Forex Factory (https://www.forexfactory.com/forum.php)
-   Commercial Content (https://www.forexfactory.com/forumdisplay.php?f=158)
-   -   My EA Gift to Traders (https://www.forexfactory.com/showthread.php?t=520258)

nkimandi May 29, 2015 7:00pm | Post# 1441

{quote} One method I have considered is to introduce another variable which sets the upper limit for risk. You know that I have programmatically set 20%, but this figure could also be optimized if it was set as a variable. Starting a .5% with a Stop Loss Multiplier of 1.5, there are 9 progressions to reach 10% equity.
Hi Dave C..most brokers automatically hedge positions if it margin falls to less than 80%..some 100%..another option you could look at other than capping
is for the user to set the limit..say 200% or 300% and so forth(whatever inputted figure)..once it dips below the user set figure then it doesn't initiate any more trades no matterl..i have seen this feature in some EAs..it also prevents inadvertent auto hedging thanks to careless or greenhorn trader error

Dave_C May 29, 2015 7:08pm | Post# 1442

{quote} Am getting your logic and the best way to answer it would be for you to pull up the Swiss franc(USDCHF) chart and scroll all the way to January 15th 2015..zoom in further to H1..now put your EA against it with your settings intact that you are using..would any EA have been able to save you the trader?what would have saved you?wouldnt it have been simulating or imagining a black swan event such as that beforehand?how would you have done it?..on that day no Stop loss could have saved you..no trailing SL..no exiting on opposite signal..nada!.....
Totally agree and in recent times the Jan 15 event is the benchmark for an EA's resilience, My primary concern is having a good look at what happened to the CCI Vol last night where nearly all the profit was given back. I do think I may have to invoke a trailing stop loss on equity such as what was demonstrated with the CloseTrades EA to protect gains of a basket of EA's. In other words, having got to 35,000+ from 10,000, there should be a mechanism such as we have on individual trades.

I was a co-incidence that it happened on the last day of the month when I advocate a new set of EA's but nevertheless it did happen before the end of the month and is something that I will really appreciate your thoughts. This is the myfxbook address for the CCI Vol.

Dave_C May 29, 2015 7:16pm | Post# 1443

1 Attachment(s)
Enclosed is the CloseTrades EA which monitors account equity and automatically closes all trades on an Account when the equity either attains or falls below preset levels. It was originally used when we looked at the concept of daily targets and whilst it did work, at that stage, I thought that relying on money management at EA level would be a better way to go but todays effort re-opens this idea.

It is a simple EA which simply closes all trades and sets up a dummy but limit order which prevents any TradersGift EA within the account from trading until the dummy order is removed.
CloseTrades.mq4

nkimandi May 29, 2015 8:02pm | Post# 1444

{quote} Totally agree and in recent times the Jan 15 event is the benchmark for an EA's resilience, My primary concern is having a good look at what happened to the CCI Vol last night where nearly all the profit was given back. I do think I may have to invoke a trailing stop loss on equity such as what was demonstrated with the CloseTrades EA to protect gains of a basket of EA's. In other words, having got to 35,000+ from 10,000, there should be a mechanism such as we have on individual trades. I was a co-incidence that it happened on the last day...
Cool..i just have had a quick and dirty look at the loss contributors via the link..i will try and take the losses compare with the entry time stamps,duration of holdings in each case and exit time stamp and set up the charts and manually 'think' like the EA with its settings at that time..a trailing stop as a take profit is one good option or have a basket take profit option set as a percentage of total equity..say basket total equity due to open profit has grown by 5%,10%,50% etc etc then it automatically closes all open trades and begins again..the user can input his/her targetted percentage growth in the EA for a basket of entries and once it reaches or touches the %age profit target then all open transactions(both in profit and loss) get cut/liquidated.

its 3am saturday back here..let me get some shut eye and freshen my brain and tackle your request first thing when i wake up..i hope to have further useful suggestions on this..until then a basket trailing stop or basket take profits can be viewed as primary frontline measures..i do agree with you on this..

cheers

Dave_C May 29, 2015 8:03pm | Post# 1445

{quote} Cool..i just have had a quick and dirty look at the loss contributors via the link..i will try and take the losses compare with the entry time stamps,duration of holdings in each case and exit time stamp and set up the charts and manually 'think' like the EA with its settings at that time..a trailing stop as a take profit is one good option or have a basket take profit option set as a percentage of total equity..say basket total equity due to open profit has grown by 5%,10%,50% etc etc then it automatically closes all open trades and begins...
Your contribution is greatly appreciated. If you could kindly send me an email, I will respond in detail.

