Forex Tsd Contest! - page 9

New comment
 
sunwest:
Hello,

Could someone post the link to the rule again , I am not sure about the no MM rule, do you mean that the size Lot should always be the same ?

But for example if I would like to open X order of 0.1 Lots at different price, is this possible, or you can only have 1 trade open at a time ?

Thanks

Profit calculated in pips for this contest. Your EA can open as many order at a time per chart as you can. With any lot size. Does not matter because all the profit is in pips.

For example: Pengie EA is openning many orders (pending orders). Codersguru EAs are the same (many orders). Igorad's EA is openning one order per chart at a time only. EAs are very different.

For example: 1.2340 with target with 1.2350 is 10 pips? Which lot size? Any.

It means that if your EA will increase or decrease the lot it will not affect on pips.

EA should trade together with other EAs in one Metatrader, should have magic numbers and so on.

Some EAs tested for June contest are having 0.1 lot size, some of them 1. It does not matter. Initial deposit is 50,000 for all the EAs in the beginning.

Codersguru selected one Metatrader for all his EAs. He is having 3 EAs now (two EAs are improved versions and one is very new one).

Igorad is using 1 Metatrader (1 EA which is winnning now) and Pengie is using 1 Metatrader as well (means his EA is working alone).

Igorad is keeping his EA without changes the 3rd month already. Pengie requested to change timeframe only (from M15 to H1) and improved his EA one time. Codersguru is giving the new EAs for almost every month (new or fully improved).

If somebody decide to participate so just create the thread here for the discussion with administrators about participation (i will ask about timeframe, settings and so on). This thread will be visible for administrators only.

 

June contest was finished.

The results you may find in elite section (together with the statements).

 

July EAs

newdigital:
June contest was finished. The results you may find in elite section (together with the statements).

newdigital,

Thank you very much for your efforts and congratulation for the 2nd position of the contest !

Is there any time to change/enhance the experts of July contest?

thnx!

 

hi NewDigital,

Can you report who's on the 1st, 2nd and 3rd on the winner list ?

Thanks

 
codersguru:
newdigital,

Thank you very much for your efforts and congratulation for the 2nd position of the contest !

Is there any time to change/enhance the experts of July contest?

thnx!

I posted the results on Saturday in elite section. So everybody had time to see and to change something.

Recently the July contest was started already. Today is 5th of July. Of course I may close all trades and restart it once again tomorrow. But I'm affraid that it will be 6th of July and it may be too late. Because July contest will be finished on Friday 28th of July.

For everybody: please, if you want to change something, or you want to participate in contest, so do it in the first 5 days of the month.

As to participants.

Igorad's EA is testing for the contest without any changes during the several month already. This EA may be profitable for two month testing without interruption. May be. But it was in small losses monthly.

Codersguru EA (very new one) won the 3rd place. This EA may win the first place because it was many opened trades with very good profit. I think this EA needed just few days to close but contest was finished.

My EA (new version) win 2nd place. But 3 days before contest started this EA did more than 600 pips. Those pips was not included in the contest pips calculation and because of that i won 2nd place only.

Pengie did not change his EA and it is continuing testing for July.

 

hi newdigital,

So Pengie EA won the 1st place ?

 
newdigital:
Profit calculated in pips for this contest. Your EA can open as many order at a time per chart as you can. With any lot size. Does not matter because all the profit is in pips.

For example: Pengie EA is openning many orders (pending orders). Codersguru EAs are the same (many orders). Igorad's EA is openning one order per chart at a time only. EAs are very different.

For example: 1.2340 with target with 1.2350 is 10 pips? Which lot size? Any.

It means that if your EA will increase or decrease the lot it will not affect on pips.

So you calculate profit in pips for every opened trade, no matter what lot size is and how many trades?

