On the website of our Automated Trading Championship, we pay special attention to statistics collected during contests. For the purpose of traders’ education and of search for new useful trading strategy estimation criteria, we do our best to consider the results obtained both by individual Expert Advisors and by all participants as a whole. This is why we offered our readers article “Be In-Phase” written by Mikhail Korolyuk, a well-known trader and fund manager. This is why it is quite reasonable to consider the results of previous Championships in the light of the published article.

The results of either Championship, 2006 and 2007, were processed separately. It should be reminded that 258 EAs participated in the Automated Trading Championship 2006. Only 237 Participants of them were selected. Their trading results were applied upon the zero returns curve. The Participants excluded of the review either made no trades or had no losing trades. In such cases, it was impossible and unreasonable to calculate the Avg.Win/Avg.Loss parameter for them.

Fig. 1. Participants of the Automated Trading Championship 2006 on zero returns curve.

The red line shows the zero returns limit and divides the chart into two areas: profitable and losing. Blue rhombs mark the Participants’ positions in the chart. Those Expert Advisors that are below zero returns curve are losing. Most EAs turned out to be losing. The green line shows the limit of the safe trading area. Expert Advisors located above the safe trading curve are in the relatively safe area.

The EAs located above the zero returns curve and below the safe trading curve are potentially profitable. As Mikhail Korolyuk puts it:

As you can see in Fig. 1, there are considerable fewer Participants in the profitable area than in the losing one. You can see 39 Expert Advisors there:

Table 1. List of Participants That Got into Profitable Area.

Only two of them got into the safe trading area: alexgomel (17 trades) and DxdCn (5 trades). However, unfortunately, we cannot classify any of them as safe ones, since they made too few trades to consider the obtained result statistically important.

Now let’s consider the results of the Automated Trading Championship 2007 in which already 603 EAs participated. 558 of them are considered in our analysis, because the correct calculation of AvgWin/AvgLoss is possible for them.

Fig. 2. Participants of the Automated Trading Championship 2007 on zero returns curve.

Now there are much more points on the chart, but the nature of their distribution has not changed much. Most Expert Advisors are still in the losing area, a small part of them is in the potentially profitable area, and only 7 EAs are in the safe trading area.

However, from Table 1 we know that many Participants located in the profitable area didn’t make even 30 trades, so we prepared a new chart displaying only the EAs having 30 or more trades made on their accounts. The new chart is shown in Fig. 3.

Fig. 3. Participants of the Automated Trading Championship 2007, which made 30 or more trades: zero returns curve.

Now there is only one Participant in the safe trading area: waaustin (56 trades). For waaustin, the ratio between the average winning trade and the average losing trade (AvgWin/AvgLoss) reaches the huge value of 8.05. In other words, the average profit of this Participant was 8 times higher than the average loss! How did he manage to do it?

Participant waaustin took only the 18th place in the Automated Trading Championship 2007. If we look through this Participant’s reports, everything will become clear. Only one of 56 trades made is losing, other 55 trades are profitable. Normally, such ratio between profitable and losing trades characterizes strategies that “sit out” losing positions. The MAE distribution diagram confirms this statement:

Fig. 4. MAE distribution for ATC 2007 Participant waaustin.

Considering that the net profit of this Participant made 13 284.10 dollars, this trading style does not appear so attractive with such drawdowns.

Thus, no Participant could approach to the creation of a “holy grail”. On the other side, the first Winner at ATC 2007, Better, whose victory was certain, is obviously located in the profitable area, although he has not got into the mentioned safe area. 65.9% of trades made by his EA were profitable (%Win = 65.93%), and the corresponding value of AvgWin/AvgLoss = (100% – 65.93%)/65.93% is equal to 0.51. This means that, at AvgWin/AvgLoss = 0.51, Better Expert Advisor would have had zero returns. Practically, his EA has the value of AvgWin/AvgLoss=1.12 that exceeds the parameter of zero returns more than 2 times!

We can draw a conclusion that stable trading algorithms can also be located outside the safety area. Or, in other words, this safety area is, indeed, a certain ideal target to work for, but you shouldn’t try to create a trading strategy only to get into that area.

In the end, we would like to consider the trading results of two more Participants of ATC 2007: Newman2 (40 trades and the 8th place) and warwickquant (44 trades and the 31st place). As you can see in Fig. 3, these Participants have not got into the safe trading area, but are located very closely to it.

Table 2. Some Parameters of Participants Newman2 and warwickquant.

For both Participants, a small percentage of profitable trades and a high ratio between the average profitable trade and the average losing trade are typical. Such parameters are also typical for trading systems that follow the trend, cut off their losses and allow their profits to grow.

Fig. 5. MFE and MAE distributions for ATC 2007 Participant warwickquant.

As you can see, the MFE and MAE charts confirm this conclusion for Participant warwickquant. The MFE chart shows in the X-direction the highest potential profit that could be gained at closing a position.

There are two methods of closing trades at a fixed level: you can either limit your losses from below using StopLoss or limit you profits from above using a Take Profit order. For trend-following systems, the first way is more typical. No Take Profit is placed, but a StopLoss is present. This allows you to take the largest profit if your trade has been successfully opened with the trend.

At the same time, some trades that have previously got into the profitable area will be closed with losses, if the trend turns around in an adverse direction. It is the MFE chart in which some trades are shown that had unfixed profits within certain periods of time. However, those trades were closed by StopLoss after the price had moved in the reversed direction due to changes on the market.

Fig. 6. MFE and MAE distributions for ATC 2007 Participant Newman2.

The conclusion is correct for Participant Newman2. In the MAE charts of both Participants, it can be clearly seen that the losses are limited at a certain drawdown-resistance level (blue line). At the same time, in the MFE chart, we can see trades that used to be profitable at their best moments, but were closed with losses by Stop Loss, since no Take Profit had been provided.

Conclusions:

• The trading strategy estimation methods proposed in Mikhail Korolyuk’s article “Be In-Phase” is certainly interesting and informative. Although we have received more evidence that there is no simple formula of creating a “holy grail” in trading, we can see that profitable strategies are easily identified using zero returns curve.
• Moreover, we can assume with a certain degree of confidence how the profits were obtained when trading with one or another strategy. For this, it will suffice to see in what chart area the strategy is located when analyzed according to the above methods. A strategy may produce frequent profits, but the average profit will be quite small as compared to the average loss. A strategy may produce rare winning trades, but then the average profit must significantly exceed the average loss. Or it can be a strategy ranging somewhere between the two above.
• This can be exemplified by the results obtained by the Winner of the Automated Trading Championship 2007, Better. His Expert Advisor made 408 trades, almost 66% of them having turned out to be profitable. At the same time, the value of AvgWin/AvgLoss was equal to 1.12.
• It means that it is unnecessary for a stably profitable strategy to have a very high percentage of profitable trades or a large AvgWin/AvgLoss ratio. The main thing is that it must be located as high as possible above zero returns curve, but without unnecessary “sitting out” losing trades.