We go on collecting and analyzing statistical data obtained in the course of the Automated Trading Championship 2007. This present report updates information published before on equities of Expert Advisors in different groups. Besides, in analyzing the new data we managed to find some interesting interrelations between behavior of Expert Advisors and their profitability.

The Most Popular Securities

The preferences given to securities traded by Participants of the current Championship have changed as compared to the preceding contest. Like in the Automated Trading Championship 2006, the most of Participants choose the currency pair of EURUSD (200 EAs, or 33%). This currency pair was traded by 141 (or 54%) Expert Advisors last year.

The second place by popularity is also kept by GBPUSD, however, the amount of EAs trading this security has greatly increased - 185 Expert Advisors (31%) this year to 70 (27%) in Championship'2006. The currency pair of GBPJPY forged in the third place. Only 7 EAs were attached to this symbol last year, whereas there are 116 of them this year. Another pair based on yen, EURJPY, has also increased in popularity. Only 2 EAs were attached to it in 2006. This year there are 39 of them.

The current distribution of Expert Advisors by currency pairs seems to pretty well reflect the real preferences given by traders to different securities. It is a surprise that the pair of EURGBP being unpopular last year (only 1 EA in ATC'2006) attracted 11 EAs this year.

The Most Popular Timeframes

The distribution of Expert Advisors by timeframes has not changed much within a year. H1 is still the most popular timeframe - 188 EAs chose it this year (91 - in ATC'2006). The 15-minute timeframe took the second place. It was chosen by 105 developers (to 33 in ATC'2006). M1 took the third place. 71 Expert Advisors are trading on one-minute timeframe to 45 ones in 2006. We can conclude that preferences concerning timeframes are rather stable.

Multicurrency Expert Advisors

Nowadays, it is a common knowledge that MetaTrader 4 Client Terminal and the development environment of MQL4 allow traders to develop and use multicurrency Expert Advisors. More and more traders use in their EAs the opportunity to analyze data obtained from any symbols or timeframes. This fact is also reflected in the Automated Trading Championship 2007.

Approximately 6% of Expert Advisors participating in the Championship in 2006 traded on several currency pairs simultaneously. The amount of multicurrency EAs increased to 82 this year, which makes nearly 14% of the total amount of Participants. Seemingly, the number of such strategies tends to grow. The amount of EAs trading on only one currency pair makes 521, or 86.4%, of the total amount of Participants. It is to be recalled that they made 94% last year. Note: Only those EAs were considered that had made at least one trade.

The Amount of Positions Closed per Day

It was also interesting to compare the position closing processes in both contests. It should be reminded that the Expert Advisors in Automated Trading Championship 2006 tended to decrease the amount of positions closed per day. This amount ranged from 1000 to 600 on the first days of the Championship, whereas the mean amount of closed positions (red line in the diagram) fell two times and became just a little more than 400 at the end of the fourth week.

The situation is different this year (Fig. 5). It is safe to say that the amount of closed positions has not changed much by the end of the first month of the contest. The most interesting thing is that, even though the number of Participants increased more than 2 times, the amount of closed positions did not practically change on the first days and ranged about 900 trades per day.

Seemingly, the results of the last-year Championship paid off - the developers understood that they should not scalp. Expert Advisors started to open positions in a more restrained and thought-out manner. This explains the very slow reduction in amount of closed positions.

Ratio between Profitable and Losing Positions

Many traders try to optimize their strategies by the ratio between profitable and losing trades. Like in 2006, we calculated how many positions were closed with profits, how many - with losses, and how many - at the Open price. Swaps were not considered at that.

20 780 orders were executed within the month. 10 769 of them (51.8%) turned out to be profitable. 9 363 of the closed positions (45%) had negative results. 648 positions (3.1%) were closed at Open price. The data of the current Championship do not really differ from those of year 2006. In Fig. 6 we can see that the ratio between three groups of orders has not changed much within a year. We can also observe that the amount of profitable trades dominates over losing ones.

