Two-Stroke System (Mechanical Price Action +300 pips per month) | Page 4 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Dreamliners’ MPAS (Mechanical Price Action Strategy) System 486 replies

Percentage vs. Pips per Trade vs. Pips per Day/Hour! 7 replies

MACD_Colored Indicator for 4 Hour Strategy (300+ Pips per Month) 157 replies

my MT4 have a stroke? 2 replies

Attachments: Two-Stroke System (Mechanical Price Action +300 pips per month)
Exit Attachments

Two-Stroke System (Mechanical Price Action +300 pips per month)

  • #61
  • Jul 30, 2007 12:04pm Jul 30, 2007 12:04pm
Hi Friends -

Just a quick update on what I am doing with this system.

I am focusing on testing several things:

1. How does the system perform over a longer period of time without any modification to the trading rules or the method for determining the "trade side"?

I am currently finishing a more detailed walk forward test of the system including 13 major pairs reaching back to 2001. I am not yet finished, but the preliminary results seem to show in every case that the system works as expected without any modification to the current trading rules. I'll discuss this later, but depending on the timeframe used to calculate the trade side, it seems that the returns can vary widely ... though they are still very positive.

2. Without any modification to the current trading rules, is there a more optimal way of dynamically determining the "trade side" that would have performed well over the past 7 years?

If the system can switch trade sides during major trend changes or protracted retracements, the returns can be dramatically improved. I am looking at various uncomplicated ways to do this. Any suggestions would be appreciated. Even without changing the current method of determining the trade side, the system would have performed very well over the past 7 years (01-07), to the tune of hundreds of pips per month. But it is easy to see that switching sides during major retracements or more quickly responding to major trend changes can make a big difference, perhaps even doubling or tripling the returns.

3. Beyond this, I am giving extra attention to the issue of drawdowns in the equity curve. Overall, the system appears to be very profitable, but I believe that the equity curve can be significantly smoothed by diversification and optimization.

As I mentioned before, I am trying to look first at factors other than optimizing the system rules or adding indicators, filters, etc etc etc. Afterwards, if the system seems to be robust over time, then I will see if it can be improved with optimized entry/exit rules, stops, position sizing, etc.

So right now, I am working on smoothing the equity curve by diversification: market diversification, system diversification, and portfolio diversification. I'm concentrating on portfolio diversification, meaning that over time the pairs this system trades will change. Managing a portfolio of pairs with simple rules for switching between pairs could improve the quality of returns and generate a smoother equity curve. I am working on something I call "hypothetical component equity curves" which I think may work. If it doesn't work ... well, I've had plenty of stupid ideas. Anyway, a picture is worth a thousand words, so let me post just a rough cut of what I'm talking about, and then I'll clarify.
  • #62
  • Jul 30, 2007 12:09pm Jul 30, 2007 12:09pm
Here are two equity curves including a handful of major pairs in this system. I hope it comes out clearly in the cropped images, but if you can see it, the drawdowns in the second equity curve are much more shallow than the original curve. As I said, this is just a rough cut on the idea, but I think it may be promising. All the colored lines on the second curve show the component equity curves of the different pairs in the portfolio.

Okay, more on that later when I'm finished testing.
Attached Images
  • #63
  • Jul 30, 2007 12:19pm Jul 30, 2007 12:19pm
Hi Inflection,

That last equity curve looks really sweet! Things seem to be promising...
  • #64
  • Jul 31, 2007 1:49am Jul 31, 2007 1:49am
Not to be a wet blanket, but if you're only going long for EUR/USD and only short for USD/CAD with two position in each and back testing, has anyone considered what would have happened if you just bought/sold and held for the five years?

EUR/USD - 4,050 pips gain, per the spreadsheet
USD/CAD - 5,243 pips gain, per the spreadsheet

If you divide by 57 months (per the spreadsheet), that's 163 per month, or 326 per month since it's two positions ... and you're not continuing to pay the spread.

My point is, everything else is just something to do and has no affect on your profitability. The only thing adding value is that we should have gone long EUR/USD and short USD/CAD in 2002.

You can't necessarily say the trend will continue for another five years (although personally I think so) just based on the last five years.

If I'm wrong and I don't understand the strategy, then I apologize for the negativity and I'm sure someone will let me know.
  • #65
  • Jul 31, 2007 2:04am Jul 31, 2007 2:04am
Quoting broug
Not to be a wet blanket, but if you're only going long for EUR/USD and only short for USD/CAD with two position in each and back testing, has anyone considered what would have happened if you just bought/sold and held for the five years?

