Trading With Moving Averages | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

MT4 Moving Averages In Their Own Frame? 3 replies

Moving averages in its own indicator window 7 replies

Moving Averages Zone Trading 10 replies

HELP! GMMA (Guppy's Multi-Moving Averages) Trading Mehod 3 replies

Shifting Moving Averages on Daily Charts 2 replies

Attachments: Trading With Moving Averages
Exit Attachments

Trading With Moving Averages

  • #41
  • Jan 14, 2011 12:39pm Jan 14, 2011 12:39pm
Quoting Davidee
I think you miss understand, it's not 1.5r - wins aren't 1.5 times the size of losers. Wins are (on average) smaller than losers, but the very high number of them makes the total number of winning pips 1.5 times the total number of pips lost.

A margin call isn't likely since I'm using leverage of around 2:1.

Let's just follow it and see what happens. This isn't how I trade usually, I'm just forward testing it with microlots because that's the next step.
I see, that sounds a bit more reasonable, but youve got it the wrong way round in the OP, "Pips won to pips lost ratio 1.50 to 1."
Still even the other way round thats more than respectable. I guess its worth a punt using 2:1 gearing,
  • #42
  • Jan 14, 2011 1:51pm Jan 14, 2011 1:51pm
Well perhaps I should have explained it better.

But please try the testing yourself and let me know if you get a different result.

You only need daily OHLC data.

It's always in the market and the open orders/take profit levels are adjusted daily.

 

  1. Go long when the price hits (Yesterday's Low+The Day Before Yesterday's Low)/2.
  2. Close the long and open a short when the price hits (Yesterday's High+The Day Before Yesterday's High)/2.
  3. Leave the trade on and adjust the open order/take profit order daily until it's hit.

Do that and you'll tend to win about twice as often as you lose, losers will be slightly larger (on average) than winners.

At least that's the results of the back-testing, now the forward testing with microlots has begun and the trades will be posted as they happen on this thread. I invite you to check my numbers and call me out on it if you sniff any BS. One of two things will happen -

1. It won't work in which case I'll lose a little bit of cash (no big deal) and know it's not a good system.
2. It'll work and people will be free to take the idea and adapt it to suit them and I wish them luck with it.

  • #43
  • Jan 15, 2011 2:42pm Jan 15, 2011 2:42pm
OK well, it's the weekend now and the take profit and go short order wasn't hit yesterday so I now have to amend it with (Thursday's High + Friday's High)/2.
The figures I have on metatrader are (0.97633+0.96867)/2 so I have to close the long and open a short if 0.97250 is hit on Monday.
  • #44
  • Jan 17, 2011 9:12am Jan 17, 2011 9:12am
The only data available for the default Amibroker is DJI and a bunch of US stocks so I decided to do a quick test on DJI.

The backtesting was done from January 2005 through March 2010. Position size 100% of equity - fully invested in every position, no leverage. LONG only.

Original system:

CAR = 21.51%
MaxDD = -9.07%
CAR/MaxDD = 2.37
Profit Factor = 3.91

Winners = 82.58 %
Avg. Profit = 2036.26

Losers = 17.42 %
Avg. Loss = -2471.74

Attached Image


Original system, with trend filter removed:

CAR = 38.82 %
MaxDD = -22.11 %
CAR/MaxDD = 1.76
Profit Factor = 3.03

Winners = 80.48 %
Avg. Profit = 3802.73

Losers = 19.52 %
Avg. Loss = -5181.64

Attached Image


While the Anual return increased so did the MaxDD. CAR/MaxDD also decreased, meaning more overall risk than the original system. Profit Factor also decreased.

The big losses taken by the "no trend filter" system were taken during 2008 when I assume it tried to catch a lot of falling knives. Compare the equity chart with the one from the original system and notice that the original system chart was flat during parts of 2008 - because of the trend filter.

A good compromise is to lower the trend filter, to a value good enough to keep the system out during strong downtrending periods.
I tested with a trend filter of 20 days. The results:

CAR = 27.8%
MaxDD = -10.28%
CAR/MaxDD = 2.71
Profit Factor = 4.73

Winners = 85.38%
Avg. Profit = 2507.08

Losers = 17.42 %
Avg. Loss = -3093.26

Much better, compared to the original system. Notice the increase in CAR by 6 percent while MaxDD increase of only 1 percent. PF also increased to 4.73. Better readings overall.