Although the most of positions are closed with a profit, the average profit for all orders per day most frequently turns to be negative. This is because the total loss is greater than the total profit of all positions. In other words, the average loss per position exceeds the average profit. Thus, a profitable position makes averagely $148 and a losing position makes losses of averagely $226.

This year, an average profitable trade makes $306.73 in October (average swap makes 0.70 dollars for profitable trades), an average losing trade makes -473.53 dollars (average swap makes -1.05 dollars for losing trades). Thus, the conclusion of that Expert Advisors trading in the current Championship are much more aggressive is completely borne out. However, the ratio between the average profit and the average loss has not really changed and makes 1.54. Perhaps, we have discovered a new constant here. The comparative data are given in Fig.7.

Let's calculate the average result of all closed positions per day. For this, we take all orders executed on a day and sum up their results. The obtained value is divided by the amount of closed positions. In such manner, we find the average profit per order for each day of the first month. The results are represented as a blue line in Fig. 8.

The graph shows that the daily average profit was negative (with rare exceptions). Thus, despite of a small domination of profitable trades over losing ones, the aggregated balance of Participants decreased day by day. The red line is an approximating least-squares trend line.

Like the last year, the average balance of Participants decreases, whereas the average profit per trade increases. The losing EAs contribute less and less to the calculation of the average profit per trade. It is likely that the total financial result of all trades will have moved to the profitable region by the end of the Championship.

Amount of Trades, Positions Holding Time, and Profitability

Like the last year, we divided all Expert Advisors into 10 groups by the value of their equities as of the 31st of October 2007. 60 EAs with the highest equities entered the first group, whereas the EAs having the lowest equities came into the tenth group.

We also found the average amount of trades made per account in each group. In the last-year Championship, we saw that the oversize amount of trades made characterized losing strategies. What will this index show this year? We were surprised with the results again.

The general conclusion remains the same: Profitable trades are characterized by a moderate amount of trades made - approximately 20-30 trades per month. However, the Participants seem to have learned the lesson of the preceding contest very well: The trading style of all Expert Advisors is more reserved this year. In the Automated Trading Championship 2006, EAs from groups 7, 8, 9 and 10 managed to make in average over 50 trades within 4 weeks.

This year, only EAs from group 10 made in average over 76 trades per month. Expert Advisors in other groups did not exceed in average 40 trades per month. The position holding time is one more index that can characterize a trading strategy. The position holding time for one-minute timeframes will be essentially shorter than that of a trading strategy that considers only data obtained from daily charts.

At the last-year Championship, we calculated the position holding time average for each group. The analysis of trades made by Participants in October 2006 resulted in a conclusion that the short position holding time is typical for losing strategies. We also calculated this value this year and compared it to the data of the previous Championship.

We can see again that the Expert Advisors participating in the Automated Trading Championship 2007 differ from those participating in the last-year Championship. The reduction of the position holding time is observed in all groups. EAs in profitable groups are still holding positions in average longer than those in losing groups. However, the difference is not really significant now.

Seemingly, the most of Participants of the current Championship have concluded that the optimal position holding time must range from 10 to 20 hours. Thus, the most of them voted for that it is this position holding time that is optimal for winning at the three-month interval of the contest. The smaller value will decrease the chance to win, whereas the larger value will increase the probability of a correction and, respectively, lowering of profits.

Conclusions

1. Multicurrency Expert Advisors gain in popularity.
2. Developers' preferences extend to more securities now.
3. The ratio between the amount of profitable and losing trades does not practically change. A little more than a half of all positions are closed with profits.
4. The average loss is still dominating over the average profit. The ratio between the average profit and the average loss is approximately 1.54.
5. The average amount of trades per month decreased, Participants don't want to open positions too frequently.
6. The position holding time average decreased, too. Now, Expert Advisors both open smaller amounts of positions and reduce their holding time.