Dave_C May 29, 2015 8:07pm | Post# 1446

2015 Data files (and Last 3 months Data files) update now available at website. Use your existing links from the welcome email or go to Post #1.

drofwarc May 29, 2015 9:04pm | Post# 1447

{quote} Totally agree and in recent times the Jan 15 event is the benchmark for an EA's resilience, My primary concern is having a good look at what happened to the CCI Vol last night where nearly all the profit was given back. I do think I may have to invoke a trailing stop loss on equity such as what was demonstrated with the CloseTrades EA to protect gains of a basket of EA's. In other words, having got to 35,000+ from 10,000, there should be a mechanism such as we have on individual trades. I was a co-incidence that it happened on the last day...
Hi Dave,

Several years ago, I conducted extensive hand backtesting of some simple breakout systems on the H1, H4 and daily timeframes. I found that, quite consistently, the worst losses came at the end of the month/beginning of a new month. Also, if the month ended on a Friday, the probability of losses increased even more. The problem was the increased occurrence of whipsaws caused by profit taking, monthly account settling, stophunts, etc. Further, take a look how often price reverses from one month to the next. Often, the reversal gets started a few days before the end of the month, which one might surmise is the result of the Big Boys surprising the market with their plans. In any case, I found in general that my backtesting results improved when I didn't open new trades on the last day of the month/first day of the new month. Unfortunately, I never got around to looking at what would happen if I closed existing trades before the end of the month. I suspect this would also have had a positive impact on my backtesting profits, but I can't say for sure.

I don't know that I'm telling you anything you don't already know. However, I can say that it doesn't surprise me to hear that a mechanical system would experience higher than normal losses at the end of the month.

I'm throwing this information out there for the active minds on this thread to consider. Perhaps it's something we could collectively monitor and address in the EA moving forward.

Kind regards,
drofwarc

Dave_C May 29, 2015 9:34pm | Post# 1448

{quote} Hi Dave, Several years ago, I conducted extensive hand backtesting of some simple breakout systems on the H1, H4 and daily timeframes. I found that, quite consistently, the worst losses came at the end of the month/beginning of a new month. Also, if the month ended on a Friday, the probability of losses increased even more. The problem was the increased occurrence of whipsaws caused by profit taking, monthly account settling, stophunts, etc. Further, take a look how often price reverses from one month to the next. Often, the reversal gets...
That's an interesting observation. At the moment I'm looking at all the EA's from the short example back test to see what I can get from them.

At any rate even though I will email out the new ones over the weekend, I wont be placing them until Tuesday and I was considering of an amendment to the CloseTrades EA to include an Account Equity Trailing Stop Loss as part of the email. The good thing about yesterday's trading is that it has got everybody's thinking cap on. I will most probably drop EA's on the shorter timeframes (except the Continuous Straddle) and may recommend restarting or even shutting down some EA's at a combined equity Trailing Stop Loss point. The giving back of $24000 profit on CCI Vol was completely avoidable.

mozhixing May 30, 2015 1:10am | Post# 1449

2015 Data files (and Last 3 months Data files) update now available at website. Use your existing links from the welcome email or go to Post #1.
Hi, Dave,

I have a question. I downloaded data using Tick Data Downloader and export them to csv file and then I used Strategy Optimizer to test CCI VOL strategy using the data downloaded, but every time when the progress was at 99%, there was a error message and the optimizer stopped there. I could chose to continue but the results didn't come up. I checked my data format, it has the same format as yours. What is the reason? Thank you very much.