EUR/USD - 4,050 pips gain, per the spreadsheet
USD/CAD - 5,243 pips gain, per the spreadsheet

If you divide by 57 months (per the spreadsheet), that's 163 per month, or 326 per month since it's two positions ... and you're not continuing to pay the spread.

My point is, everything else is just something to do and has no affect on your profitability. The only thing adding value is that we should have gone long EUR/USD and short USD/CAD in 2002.

You can't necessarily say the trend will continue for another five years (although personally I think so) just based on the last five years.

If I'm wrong and I don't understand the strategy, then I apologize for the negativity and I'm sure someone will let me know.
You are correct in what you say.

I would have to look back at the charts to be sure, but I suspect your drawdowns would have been quite a bit larger with a buy/sell and hold approach. Your final profit would have been greater.
  • #66
  • Jul 31, 2007 6:21am Jul 31, 2007 6:21am
Hi broug -

The issue you have raised is important and much appreciated. Here are a couple quick responses.

Quoting broug


... we should have gone ... in 2002.

You can't necessarily say the trend will continue ...
And there lies the rub with buy & hold. It is highly dependent on the specific entry point. Pick a good day, near a local minimum, with the trend ahead of you, and you're golden. Pick a bad day, near a local maximum, at the end of a trend, and the difference is significant. I recently saw someone's tagline on FF which sums it up well: "The trend is your friend ... until it bends in the end." Besides, you need some means to indicate when you should get out. We can't just cherry pick days on a historical chart. If the exit strategy is a signal that is less sensitive to shorter term fluctuations, then you will be in the market longer, but then you will suffer larger retracements and a potentially large loss of profits in the final days/weeks/months. If the exit strategy is a signal that is more sensitive to shorter term fluctuations, then you may end up exiting on a retracement rather than a reversal, and you have to judge how to get back into the trend.

Quoting broug

... everything else is just something to do and has no affect on your profitability.

We absolutely must consider whether any active trading approach is better than a longer term buy & hold. We should not just assume that turning over capital more frequently will generate more profits. However, there are in fact active approaches that beat buy and hold approaches. It remains to be seen whether or not this system will hold up. I'll know when I'm finished testing, take a step back, and ask myself exactly the question you have asked, "Is this any better?" That's one of the reasons I've extended my study back to 2000. Many of the major pairs saw wide swings or significant reversals between then and now. If this system is robust, then it wouldn't matter what particular day or price we begin with. It should deliver somewhat consistent returns and be able to handle trend reversals without giving up a big chunk of the profits.

I don't mind wet blankets. They're good for putting out the fires of stupidity. I appreciate your critique, so keep it coming. I'm not interested in being right. I'm interested in making money.
  • #67
  • Jul 31, 2007 8:57am Jul 31, 2007 8:57am
One of the key advantages of entering and leaving positions over the buy-and-hold approach, is that it is possible to use compounding. So, while the number of pips may be comparable, or even better with buy and hold, but the overall monetary gain, which is afterall what we want, will be greater if you make more trades.
  • #68
  • Jul 31, 2007 9:07am Jul 31, 2007 9:07am
Quoting sgrig
One of the key advantages of entering and leaving positions over the buy-and-hold approach, is that it is possible to use compounding. So, while the number of pips may be comparable, or even better with buy and hold, but the overall monetary gain, which is afterall what we want, will be greater if you make more trades.
Maybe. You also have to consider the constant drain of the spreads you are incurring.
  • #69
  • Jul 31, 2007 9:31am Jul 31, 2007 9:31am
Great post inflection and a nice unclutered strategy.....so far however all I am seeing is backtest data ... Is anyone doing a forward test on this strategy yet?
Same Whore .... Different Dress
  • #70
  • Jul 31, 2007 9:34am Jul 31, 2007 9:34am
I am still concerned about picking "the side" of the market to enter. You have the benefit of hindsight in formulating your system and picking the winning side.

There is really no way to preconceive the market's bias for the upcoming year, so any success is as just a flip of the coin.

Have I missed something here? -Thanks
  • #71
  • Edited 9:55am Jul 31, 2007 9:44am | Edited 9:55am
How about using a 14 or 30 day MA.......trade the side of that...if the candles are on top trade long.....if the candles are below the MA trade short.
Also at the risk of being redundant....is anyone forward test this strategy yet?
Same Whore .... Different Dress
  • #72
  • Jul 31, 2007 1:51pm Jul 31, 2007 1:51pm
Quoting billbss
Maybe. You also have to consider the constant drain of the spreads you are incurring.
That's true, so in general it is necessary to strike the right balance - not to overtrade in order not to incur excessive spread costs, and at the same time give enough opportunities to increase the position size for compounding.
  • #73
  • Jul 31, 2007 7:09pm Jul 31, 2007 7:09pm
Quoting Inflection

2. Without any modification to the current trading rules, is there a more optimal way of dynamically determining the "trade side" that would have performed well over the past 7 years?

