Thread: Forex Robot World Cup

The subject competition has been under way for about a month now and there have been a number of surprises. Already there are 6 competitors (of 24) down > 60% and one was bust after 11 days. I'm not sure what this says about the competition's qualifying rules and about backtesting in general. It is probably more a reflection of the flaws in backtesting than the qualifying rules.


The leaders have a variety of ways to suit every style. There is one with a 97% win rate. Some trade frequently, others infrequently. Some go for the big wins and losses and others want smaller amounts. Here is some statistics I put together on the weekend.



.

Looks interesting. I didn't even know there was a competition going on. Let us know how the results look.

Originally Posted by Kenny Rogers

Looks interesting. I didn't even know there was a competition going on. Let us know how the results look.

I'm not sure about posting links on these forums, so if you search for Forex Robot World Cup, you will find it. You can follow the results there and get other information such as the rules. I will be doing a followup analysis at the end of January, just before they go on sale.

The competition completed this week and the better robots will be offered for sale in a couple of weeks. I did some statistic analysis of the robots' performance and have included the results below. The records for SuperVolcano for Jan were messed up.

Another competition will be starting sometime in March.

If you examine the equity graphs of the leaders, the winner's equity is all over the road, while #2 is a nice straight line. The rank beginner would choose #1 as the best EA. The slightly more experienced person would realize that ending up in the #1 spot doesn't do you much good if the equity curve passes through 0 on the way. So the more experienced would tend to favor the second best because the curve is much straighter.

And so there's less risk. Or is there? I examined the trade-by-trade of HiRIDER, the #2 robot, and there's a hidden gotcha that'll get ya if you miss this. The trade sizes are small, typically 0.03 lots, which gives it that nice smooth equity curve. But the problem is that the EA takes multiple small trades in the same direction (all on EURUSD) all within a minute or two. So ten small trades stack up to one large trade. Doing the arithmetic shows that this EA risks 50% of the equity of the account on each aggregate position. A single adverse move will drop the account equity by 50%. By luck and by having the stops set well away from the market, this EA didn't encounter a large enough adverse move within the time allowed in the competition. But as any trader knows, no matter where they are set, all stops eventually get hit, its only a matter of time.

Move along folks, there's nothing to see here.

Originally Posted by chipdude

If you examine the equity graphs of the leaders, the winner's equity is all over the road, while #2 is a nice straight line. The rank beginner would choose #1 as the best EA. The slightly more experienced person would realize that ending up in the #1 spot doesn't do you much good if the equity curve passes through 0 on the way. So the more experienced would tend to favor the second best because the curve is much straighter.

And so there's less risk. Or is there? I examined the trade-by-trade of HiRIDER, the #2 robot, and there's a hidden gotcha that'll get ya if you miss this. The trade sizes are small, typically 0.03 lots, which gives it that nice smooth equity curve. But the problem is that the EA takes multiple small trades in the same direction (all on EURUSD) all within a minute or two. So ten small trades stack up to one large trade. Doing the arithmetic shows that this EA risks 50% of the equity of the account on each aggregate position. A single adverse move will drop the account equity by 50%. By luck and by having the stops set well away from the market, this EA didn't encounter a large enough adverse move within the time allowed in the competition. But as any trader knows, no matter where they are set, all stops eventually get hit, its only a matter of time.

Move along folks, there's nothing to see here.

So you are looking for an EA with a perfect equity curve with a small SL , decent RR ratio and with no drawdowns or correction phases?

will anyone buy these EA for 999$ ??

Originally Posted by forexhr

So you are looking for an EA with a perfect equity curve with a small SL , decent RR ratio and with no drawdowns or correction phases?

I sense some sarcasm there...

It is a stupid competion that will not reflect how good a programmer or develloper can devellop.
Because if one wants to join the competion he needs to sign a contract that the EA that he enters in the competion is no longer his EA but that the EA becomes automaticly the full property of the organizers of the competion and that the person who entered the EA looses all commercial right on his EA.

So If you have a good EA you will not enter that competion because if you don't make into the top 3 you don't have anything. No prize money and you loos all right and intellectual property.
So which kind of EA will people enter ?... An EA that is over-optimized and hoping on some lucky period where the EA will perform quite good and if not that you are not bothered that you lost that EA.

If you would ever buy an EA that performed well in the ranking be awere that it will be an over-optimized piece of junk and don't let the results over a couple of months fool you.

Friendly regards...iGoR

iGor is damn right!