Attached Image


In conclusion I believe David's idea is a valid one and there seems there really is a nice edge around this system.
It doesn't work on everything, I also tested it on a few random US stocks and the results were less than satisfacatory.

I don't yet have EURUSD data but when I have it I'll come back with more testing.
Trading = a mirror to your human flaws. Fix them or be fixed.
  • #45
  • Jan 17, 2011 9:27am Jan 17, 2011 9:27am
Quoting soso
The only data available for the default Amibroker is DJI and a bunch of US stocks so I decided to do a quick test on DJI.

The backtesting was done from January 2005 through March 2010. Position size 100% of equity - fully invested in every position, no leverage. LONG only.

Original system:

CAR = 21.51%
MaxDD = -9.07%
CAR/MaxDD = 2.37
Profit Factor = 3.91

Winners = 82.58 %
Avg. Profit = 2036.26

Losers = 17.42 %
Avg. Loss = -2471.74

Attachment

Original system, with trend filter removed:

CAR = 38.82 %
MaxDD = -22.11 %
CAR/MaxDD...
work. Thanks for this great contribution.
Vincit qui se vincit.
  • #46
  • Jan 17, 2011 1:33pm Jan 17, 2011 1:33pm
Quoting soso
The only data available for the default Amibroker is DJI and a bunch of US stocks so I decided to do a quick test on DJI.

The backtesting was done from January 2005 through March 2010. Position size 100% of equity - fully invested in every position, no leverage. LONG only.

Original system:

CAR = 21.51%
MaxDD = -9.07%
CAR/MaxDD = 2.37
Profit Factor = 3.91

Winners = 82.58 %
Avg. Profit = 2036.26

Losers = 17.42 %
Avg. Loss = -2471.74

Attachment

Original system, with trend filter removed:

CAR = 38.82 %
MaxDD = -22.11 %
CAR/MaxDD...
You can't trust the High's and Low's on DJI data - I think the index is still calculated similarly to how it was in 1851! I think the high is each stock's high / 30, and the low is each stock's low / 30. Since, in reality, the individual components don't all hit their highs and lows at the same time the highs and lows aren't always real; i.e. what it's telling you the daily highs and lows are weren't really 'hit' so in reality you could never really have bought or sold at these levels. Try it on the S&P 500 and compare the difference.
  • #47
  • Edited 3:22pm Jan 17, 2011 3:11pm | Edited 3:22pm
How about the QQQQ? Have you looked at it?
Seems it would be a good candidate.
  • #48
  • Jan 17, 2011 3:16pm Jan 17, 2011 3:16pm
OK, well the high wasn't hit today and it's nearly tomorrow (or at least it is here) so I have to adjust my profit take/open short order.

Friday's high was 0.96867, Monday's high was 0.96757. So that's (0.96867+0.96757)/2 = 0.96812.

Order adjusted to take profit and then go short at 0.96812.

Wish me luck...
  • #49
  • Jan 17, 2011 3:19pm Jan 17, 2011 3:19pm
Quoting Davidee
You can't trust the High's and Low's on DJI data - I think the index is still calculated similarly to how it was in 1851!...
Don't have SP data, yet. Anyway there's clearly an error in my backtesting code. Will get back when sorted out...
Trading = a mirror to your human flaws. Fix them or be fixed.
  • #50
  • Jan 17, 2011 3:26pm Jan 17, 2011 3:26pm
Quoting Davidee
OK, well the high wasn't hit today and it's nearly tomorrow (or at least it is here) so I have to adjust my profit take/open short order.

Friday's high was 0.96867, Monday's high was 0.96757. So that's (0.96867+0.96757)/2 = 0.96812.

Order adjusted to take profit and then go short at 0.96812.

Wish me luck...
Guess I missed something... Are you currently in a trade? and could you post a chart with it? Thanks.
  • #51
  • Jan 17, 2011 3:30pm Jan 17, 2011 3:30pm
Quoting soso
Don't have SP data, yet. Anyway there's clearly an error in my backtesting code. Will get back when sorted out...
I don't think there necessarily something wrong with your code. It's just the highs and lows on the DJI aren't real. Just test any other index. I don't imagine indices will work too well though.
  • #52
  • Jan 17, 2011 3:31pm Jan 17, 2011 3:31pm
Quoting soso
Don't have SP data, yet. Anyway there's clearly an error in my backtesting code. Will get back when sorted out...
OK, I'm not sure how David trades, but the way I backtested is that for LONG positions I have a BUY STOP order at previous days average lows as he described. If the entry is triggered then I have a SELL STOP for closing at previous days verage highs. Opening and closing can take place in the same day.

