gets above about 80, that’s sometimes called an overbought region. Now it’s meaningful when it’s overbought, if at the same time the that these exchange rate here is forming higher highs, if the so Kasich’s indicator is forming lower highs at the same time, so there’s a bearish divergence then would look at this as a bearish entry signal.
So that would be part of a rule set here for a short entry or potentially as an exit signal. Now, a long entry would be similar, except just to in Reverse. So imagine here, for example, that we have a price that has been declining in value. The exchange rate and it’s been forming lower lows, while at the same time we actually have higher lows on the oscillator, so in this case the stochastics. So if we get a lower lows on the price of the exchange rate, higher lows on the stochastic oscillator, well, that’s a bullish divergence.
So a trader would use that as a signal to go long, that particular exchange rate or alternatively, this could also be seen as an exit signal from a traders in a short position and the bullish divergence is basically telling them that their expectation is that the market Is going to continue to the upside now, in both cases, there’s a rule of thumb here that will actually help improve the overall performance of a system like this, and I bring it up because it should generally be a component of most trading systems, especially those that, Where you’re trying to take advantage of the trend – and it is in fact the trend so if the trend is very supportive for the trade entries.
So let’s say in this example that you get a bearish divergence like this and although the market has been recently moving higher, the long-term trend was actually negative, but the longer-term trend was actually negative, or in this case the longer-term trend was actually very positive and it Had just recently pulled back into this little divergence that helps to put the odds a little further in favor of a trader who is looking to be able to enter a long position there or in this case a short position here is: are those signals conforming to The bias of that longer-term trend, so, let’s take a look at a couple of specific examples on the charts of both a bearish divergence and a bullish divergence. You can see the set up that I was discussing at the white board here in this chart with the Australian dollar to the US dollar.
Currency pair long-term trend was in fact negative. Now the currency pair had started to rally a little bit in April and May of 2015, where the price was actually making slightly higher highs, while at the same time as you can see here on the stochastic syndicator, it’s actually making lower highs. Now that’s a classic example of a bearish divergence and it was completed on the 18th of May when the stochastics fast line crossed below its slow line.
Now, ultimately, the exchange rate dropped by about 560 pips to the downside before it started finding a little bit of support, but ultimately it the exchange rate actually dropped at 1257 pips before finding a temporary bottom in September. Now the example here that we have of a bullish divergence once again we’re going to defer to that longer-term trend. So you can see here with the pound to the yen exchange rate. We have a longer-term trend to the upside now between the end of march and about the middle of April, we started to see a couple of bottoms forming on the price or the exchange rate itself.
Now these bottoms, the second one, was actually deeper than the first one. However, that was not matched by the bottoms on this pick, a sticks oscillator. So, as you can see here, the oscillator was actually forming higher lows. Well, those lower lows in the price and those higher lows on the oscillator: that’s the nature of a bullish divergence now, eventually, the exchange rate actually rallied about 2000 pips before reaching resistance. In June now, once we’ve come up with a trading idea, something that seems to make sense. What we need to do is to move on to step number two, which is back testing now back testing is basically what we do when we are going back into historical data and we’re applying our rule set.
So our trading idea to that historical data to see well, how would that have worked in the past now back testing can contain certain risks and problems in that investors, sometimes over optimize a little bit or they make assumptions about the data in the past. That may be so refined and so precise that they’re not very good representations of what we should expect in the future. So, generally speaking, what I would suggest, when you’re back testing, to try to eliminate some of those problems is to do essentially three things so number one is to test your trading system over different time periods.
So, for example, if you have a trading system that relies on daily bars or something like that, well try to apply the same trading system to hourly bars or even 10 or 15-minute bars. Something like that to see whether or not this may be just a fluke or an anomaly now not only that, but also test your system over different market conditions. So when we say time periods well, you might have found that something worked really well. This year, we’ll go back five years in the past and look at that historical data and see. Did it still work five years ago as well, because we don’t know when the market conditions of five years ago may actually wind up reasserting themselves? Number two is: if you have a system that you’ve developed on – let’s say a particular currency pair like the euro dollar, or something like that well test it on different pairs.
So by doing this, you’ll be able to improve your confidence that this isn’t something that may just be an anomaly or, where you’re doing a little bit of overfitting if it works, for example, in the euro dollar, but does not work for whatever reason on the dollar. Yen or the euro yen, or something like that.
Well, that may be telling you something about the validity of your rules and potentially, of course, this the potential for success in the future and then finally make sure that you’re doing a large enough test in a sense of sample size. So a large sample size doesn’t have to be hundreds and hundreds of trades, but it should certainly be more than just a couple. So, for example, we might defer to some of the traditional things that we know about statistical sampling, so looking at trades are a sample set of at least more than thirty instances is going to help to improve your confidence level that whatever system that you’re back testing.
That the results that you’re getting from historical data is likely to essentially be replicated into the future, but the larger your sample size is the more confident you can be that that is in fact going to be the case. The smaller your sample size is the more likely it is. You may be dealing with an outlier or some kind of an anomaly now in step number three we’re gonna talk about risk control. Now I think that there are really two important components of risk control that we need to think about. The first one, of course, is what you might imagine, which is where’s my stop-loss. At what point do I assume that if the trade has gone against me, I want to get out of this trade and just assume that the analysis is dead, that it’s no good?
