More or less profit for the bot competition?

[Leg0nd]

Been playing around with a strategy tester on Metatrader 5's platform.
I think I have a strategy to win the competition with, but I'm unsure.

The rules for the competition want the bots to run using $750,000-$900,000

In this setup i used 700k. It works just as well with 900k.

Do I need to make it risk more to make more money? Or is 1.5+ Million good enough? I have a version that will make over 3 million using 900k... i just don't know if i should submit this one or that one.


Here's what the back-test looks like on the 1.5 mil version. (Tested it from 01/01/2012- as far forward as it would test, 04/--/2012)




Here's the trades it makes and some background info.

History Quality: 100%
Bars: 789 Ticks: 6171203
Total Net Profit: 1 623 700.00 Gross Profit: 1 623 700.00 Gross Loss: 0.00
Profit Factor: 853 Expected Payoff: 50 740.63 Margin Level: 1045.39%
Recovery Factor: 3.06 Sharpe Ratio: 0.63 OnTester result: 0
Balance Drawdown:
Balance Drawdown Absolute: 0.00 Balance Drawdown Maximal: 0.00 (0.00%) Balance Drawdown Relative: 0.00% (0.00)
Equity Drawdown:
Equity Drawdown Absolute: 165 500.00 Equity Drawdown Maximal: 503 200.00 (29.72%) Equity Drawdown Relative: 34.59% (282 600.00)
Total Trades: 32 Short Trades (won %): 21 (100.00%) Long Trades (won %): 11 (100.00%)
Total Deals: 64 Profit Trades (% of total): 32 (100.00%) Loss Trades (% of total): 0 (0.00%)
Largest profit trade: 281 200.00 Largest loss trade: 0.00
Average profit trade: 50 740.63 Average loss trade: 0.00
Maximum consecutive wins ($): 32 (1 623 700.00) Maximum consecutive losses ($): 0 (0.00)
Maximal consecutive profit (count): 1 623 700.00 (32) Maximal consecutive loss (count): 0.00 (0)
Average consecutive wins: 32 Average consecutive losses: 0



this strat only traded the euro/usd.
What do you think? Is this enough back testing to leave as is, or should i make it more profitable?

I can post the trades if you guys want to see them

[SithJawa]

What competition :O?

And how does this bot work? to be honest it looks like it is waay over optimized, for a system to have no losses is near impossible :3 the key is to minimize losses more so than maximizing profit

Automated Trading Championship 2008

they have an automated trading championship, this is the link to the 2008 one, as you can see the winner tripled his account in three months with his ea but he had many losses and large ones too

Another thing is the testing data should be much greater, atleast 2 years i'd say, you want to know if the ea can perform overall profitably through all market conditions so a large period of time is needed for the tests

[Leg0nd]

I don't know how you can over optimize a bot. This bot works even if backtested all the way back at the beginning of the market.

Have you made a bot that runs 96%-100% winning average starting at the beginning of this year til April? I have tested so many different strategies. But none have come close to this bot. I used the macd to make it

[SithJawa]

You can over optimize a bot by tweaking it to specific periods of time that reduces its overall profitability while increasing it during the period you are testing it in

And no aha i have not made a bot that profitable if it is indeed that accurate, you must have a lot of tight filters on it then to have no losses, i dont suppose you would allude to the theory behind the strategy? If i were you i wouldn't give out specifics :P

EDIT: if you could post some of the trades that would be interesting to see i'd love to theory craft

[Leg0nd]

You can over optimize a bot by tweaking it to specific periods of time that reduces its overall profitability while increasing it during the period you are testing it in

And no aha i have not made a bot that profitable if it is indeed that accurate, you must have a lot of tight filters on it then to have no losses, i dont suppose you would allude to the theory behind the strategy? If i were you i wouldn't give out specifics :P

I'm going to forward test it as well, no funny filters or anything. Just the right chart matched with an indicator that has the right settings. 3.5 mill isn't enough even. I'm going to up the lots so that I will make 10+ mil in this time frame. The automated trading championship lastsfor 3 months. Whoever makes the most profitable bot wins. So tweaking the bot to take a lot of risk, that makes the most money will win. Unfortunately if I don't do it, someone else will get lucky and will do it to win. It's not the safest one who gets the most profit. It's the one who called the right trades with the most amount of money in those trades.

I might share the settings if I make a new bot. One that does this with less risk. All I have to do is Lower the lots per trade and the risk will evaporate

[goldylox]

Hi, nice job.

If it is for a competition, you might need to set the risk high. You probably will not win by using safe money management.

Most bots that win competitions are fairly useless for real trading. They just get lucky with the market conditions.

Maybe try to predict the market conditions at the time the bot is trading. What will happen with the euro, etc. You could set up a bot that only trades short on the euro if you think it is doomed for eg.

[Leg0nd]

Thanks. I have been thinking the same thing. I have two or threes bots that run 100% winning average. One of them would have won 75 trades in a row from January until the end of march.

The challenge is how do you make the bot backtest positive at least 8 trades in the beginning of the year, but not lose and still win for the October 2nd trough December 28th competition? 100% isn't hard to achieve, but I want it to stay 100% all year starting in january and starting in October. any ideas? Plus using 700k or 900k what do you think the winner is going to be able to put out? How many million should I keep trying to go for?

[PerchTird]

Sorry I am confused - I start many posts like this - If you have a bot as you say you do? why not open up live with $50 and make the $2000 prize money you would have won? -- Also - read fine print;: if you have a winning bot - does the competition required you submit the code - before they will award the winnings? That in itself would keep me at bay - - I'd want to keep my 'winning bot' to myself - at least for the short term.

Anyhow goodluck-

[Leg0nd]

Sorry I am confused - I start many posts like this - If you have a bot as you say you do? why not open up live with $50 and make the $2000 prize money you would have won? -- Also - read fine print;: if you have a winning bot - does the competition required you submit the code - before they will award the winnings? That in itself would keep me at bay - - I'd want to keep my 'winning bot' to myself - at least for the short term.

Anyhow goodluck-

Thanks. I'm working on a different bot using an indicator I developed with Zechat doc to enter into the competition with. I don't think I need to submit the code, but either way as long as zechatdoc is cool with it no big deal.

As far as this bot goes, I don't think it will continue to win forever. Apparently most bots have this problem. I'm going to pull up the specs and share them with y'all. Bout to take a picture

[MoneyNVRSleeps]

Nice bot,

Raise risk to 65% per trade, compounded

If you want the code, let me know