Anyhow I need more data to backtest. Will have.

tradestar1, I don't have QQQQ data yet.
Trading = a mirror to your human flaws. Fix them or be fixed.
  • #53
  • Jan 17, 2011 3:32pm Jan 17, 2011 3:32pm
Quoting Davidee
I don't think there necessarily something wrong with your code. It's just the highs and lows on the DJI aren't real. Just test any other index. I don't imagine indices will work too well though.
You may be very right David. Individual liquid stocks that I have don't offer good results.
Trading = a mirror to your human flaws. Fix them or be fixed.
  • #54
  • Jan 17, 2011 3:40pm Jan 17, 2011 3:40pm
Quoting tradestar1
Guess I missed something... Are you currently in a trade? and could you post a chart with it? Thanks.
Yep, I've been trading the system live on this thread with microlots.

If you go back and read the thread you'll see I've been doing.
Attached Image
  • #55
  • Jan 17, 2011 6:06pm Jan 17, 2011 6:06pm
Quoting Davidee
Yep, I've been trading the system live on this thread with microlots.

If you go back and read the thread you'll see I've been doing.
OK here is my USD/CHF chart. It appears to be a bit different than yours. This is part of the problem with FX ...... the broker and chart times, I trade at Oanda and their new daily candle forms at 8:00 PM PST, then depending on how you have the chart set up.....Hiding weekend data or not and if so what time you have set for end of week and start of week will and can cause quite a variance on the chart.

Someplace I am missing something in the translation.
Attached Image (click to enlarge)
Click to Enlarge Name: Screen shot 2011-01-17 at 3.57.13 PM.png Size: 27 KB
  • #56
  • Jan 18, 2011 3:48am Jan 18, 2011 3:48am
Well today's (Tuesday's) target is close the long and go short at (Monday's High + Friday's High)/2. I make that target figure 0.96812. What values do you get off your chart? It must be pretty similar, right?

I got the data from Metatrader 4.
  • #57
  • Edited 9:12am Jan 18, 2011 8:25am | Edited 9:12am
I have .96820 since I am still using the 4 decimal point view.
So now we are looking to go short at that price correct?
Also are you using the 80 day filter?
  • #58
  • Jan 18, 2011 11:19am Jan 18, 2011 11:19am
Quoting tradestar1
I have .96820 since I am still using the 4 decimal point view.
So now we are looking to go short at that price correct?
Also are you using the 80 day filter?
OK, so we have pretty similar numbers then.

You are correct, I've set an order to close the long and open a short if the price hits 0.96812 today.

But if that level is not hit today, then I will amend the order using (today's high + yesterday's high)/2 and hope that the new amended level is hit tomorrow.

You can only do one trade per currency pair per day though. If the order is hit and the long is closed and the short is taken today, then it won't be until tonight/tomorrow that I set a take profit and open a new trade order. Does that make sense?


I am trading this live on this thread using microlots because I've done a lot of back-testing and I think I need to do some forward trading with real money whilst keeping a written record of the results to move forward with the concept.
  • #59
  • Edited 12:29pm Jan 18, 2011 11:24am | Edited 12:29pm
Quoting Davidee
OK, so we have pretty similar numbers then.

You are correct, I've set an order to close the long and open a short if the price hits 0.96812 today.

But if that level is not hit today, then I will amend the order using (today's high + yesterday's high)/2 and hope that the new amended level is hit tomorrow.

You can only do one trade per currency pair per day though. If the order is hit and the long is closed and the short is taken today, then it won't be until tonight/tomorrow that I set a take profit and open a new trade order....
OK, thank you. It's starting to gel now...

Also I have been watching the EUR/USD...
At this time I have 1.3381 for a sell .... Does that sound about right?
  • #60
  • Jan 18, 2011 3:50pm Jan 18, 2011 3:50pm
I'm still trading USD/CHF, I'm adjusting the stop loss (0.96757+0.96557)/2=0.96657.

For the EUR/USD I have (1.34659+1.33824)/2=Sell @ 1.342415
0 traders viewing now