Dave_C May 30, 2015 5:20am | Post# 1450

{quote} Hi, Dave, I have a question. I downloaded data using Tick Data Downloader and export them to csv file and then I used Strategy Optimizer to test CCI VOL strategy using the data downloaded, but every time when the progress was at 99%, there was a error message and the optimizer stopped there. I could chose to continue but the results didn't come up. I checked my data format, it has the same format as yours. What is the reason? Thank you very much.
Can you post an example problem data file in zip format and I will examine it.

Dave_C May 30, 2015 5:33am | Post# 1451

I'm sorry I haven't posted a daily topic today but I will double up tomorrow. A social day!

Shawn29 May 30, 2015 6:59am | Post# 1452

My CCI+Vol didn't really take any beating, though it never did get the gains Dave's did. I was using my own optimization though so I'm sure the settings are different. Also Dave I like the idea of being able to adjust the risk % from 20% to lower as the margin talk that's been going on has been a concern of mine using smaller amounts like $500 or $1000. Though using only 7 pairs I haven't had any problems yet, some that I've tested where I've done more then that I've run in to insufficient funds notices.

Edit: Was just thinking about the margin stuff, and something that also came to my mind is the strategy used has a pretty high impact on this since depending on what you're using it might not take all that many trades. An EMA and MACD combo might only take a few trades a month per pair so the chances of having multiple trades or back to back losses opening around the same time is less likely.

mozhixing, try renaming the file exactly like his ex: EURUSD_M1_yyyymmdd_yyyymmdd I had the same problem and that fixed it for me.

demalichcha May 30, 2015 1:04pm | Post# 1453

I like if I can give a contribution to the thread in betterment of the improvement of EA. But lack of knowledge in the subject and lack of knowledge in English prevent me from giving a significant contribution.

Anyway, I believe if the EA is improving for basket trading method, we should have an overall strategy to face the worst scenario without getting a margin call. Definitely, that kind of strategy will slower the gains, but sustainable. (I don't know, whether trading in lower time frames will help to maximize the gains). The overall formula would consist with equity, lot size and number of currency pairs.

demalichcha May 30, 2015 1:08pm | Post# 1454

My CCI+Vol didn't really take any beating, though it never did get the gains Dave's did. I was using my own optimization though so I'm sure the settings are different. Also Dave I like the idea of being able to adjust the risk % from 20% to lower as the margin talk that's been going on has been a concern of mine using smaller amounts like $500 or $1000. Though using only 7 pairs I haven't had any problems yet, some that I've tested where I've done more then that I've run in to insufficient funds notices. mozhixing, try renaming the file exactly...
I think 40% per month is a great success.

Dave_C May 30, 2015 6:36pm | Post# 1455

Friday was a bad day for nearly all of the strategies and next month (June) I am going to try an experiment and make a general post of 3 systems all using the CCI Vol Strategy optimized for the last 3 months. (one using a 10% limit on equity with resultant different variable values)

One will do all the wrong things

That is -
Start tomorrow - Monday and first trading day of the month
Finish on Tue 30th June with no trailing stop loss protection on equity.

The second system will start on Tuesday 2nd and finish on Monday 29th June and employ a Trailing Stop Loss protection of $5000 ($10,000 opening balance) and a limit of 20% on equity

The third system will start on Tuesday 2nd and finish on Monday 29th June and employ a Trailing Stop Loss protection of $3000 ($10,000 opening balance) and a limit of 10% on equity

During the course of the month I will highlight the 5,000 profit level (50% return) but allow the profits to run.

I'd be interested in your thoughts because I could alter the concept if there is a better idea.

zakis May 30, 2015 7:15pm | Post# 1456

Demo is demo. Lets try with real money

Dave_C May 31, 2015 12:05am | Post# 1457

Todays Discussions: (Refer Post #1301)

This is the beginning of the daily topics on the 12 Strategies and 9 Filters.

Common to all strategies is the ability to specify when the condition will be fulfilled. The default setting is a cross on the last completed bar which gives the buy or sell signal, but the signal can be satisfied within the last n bars (the end of the range defaults to 5 but can be extended if required). If this value is set to, for example 5 , any cross within the last 5 bars will provide the signal. If n=0 then the current bar will be used for the calculation and is not recommended.

If multiple strategies are being used, ensure that the look back bars of at least one indicator is expanded out beyond 1, because simultaneous crosses really restrict the number of trades taking place.

