Simple trading of last day's High and Low

To add comments, please log in or register
Yuri Predborski
561
Yuri Predborski  

Hi!

I want to share the system I made recently. You can have the EA too (its time-limited).

The reasons I'm going to share it:

1. I want to make it better.

2. I want to change it so it can trade all currency pairs.

3. I only have 1 PC and can't test everything.

4. Having only 40% wins makes me depressed.

Even though average loss is 22 pips and average win is 65 pips. Max loss is 30 pips. Max win was a huge trending move of 390+ pips. Even though its a great thing, it still makes me depressed. 45% wins for 2006-2007 and almost 40% for 2004-2007. I want to make it better.

Ok. Details on the system.

1.

You calculate the High and Low of previous day (take the highest and lowest point of yesterday, yesterday is from 0:00 yesterday till 0:00 today).

2.

You set 2 breakout orders - place a buystop if the price (bid) is above MA (place buy-stop 2 pips higher than last day's high) and sellstop if the price is below MA (place sell-stop 2 pips lower than last day's low). I place orders from 7:00 to 8:00 GMT. I don't place orders if the price is closer than 3 pips to MA. I use simple moving average and open-prices model. I checked on others and the result was the same for (high+low)/2 and typical (H+L+C)/3 and weightened close (H+L+C+C)/4.

If you have a better trend-following system, please let me know. Under trend-following I mean filtering out the wrong-way trading. Also I noticed that sometimes the price may move from one side of the MA to the other side. If that happens, both buystop and sellstop are opened. But once the buy-stop is triggered and becomes buy, we delete sell-stop. That helps. A very tiny bit.

3.

You delete orders if nothing happened near the end of the day. I delete them at 21 GMT. Please note that DelTime parameter does NOT use GMT shift. Be careful.

4.

If one of the orders gets opened, you trail it till either TP is hit or SL is hit. My TP of 1000 didn't get hit a signle time - the best result was almost 400 pips. I use TrailingStop of 45 and StopLoss of 30 (trailing stop is used as soon as we are in profit, no wait).

Winning ratio is only around 40-45% but the good thing is (as I said) that every win is 3 times more than 1 loss. And you get 3 losses and 2 wins in a row on average which gives this system profit factor of 2 (1.8 without MM, 2.08 with MM).

I tested only with GBP/USD, H1 timeframe. It won't work with any other timeframe. One of my goals is to find parameters or make changes so it makes profit with other currencies. I'm fine with H1 TF.

EA parameters:

GMT.Shift=2; //no need to explain. Can be negative. EST is negative.

Start.MS=7; //starting hour, affected by GMTshift

End.MS=8; //ending hour, affected by GMTshift

Hours.MS=24; //number of hours to find high/low

BuyS.MS=2; //shift up from last high, can be negative. Don't change

SellS.MS=2; //shift down from last low, can be negative. Don't change

DelTime.MS=21; //hour when you want your orders deleted, unaffected by gmt.shift

TP.MS=1000; // takeprofit. target. unreachable dream (I have systems that CAN reach this target but this one won't)

SL.MS=30; // small stoploss

TS.MS=45; // trailing stop. medium one. 85 seemed to work but this is much better. smaller ones decrease profits greatly which increases drawdown

MA.shift.MS=3; // shift goes for shift from MA to price

MA.Shift.MS=8; // this shift goes into MA's settings, I'm not sure how this is used so don't ask.

MA.Period.MS=5; //MA period - how many bars of MA are used to calculate it. I used 24 but 5 works better.

MA.Mode.MS=0; // mode 0 - simple MA

MA.PriceType.MS=1; // price type 1 - open price

MM.start="----------MoneyManagement---------";

MM="0 for basic lot, 1 for % of deposit";

MMformula=1; //see above for comments

BasicLot=0.1; // if you don't use money management, you'll use this instead.

MMRiskFactor=0.1; //0.1 = 10% of deposit in each position. Only one new position is opened each day.

LotPrecision=2; //easy to guess? how small your lot can be, digits. I have 0.01 lots, if you have 0.1, use 1 instead. For minilots where 1 lot is 10.000 units, use 1. You'll have to increase risk by 10 times I guess.

MinPips.MS=5; //don't use this one.

Used parameters are default. Default parameters seem to be the best, obtained via lots of optimizations and re-testing of optimized parameters on a different time period than optimization.

Two uploaded statements are for 2004-2007 test (3 years) with MM and without MM. I do not have any data for longer periods of time (Alpari data, 3 years, since the middle of 2004 till now)

If you get any ideas on changing the system to make more wins, post here.