This is absolutely wrong. E.g. one EA opened USDCHF 1 lot @1.2340, closed @1.2350 and won 10 pips or 100 usd, other EA opened 10 trades by 0.1 lot @1.2340, closed @1.2350 and also won 100 usd, but in pips you calculate as 10*10=100 pips. This is ridiculous to determin winners in such way (try also count 100 trades by 0.01 lot ).

Do I correctly understand your profit calculation rules?

I already suggested changing rules for contest profit calculations, but nobody listened to my words.

Current rules also make no sense to participate for quite big part of strategies.

The only universal and fair way of measuring trading perfromance is by calculating % of equity growth, this is how entire world is estimating trading strategies, cause it counts in everything - from risk involved to return of investments. You can also make a rule to limit opened risk about 5-20% - this will help to get estimation of strategy very close to real life conditions.

At all, the purpose of the contest is to find the best trading system that can be applied to real trading, and not the fictitious pips profit.

 
Beluck:
So you calculate profit in pips for every opened trade, no matter what lot size is and how many trades?

This is absolutely wrong. E.g. one EA opened USDCHF 1 lot @1.2340, closed @1.2350 and won 10 pips or 100 usd, other EA opened 10 trades by 0.1 lot @1.2340, closed @1.2350 and also won 100 usd, but in pips you calculate as 10*10=100 pips. This is ridiculous to determin winners in such way (try also count 100 trades by 0.01 lot ).

Do I correctly understand your profit calculation rules?

I already suggested changing rules for contest profit calculations, but nobody listened to my words.

Current rules also make no sense to participate for quite big part of strategies.

The only universal and fair way of measuring trading perfromance is by calculating % of equity growth, this is how entire world is estimating trading strategies, cause it counts in everything - from risk involved to return of investments. You can also make a rule to limit opened risk about 5-20% - this will help to get estimation of strategy very close to real life conditions.

At all, the purpose of the contest is to find the best trading system that can be applied to real trading, and not the fictitious pips profit.

I do calculation in pips irrespective of the lot size. And no one EA is openning the trades as you described above. Some EAs opens one order per chart at the time. The other EAs are reacting on every signal to open and it is not like 10 orders on the same time. Nothing like this.

Programmers are different and programmers are coding differently. And EAs are very different because of that. Not because of good or bad coding. Because of different "forex programming schools". Forex world is devided. Just an example: one old EA (not participating in the contest) was discussed in western forex world during the more than 1 year already as very good EA. This EA is openning many orders per chart. And this EA was never (and will never be) discussed in Russian forex world only because of that. Different "forex schools".

As to the rules so we got many suggestions to change the rules so you are not alone with this feeling. For example some people (programmers) proposed to calculate profit in pips per order but the other programmers did not agree with this. And many other proposals mostly in contradiction with each other. I mean that many people proposed the different rules.

Most reasonable rules were suggested by Igorad. This rules are related to your opinion.

I have the same feeling with you: the rules may be changed. But nobody suggested on how to do it practically (just for me for example).

So we have one rule only till now: EA should be profitable in pips calculation for all testing pairs in total. And no any limitation concerning hedging, grid and so on.

As to me so I am not very happy because of one month contest duration. It should be at least 2 or 3 months. And we do not have EAs for contest enough. If we have 7 or 10 participants so we will have to improve the rules (to filter some EAs). But it was a talking with some programmers that "we have good EAs and our EAs may win but we do not want to post mql4 code even in elite section".

As to GridMACD so I was very sceptical about this EA but I looked at the trades and performance and saw that this EA is placing pending orders first, then orders were executed (not buy and sell orders simultaniuosly!) and after one or two weeks all the orders are closed. Then 1 or 2 days without trading and then it will be started once again: pending orders and so on. So this EA is not very simple as I was thinking before.

 
But nobody suggested on how to do it practically

I can suggest you, it is very simple - write down equity at the start of contest period and at the end, calculate how many percents it has grown and that's it. No need to interrupt experts and reset starting capital, etc.