If the system can switch trade sides during major trend changes or protracted retracements, the returns can be dramatically improved. I am looking at various uncomplicated ways to do this. Any suggestions would be appreciated. Even without changing the current method of determining the trade side, the system would have performed very well over the past 7 years (01-07), to the tune of hundreds of pips per month. But it is easy to see that switching sides during major retracements or more quickly responding to major trend changes can make a big difference, perhaps even doubling or tripling the returns.

why dont you run a test each week or month to check the trade side you should be on and then trade on that side for the whole month, then next month you do the same and over again in the next months. dont know if its a good idea, but its something from the top of my head.
  • #74
  • Edited 8:11pm Jul 31, 2007 7:59pm | Edited 8:11pm
http://img294.imageshack.us/img294/1...enshot1ls6.png

Seems to me the side to trade is pretty obviouse with few exceptions, SMA moving up and price is above the SMA line then trade long.....SMA moving down and price is below SMA line then trade short.....not exactly rocket science
Same Whore .... Different Dress
  • #75
  • Aug 1, 2007 3:17pm Aug 1, 2007 3:17pm
Soooo..... hows the forward testing going?
Same Whore .... Different Dress
  • #76
  • Aug 2, 2007 7:41am Aug 2, 2007 7:41am
Hi johnedoe -

Thanks for your comments.

I really am not trying to make anything complicated. At least, I think that's what I'm doing. I like simple things, less of a rocket to the moon, and more of a bicycle to the grocery store.

Unfortunately, the trade side is not that obvious. If you use a short MA, and really look at the prices with more than a cursory glance, you will see that the prices often vacillate around the MA. Unless you have really good way to manage this, you will lose all your profits from the whipsaws. Everyone who has tried simple MA cross systems has experienced this, which is why you will find almost every MA system posted at FF has more than a few rules, confirmations, and other filters.

As far as forward testing goes, that comes later. First, I have an idea. Then I test that idea on some historical data. Then I expand that to include more historical data. After I'm confident with that, then I forward test in realtime demo mode for one to three months. Then I trade small size. Then I go full steam ahead.

Forward testing takes time. Unless you think that a week or two of "good" results makes a good system, which I do not. Backtesting means you can look at thousands of trades. I would rather spend a few weeks backtesting a couple thousand trades, than base my confidence on 15 days of trading. Don't get me wrong, I believe in forward testing, demo trading, ramping up, etc. It just comes second, not first, in my approach. I prefer to lose money on paper rather than blow up real accounts, however small they may be. I am definitely *not* a fan of the "hey-I-blew-up-my-forex-account-I-guess-it's-just-the-price-of-my-education-now-I'll-try-some-other-system" school of thought.

I'll be back with more soon. It just takes some time to test and validate the trades.
  • #77
  • Aug 2, 2007 8:27am Aug 2, 2007 8:27am
Quoting Inflection
Hi johnedoe -

Thanks for your comments.

I really am not trying to make anything complicated. At least, I think that's what I'm doing. I like simple things, less of a rocket to the moon, and more of a bicycle to the grocery store.

Unfortunately, the trade side is not that obvious. If you use a short MA, and really look at the prices with more than a cursory glance, you will see that the prices often vacillate around the MA. Unless you have really good way to manage this, you will lose all your profits from the whipsaws. Everyone who has tried simple MA cross systems has experienced this, which is why you will find almost every MA system posted at FF has more than a few rules, confirmations, and other filters.

As far as forward testing goes, that comes later. First, I have an idea. Then I test that idea on some historical data. Then I expand that to include more historical data. After I'm confident with that, then I forward test in realtime demo mode for one to three months. Then I trade small size. Then I go full steam ahead.

Forward testing takes time. Unless you think that a week or two of "good" results makes a good system, which I do not. Backtesting means you can look at thousands of trades. I would rather spend a few weeks backtesting a couple thousand trades, than base my confidence on 15 days of trading. Don't get me wrong, I believe in forward testing, demo trading, ramping up, etc. It just comes second, not first, in my approach. I prefer to lose money on paper rather than blow up real accounts, however small they may be. I am definitely *not* a fan of the "hey-I-blew-up-my-forex-account-I-guess-it's-just-the-price-of-my-education-now-I'll-try-some-other-system" school of thought.