The second level, of course, is where do I think that the trade is done in the sense that I’ve reached a likely profit objective, and I want to remove my risk by taking my profits off the table, so this would be the first component here of risk Control and let’s take a look at a specific example of what we mean by a stop-loss with this system that we’ve been evaluating here in this video. So if we revisit this trade on the pound yen, just after that, divergence was actually confirmed on the 15th. Let’s imagine that an investor was evaluating an entry on the 16th well, if the price had not actually risen, but instead had reversed and begun to decline and had taken out the low of that divergence which occurred on the 14th of April.
You can see that here the low price at the time was 174 0.866, so their stop-loss would have been placed a little over 200 pips to the downside from their entry point, because the assumption is that at that second low had actually been taken out with a Third, lower low, well that divergence is apparently not working out now. The second aspect here of risk control is position sizing position.
Sizing is oftentimes overlooked, but it helps to stay consistent. So it’s not uncommon for a trader to feel really confident about particular trades. So they get in very heavy or to feel very scared of a particular trade, and so they get in very very light and they, since they can’t predict the future, they wind up sabotaging themselves a little bit or at least risking the sabotaging themselves, because they don’t
Know which of those two trades ultimately could be their big winner, so position sizing becomes really key. Now we can approach this very simply. It’s not very complicated and one of the common ways to do this is by evaluating each trade as a fractional percentage of your overall account equity. So, let’s put ourselves in the example of a hypothetical trader here: let’s say that there was a hypothetical trader and she had account equity of $ 10,000.
Now she had decided that in any single position she was willing to put at risk 5 % of her overall account equity. So we’ll say here the upper limit. As far as the position size is 5 % of 10,000. Well, if we times 5 % by 10,000 – and we find out that this trader at least right now is willing to put at risk $ 500. Now that’s part of our question here we want to be able to answer what is the amount of money in notional terms that this trader is willing to put at risk in a particular trade, but we need to know what the stop-loss is or what the expected Loss is in a particular trade before we can translate this into an actual position size in terms of how many Lots. So, let’s say that a trader here that their stop-loss they’ve decided that they’re willing to give up a hundred and ten pips so 110 pips.
Now to keep our math nice and easy. Let’s say that, where this trader is using many Lots where each pip and in this case is worth $ 1 so because of the currency pair, that they’re actually trading will say that the dollar is on the base side of that, so that each pip will be Worth $ 1, so what we would do is we would take this 500 dollars we’re gonna divide it by a hundred and 10 pips, because we know that’s where the stop-loss is now. If we do that, of course, we’ll find out that we wind up being about going into from 110 going into 500 goes in about 4.5 times.
So that means that a trader here, if they wanted to maintain a consistent position, size of 5 % of their account equity and as their account equity, grows. Of course, their position, size and dollar terms is going to continue to go up. But this way this trader knows that this position deserves a size of, let’s say for mini lots and 5 micro lots, so that would equal exactly what their position sizes planned to be in notional terms now. Our last objective here is to develop an expectancy and to actually keep that expectancy up to date. So what does this actually mean? Well expectancy is a way for us to think about what to expect over time from our average winners and losers, and what’s the difference between the two.
So, let’s think about this now we need to know two things about our at least our back-testing. When we initially start and then on an ongoing basis well in our forward testing, if you will, what are these two ratios of looking like then as well? The first one is a win to loss ratio. Now our win the loss ratio. What does this mean? Well, it basically means what percentage of the time do we win and what percentage of the time do we lose? So, let’s just do a hypothetical here. Let’s say that a trader is winning about 30 percent at a time and they are losing about 70 percent of the time all right now. That may not look very good, but we have another ratio to cover here. So what does the risk to reward? Look like now.
This can definitely change the way that a particular trading system is really performing, regardless of what its win-loss ratio looks like. So let’s say that when a trader is right that their right to the tune of about $ 2,000, so, let’s just say hypothetically here so hypothetically, they stand to be rewarded on average about $ 2,000 for a winning trade and when they’re wrong, they wind up losing About $ 500, now we have all the information that we need to develop an expectancy and the way that we do that is. We multiply the potential loss by the amount at risk, the potential for a winner and the amount of reward, and then we’re going to net those out. So, let’s do that we’ll do the loss in the risk first, 3.74, 70 % times 500 gives us an average loss here so on any given trade.
We think on average that that’s going to lose 350 dollars. If we wait this, assuming that the other 30 percent are winners all right now in a given trade, we would assume that the winners so we’re gonna times 0.3 for 30 % times 2,000 times 2,000 equals $ 600 all right now. What we want to do at this point is, we want to make sure, and our objective here is to make sure that our expectancy is positive. So how do we know? Well, all we need to do is just net these two numbers out so over time. If we have a positive expectancy, then that means that our weighted winner is larger than our weighted loser. So in this case it is to the tune of about 250 dollars per trade. So what does that mean that basically means that, on an average basis, an average trade, a trader would have an expectancy of earning 250 dollars on average?
Doesn’t that $ 250 not equal to their winners, which are much larger and it’s obviously not the same as their losers, which are negative? What this tells us is, if it’s positive, then a trader would assume that there’s a certain likelihood that they’ll be able to be profitable in the future.