If you get any questions, feel free to ask!

If you get bugs or your backtest result differs greatly from mine, I'll investigate this matter. And please don't use on a real account. I didn't forward-test it yet.

PS: Thanks for fixing!

Yuri Predborski
561
Yuri Predborski  

I'll post a picture of how this system works so you could understand it better.

The chart shows GBP/USD with MA. MA parameters are the same as in the EA.

As you can see, this way we can trade only current trend.

If you optimize the EA for other currencies, you will find different values.

Before optimizing anything, please post so I would know who is doing what and noone would be doing the same thing as others.

I need people who will optimize entry hour and MA parameters, as well as TP,SL,TS for other currency pairs. The more the better. Also I'd love if you only optimized it for 2006 and then tested the best results (high profit, small drawdown. I think the best is 5-10% risk and 20% max-relative drawdown) on longer periods of time. Test in 2007 should show not higher drawdown as well as at least 50% profit (I want 3:1 ratio with MM. 3 times profit compared to DD. or more). Test in 2004-2007 should show no more than 1,5 times higher drawdown as well as profit that is at least 200-300%.

Other results may be discarded as useless.

Picture - GBP/USD with MA, which is using parameters of the EA's MA.

Files:
ma_line.gif 33 kb
Yuri Predborski
561
Yuri Predborski  

Another post reserved for further updates

karmizi
23
karmizi  

suggestion

Dear shinigami,

Do u ever consider about time entry. I found it really helps during trading. Around 2200-2300hrs GMT are the most quiet time. This hi n low calculation have been disscussed in thread started by fxoperater and newdigital.

Another observation i make for the last 2 weeks. If possible make the entry not above the hi and lo target but below ( in the channel of hi n lo ) around 15 pips of current price at 2100-2300hrs GMT. TP between 10-15pips but TS can be use as u said when +ve pips or trend appear.

Just my 2 pips suggestion.

p/s: I been waiting for an EA like u did. Thanks.Hope to see some improvement.

Yuri Predborski
561
Yuri Predborski  

I already tried all starting hours from 0 to 23 for the last 3 years and these work the best.

I tried using small TP but that usually ruins the system.

I tried using small, big and other TS's but only 12, 25 and 45 seem to give some result.

But I'll give it a try!

Okie. I'm done testing.

The best result with 19-21 (GMT) with deleting old orders one hour before the start is 946 profit, 368 drawdown (pips).

The best result with 7-8 (GMT) with deleting orders at 21 is 1186 with 275 drawdown (pips, standard lots were used - 0.1).

Besides 19-21 only works for 2007 and doesn't give any good result for 2004-2007 (the best result is with defaults - 5,5k pips; you can get a better result but it will come at a cost of drawdown).

I'll try making an entry inside the channel now.

karmizi
23
karmizi  

exprience from 2 days live trade.

Dear Shinigami,

I`m sorry for not able to help u in the backtesting. I`m quite new in the forex market trading. Still exploring the MT4 program.

Attached are my result for 2 days trading. Looking at the profit, it is good but to early to say that this is a winning system. But perhaps i can share something with u.

I have enter a trade that are too close to previous day high and low price and the trade still floating. Maybe i need to learn how to use buy and sell limit for the pending orders also. Some trade, I close manually because its taking to much time.

I`m putting my pending orders at 0600hrs GMT +8. U can also see the forex market hours monitor. Start at 0600hrs off to work and come home on pc and see few pips taken.

thanks.

Now I remember why the backtesting not so favourable to my time. U can read the post 428 at asian high and asian low here. Because of this i decide to trade within the price channel if the price is near to previous hi and lo.

Yuri Predborski
561
Yuri Predborski  

Yep, thats too early to tell. But I've been trying your proposition (-15 pips shift) and that makes profit as well. It fails on 2004-2006 but works for 2007.

I'll add a second session and see what I can see.

Thanks for sharing your experience! I hope we can use it to make a great trading system!

prasxz
1263
prasxz  

hi

Shinigami:
Hi!

I want to share the system I made recently. You can have the EA too (its time-limited).

The reasons I'm going to share it:

1. I want to make it better.

2. I want to change it so it can trade all currency pairs.

3. I only have 1 PC and can't test everything.

4. Having only 40% wins makes me depressed.

Even though average loss is 22 pips and average win is 65 pips. Max loss is 30 pips. Max win was a huge trending move of 390+ pips. Even though its a great thing, it still makes me depressed. 45% wins for 2006-2007 and almost 40% for 2004-2007. I want to make it better.

Ok. Details on the system.

1.

You calculate the High and Low of previous day (take the highest and lowest point of yesterday, yesterday is from 0:00 yesterday till 0:00 today).

2.

You set 2 breakout orders - place a buystop if the price (bid) is above MA (place buy-stop 2 pips higher than last day's high) and sellstop if the price is below MA (place sell-stop 2 pips lower than last day's low). I place orders from 7:00 to 8:00 GMT. I don't place orders if the price is closer than 3 pips to MA. I use simple moving average and open-prices model. I checked on others and the result was the same for (high+low)/2 and typical (H+L+C)/3 and weightened close (H+L+C+C)/4.

If you have a better trend-following system, please let me know. Under trend-following I mean filtering out the wrong-way trading. Also I noticed that sometimes the price may move from one side of the MA to the other side. If that happens, both buystop and sellstop are opened. But once the buy-stop is triggered and becomes buy, we delete sell-stop. That helps. A very tiny bit.

3.

You delete orders if nothing happened near the end of the day. I delete them at 21 GMT. Please note that DelTime parameter does NOT use GMT shift. Be careful.

4.

If one of the orders gets opened, you trail it till either TP is hit or SL is hit. My TP of 1000 didn't get hit a signle time - the best result was almost 400 pips. I use TrailingStop of 45 and StopLoss of 30 (trailing stop is used as soon as we are in profit, no wait).

Winning ratio is only around 40-45% but the good thing is (as I said) that every win is 3 times more than 1 loss. And you get 3 losses and 2 wins in a row on average which gives this system profit factor of 2 (1.8 without MM, 2.08 with MM).

I tested only with GBP/USD, H1 timeframe. It won't work with any other timeframe. One of my goals is to find parameters or make changes so it makes profit with other currencies. I'm fine with H1 TF.

EA parameters:

GMT.Shift=2; //no need to explain. Can be negative. EST is negative.

Start.MS=7; //starting hour, affected by GMTshift

End.MS=8; //ending hour, affected by GMTshift

Hours.MS=24; //number of hours to find high/low

BuyS.MS=2; //shift up from last high, can be negative. Don't change

SellS.MS=2; //shift down from last low, can be negative. Don't change

DelTime.MS=21; //hour when you want your orders deleted, unaffected by gmt.shift

TP.MS=1000; // takeprofit. target. unreachable dream (I have systems that CAN reach this target but this one won't)

SL.MS=30; // small stoploss

TS.MS=45; // trailing stop. medium one. 85 seemed to work but this is much better. smaller ones decrease profits greatly which increases drawdown

MA.shift.MS=3; // shift goes for shift from MA to price

MA.Shift.MS=8; // this shift goes into MA's settings, I'm not sure how this is used so don't ask.

MA.Period.MS=5; //MA period - how many bars of MA are used to calculate it. I used 24 but 5 works better.

MA.Mode.MS=0; // mode 0 - simple MA

MA.PriceType.MS=1; // price type 1 - open price

MM.start="----------MoneyManagement---------";

MM="0 for basic lot, 1 for % of deposit";

MMformula=1; //see above for comments

BasicLot=0.1; // if you don't use money management, you'll use this instead.

MMRiskFactor=0.1; //0.1 = 10% of deposit in each position. Only one new position is opened each day.

LotPrecision=2; //easy to guess? how small your lot can be, digits. I have 0.01 lots, if you have 0.1, use 1 instead. For minilots where 1 lot is 10.000 units, use 1. You'll have to increase risk by 10 times I guess.

MinPips.MS=5; //don't use this one.

Used parameters are default. Default parameters seem to be the best, obtained via lots of optimizations and re-testing of optimized parameters on a different time period than optimization.

Two uploaded statements are for 2004-2007 test (3 years) with MM and without MM. I do not have any data for longer periods of time (Alpari data, 3 years, since the middle of 2004 till now)

If you get any ideas on changing the system to make more wins, post here.

If you get any questions, feel free to ask!

If you get bugs or your backtest result differs greatly from mine, I'll investigate this matter. And please don't use on a real account. I didn't forward-test it yet.

PS: can anyone fix the name please? last instead of lasd...

Thanks Shinigami for share your system...I'll try manual first

===================

Forex Indicators Collection

Yuri Predborski
561
Yuri Predborski  

Hi. I've added a picture to illustrate how MA works in the EA. See second post.

If you have any questions, feel free to ask!

To add comments, please log in or register