The concept is that we are looking to optimize the variables which determine the buy or sell signal over a particular time period in the past. I do recommend that you only look at the immediate previous 3 months as the behaviour of the market is constantly changing and any information you draw from 1 year ago may be no longer relevant. You just need to look at all the EA's that have been marketed and the majority will not stand the test of time. By default only the variables used in the calculation are selected for optimization when a particular strategy is selected.

The look back periods are not selected by default for optimization and the time should normally be left at 0 which is the currently selected timeframe. I may in the future give you the ability to select multiple time frames for the same indicator. If you need a different timeframe than the selected timeframe for the overall optimization, you can nominate another timeframe in minutes. Beware that optimizations only indicate what has happened in the past (hence the term "curvefitting"), and forward testing is recommended for confirmation

EMA Cross
EMA Cross is the cross of a short exponential moving average and a long exponential moving average. By itself, it is not very effective, since EMA cross is a lagging indicator (it is representative of past price action) but can be used in conjunction with other indicators as confirmation of other signals.

Price EMA Event
The terminology was changed from cross to incorporate signal confirmation by referencing the position of price in relation to a exponential moving average. Some strategy considerations reflect for example, the price being above of below the 200EMA and the position selection accommodates this type of consideration and can also be optimized for value of the EMA. The default option is to provide a signal on the basis of a price and EMA cross.

There is also provision to shift the value EMA signal forward to reflect a more real time situation rather than a past situation. The most common shift value is 2.

I received a request very early in the thread to include the Heiken Ashi calculation in the value of the close price which is calculated on the basis of (Open + High + Low) / 3.

As with EMA Cross, Price/EMA Cross is not really a profitable strategy on its own and Price/EMA Position is normally used in conjunction with other indicators.

Dave_C May 31, 2015 6:30pm | Post# 1458

Friday was a bad day for nearly all of the strategies and next month (June) I am going to try an experiment and make a general post of 3 systems all using the CCI Vol Strategy optimized for the last 3 months. (one using a 10% limit on equity with resultant different variable values) One will do all the wrong things That is - Start tomorrow - Monday and first trading day of the month Finish on Tue 30th June with no trailing stop loss protection on equity. The second system will start on Tuesday 2nd and finish on Monday 29th June and employ a Trailing...
Started first CCI_Vol Strategy.
Exactly the same second CCI_Vol will be started tomorrow with $5000 Trailing Stop Loss
Third CCI_Vol tomorrow with 10% Wager limit and $3000 Trailing Stop Loss

All will be at the top of each page and at the following link.
http://tradersgift.com/Established_Strategies.html

demalichcha May 31, 2015 8:53pm | Post# 1459

{quote} Started first CCI_Vol Strategy. Exactly the same second CCI_Vol will be started tomorrow with $5000 Trailing Stop Loss Third CCI_Vol tomorrow with 10% Wager limit and $3000 Trailing Stop Loss All will be at the top of each page and at the following link. http://tradersgift.com/Established_Strategies.html...
Thank you very much Dave. I see link for the set files are also there. You don't have to worry about set file requests this time

Dave_C Jun 1, 2015 12:01am | Post# 1460

Todays Discussion: Refer to Post #1457 for discussion on Lookback Bars and Timeframe

Relative Strength Index
RSI is a measurement of the Exponential Moving Average of the Average Gain / Average Loss (expressed as a percentage) over of a particular number of bars.

For example: If in a 14 bar period there were all gains, then the RSI(14) would be 100 and with all losses the value would be 0. This type of indicator is known as a momentum indicator and is a good measure of the strength of the market. Of particular importance is the direction of movement and it is generally accepted that the indicator moving down from a high level indicates a selling trend and moving up from a low level indicates a buying trend. A reading over 50 indicates buying pressure and a reading below indicate selling pressure. Peak and troughs moving in the opposite direction to price is known as divergence and can indicate imminent price reversal.

When carrying out optimizations the RSI periods, Bottom Level and Top Level are variable and the RSI passing through the levels form the basis for Buy and Sell signals and this indicator is recommended as a reliable indicator for buying and seeling pressure particularly over longer timeframes.


© Forex Factory