I'll be back with more soon. It just takes some time to test and validate the trades.
Thanks for the info. so far so good wish you good fortune and I'll be watching your progress, keep up the good work......
Same Whore .... Different Dress
  • #78
  • Aug 2, 2007 11:45am Aug 2, 2007 11:45am
Quoting Inflection
EURUSD

1. Enter LONG at the open of the day with 2 positions.
2. If %R indicates an UP trend just before the close, hold 1 position.
3. If the day will close lower than the open or prev close, then exit one or both positions.
4. If the day will close higher than the open and prev close, then hold one or both positions.
5. If the prev bar was an EXIT, re-enter at the open of the next bar with 1 or 2 positions.

USDCAD

1. Enter SHORT at the open of the day with 2 positions.
2. If %R indicates a DOWN trend just before the close, hold 1 position.
3. If the day will close higher than the open or prev close, then exit one or both positions.
4. If the day will close lower than the open and prev close, then hold one or both positions.
5. If the prev bar was an EXIT, re-enter at the open of the next bar with 1 or 2 positions.
Hi,
Sorry to be so late to the thread. Anyhow I do not understand what you are doing.
FX is a 24hrs 5 1/2 days per week market - right?
So why exit at the close , to then re-enter at the open (about 1 second later - or less if you are fast).
Or are you basing yoursystem on the New York session open/close , or European session open/close, or the Asian session open/close?

And how do you define a day anyhow - midnight to midnight EST - or GMT or any other time zone?

The only time you are likely to find a difference in the close of one day to the open of the next is over a weekend.

I am well and truly puzzled.

Ian
  • #79
  • Aug 3, 2007 8:58am Aug 3, 2007 8:58am
Hi ian -

Thanks for posting. You have a fair question, or rather two questions.

The "day" can be what you want it to be, though you might get different results with different start times. My studies are done considering the open of the Sydney market as the open of the Forex "day." This is the same time frame FF uses on its home page and that Oanda uses for its FXMarketHours application. Other brokers/platforms use different times.

The other question raises an important issue. Why exit and re-enter if there is no "signal" except the end of the day? The answer is that this system combines two parts ... a daily scalping method and a trend following method. Getting in and out on the daily movements is simply a way to book profits by scalping on the side of statistical probability. There may be much better ways to do this than what I have proposed. In fact, I'm sure there are. But the main idea is to already be in the market (via the daily "HOLD" signal) when a trend engages (as indicated by a Wm%R signal). The advantage to being early in a trend, even by a day or two, is hugely significant. Even if all the daily hocus pocus and paying spreads and so on was a wash, you would still bank a big profit by being early in these trends.

This may in fact be a completely bone-headed way of supporting my broker, but as I've said a hundred times before, it's just an idea that seems to have merit and appears to be proving out the more I study it. I'm still working on it. It's a work in progress.

Hope that helps.

I'll be back with more later.
  • #80
  • Aug 3, 2007 9:34am Aug 3, 2007 9:34am
Quoting tdion
I am still concerned about picking "the side" of the market to enter. You have the benefit of hindsight in formulating your system and picking the winning side.

There is really no way to preconceive the market's bias for the upcoming year, so any success is as just a flip of the coin.

Have I missed something here? -Thanks

Hi tdion -

Thanks for your comment. Well, you've stepped into the stickiest part of this system, and as I've expressed before, the part that I myself am most uncomfortable with ... the method for determining the tradeside.

The problem we all face is the "hard right edge." There is no future. There is only the past. That's all we have to go by. And to say that there is no way to preconceive the market's bias strikes at the heart of all fundamental and technical analysis. But you are right. There is no certainty. Only probabilities.

But that's all I've done. I've gone back in time many years, calculated the statistical frequency (i.e. probability) of up days and down days from prior history, and then used that as a method for moving forward in time. I've tested various historical time frames for the tradeside calculation (e.g. 3/6/12/24/36 months) and looked at the results moving forward through the historical data. Right now, I can only say that it *appears* to work fine, though the results vary from pair to pair with different time frames. I still have to validate the trades manually or hunker down and write a bunch of calculation formulas.

I think it's not hard to prove that if you scalp a small daily range on the side of this calculated probability, you will win most most of the time. The question is how to turn this observation into a high-performing strategy. That's what I'm trying to do.
0 traders viewing now