If you want to apply rule for limiting exposed risk, drawdown, etc., then you should use some statements analyzing tools at the end of period (if existent tools does not have such calculations, then we can make them).

Otherwise, with current rules, I can easily write EAs that will win every time - on some breakout condition EA should open just one set of 100 trades by 0.01 lot with 100 pips takeprofit and at the end of month it will be 100*100=10000 pips profit. I don't think people on this forum will be happy to get such winners EAs as they are useless in live trading.

Also, if you count only closed trades profit and do not count opened trades, then it is even more simle to exploit current rules - open 100 orders sets both buy and sell, at the end of period you have 10k profit of closed and 10k loss of opened.

You see, that counting equity % growth is simple and effective way to have unified rule for all without exploitable weaknesses. And, after all, I repeat, this is fully in accrodance with live trading conditions.

As for longer term EAs I also already suggested to have some "grand prix" - quorterly or even yearly prizes.

 
Beluck:
I can suggest you, it is very simple - write down equity at the start of contest period and at the end, calculate how many percents it has grown and that's it. No need to interrupt experts and reset starting capital, etc.

If you want to apply rule for limiting exposed risk, drawdown, etc., then you should use some statements analyzing tools at the end of period (if existent tools does not have such calculations, then we can make them).

Otherwise, with current rules, I can easily write EAs that will win every time - on some breakout condition EA should open just one set of 100 trades by 0.01 lot with 100 pips takeprofit and at the end of month it will be 100*100=10000 pips profit. I don't think people on this forum will be happy to get such winners EAs as they are useless in live trading.

Also, if you count only closed trades profit and do not count opened trades, then it is even more simle to exploit current rules - open 100 orders sets both buy and sell, at the end of period you have 10k profit of closed and 10k loss of opened.

You see, that counting equity % growth is simple and effective way to have unified rule for all without exploitable weaknesses. And, after all, I repeat, this is fully in accrodance with live trading conditions.

As for longer term EAs I also already suggested to have some "grand prix" - quorterly or even yearly prizes.

I count closed and opened trades. You know that.

As to equity so some EAs are testing with the other EAs in one Metatrader so it is not easy. I am not using "normal" statements because of that. But I am testing the about 30 EAs in elite section so it is not possible to draw anything manually. No time for that. It should be implemented practically but no any tools for that.

And as I understand you are talking about MM but it is big limitation for some programmers.

You know that some EAs are tested during the more than 6 months already in elite section. Forward testing. And some of them are performing more good than contest winners.

Igorad suggested to calculate pips per order.

You suggested to calculate profit in deposit currency irrespective of how many orders were opened.

Codesguru likes EAs based on MA crossing with many orders opened and without MM.

Pengie is the same with Codersguru's opinion: he is coding the EAs which open the order on every single buy/sell signal. He even created the thread here in public forum about it and suggested for all the coders to use it.

You have great programming ability so create EAs which is generating many pips totally even if it is not your style of coding. I know that you can because some your EAs are forward testing in elite section during the 6 months with good results.

So create EA! Why not?

Otherwise Pengie will win every month.

As to your proposals so I think we should have the other contests concerning MM and so on.

Or your idea should be elaborated to be implemented (by me for example) and we will take into account this MM and so on.

You created one script (for pips calculation) which i am using sometimes. You may see Igorad's scripts. So create some script to summarize everything:

- pips in total,

- MM,

- profit in deposit currency,

- graphics,

- risk,

- opened/closed trades,

- profit per one order,

and (for example):

- efficiency of EA (in pips) = profit (in pips) / number of orders;

- efficiency of EA (in deposit currency) = profit (in deposit currency) / number of orders;

- total efficiency of EA = efficiency of EA (in pips) * efficiency of EA (in deposit currency).

- and drawdown, risk and whatever practically usefulness.

Just create thread in elite section and code some tools (based on some other scripts there) to implement all the rules to analyze everything easy.

12345678910111213
New comment
Reason: