Time frame charts are for the 95% losers? | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Winners scale winners, losers scale losers 12 replies

Time Frame Vs Personality, Which to choose as Main Time Frame? Who Are You? 10 replies

If losers average losers, then I am a loser 10 replies

Time frame and Enter/Exit, Time frame and S/L, T/P relationships. 91 replies

Question on Lower Time Frame Indicators Vs. Longer Time Frame 7 replies

Attachments: Time frame charts are for the 95% losers?
Exit Attachments

Time frame charts are for the 95% losers?

  • #1
  • First Post: Sep 6, 2010 7:22am Sep 6, 2010 7:22am
Heres something interesting. This is one of my simple "stop and reverse" strategy.

In the normal timeframe charts (seems most traders are using), you get crap results and when not using "normal" charts you get way better results.

I dont think this is curve fitted, It have ran it on a same values on differend pairs, and it seems to have positive profit curve on all of them. It seems to be a winning strategy at least in the simulator(Ninjatrader). Also slippage of 4 points are calculated on every trade to make it more realistic. Comissions are not calculated but that doesnt make that big of a deal. Each trade is one mini lot. Currently forward testing this on a live demo account.

Guess which one of these backtest results are based on the charts that 99% or traders are using?
Attached Image(s) (click to enlarge)
Click to Enlarge Name: resul1.png Size: 81 KB
Click to Enlarge Name: result2.png Size: 75 KB
  • #2
  • Sep 6, 2010 7:30am Sep 6, 2010 7:30am
great go trade it and come back when your a millionaire

What timeframe are you talking about anyway?
Time hides Nothing
  • #3
  • Sep 6, 2010 7:33am Sep 6, 2010 7:33am
what do you mean by "normal" charts? Renko, range bars, ticks?
  • #4
  • Sep 6, 2010 7:55am Sep 6, 2010 7:55am
  •  Ideasmiths
  • | Joined Mar 2009 | Status: 一将功成万骨骷 Forex war! Win pips or die! | 30 Posts
Interesting, can you please elaborate more on the strategy? thank you
  • #5
  • Sep 6, 2010 8:08am Sep 6, 2010 8:08am
Quoting Dezil
what do you mean by "normal" charts? Renko, range bars, ticks?
Ding ding! we have the winner! Yes i was referring to a non time based bars. I'm using rangebars(RangeNoGap) in this strategy.
Important is to cut your losers short, and ride the trend. Hit rate is about 40% but winners are about 2x bigger in avarage(some way larger) than losers. At least in the backtest.
  • #6
  • Sep 6, 2010 8:09am Sep 6, 2010 8:09am
um... so you ran the other test on which TF? and this proves what exactly?

I dont know hw anyone could look at one TF and trade it profitably myself
Time hides Nothing
  • #7
  • Sep 6, 2010 8:11am Sep 6, 2010 8:11am
Quoting CindyXXXX
um... so you ran the other test on which TF? and this proves what exactly?

I dont know hw anyone could look at one TF and trade it profitably myself
The timeframe version was using 15 min, but that was just an example. I picked that timeframe because, there was some stop and reverse strategy on a newbie section. Posted the same post there to give it as pointer
  • #8
  • Sep 6, 2010 8:14am Sep 6, 2010 8:14am
Quoting CindyXXXX
um... so you ran the other test on which TF? and this proves what exactly?

I dont know hw anyone could look at one TF and trade it profitably myself
You are still locked in thinking in "time". Time has no importance in profits, only price matters!
  • #9
  • Sep 6, 2010 8:33am Sep 6, 2010 8:33am
Can you tell as something about yours strategy?

Cheers
  • #10
  • Sep 6, 2010 8:42am Sep 6, 2010 8:42am
And the search for the elusive HG continues......
  • #11
  • Sep 6, 2010 9:03am Sep 6, 2010 9:03am
johto: Did you test range bars, but how much points of range you tested? 10 points, 20 points?
  • #12
  • Edited 11:33am Sep 6, 2010 11:15am | Edited 11:33am
Quoting wskeal
johto: Did you test range bars, but how much points of range you tested? 10 points, 20 points?
The range depends how frequent you want the system to trade and whats the risk apetite. Range much dictates your "stop"(stop and reverse) size. That again depends how big lots you can trade. Too small range, and you will get hit by the ranging chop(but not as bad as time based charts). I also have narrowed the time to trade to the london and us sessions(more movement). I got somewhat better results when having a "session" to trade versus 24h. So in that sense, it is "curvefitted" using TIME as factor. It did also produce positive curve even with trading 24/5.

Looking at backtests, there are times it can have many losing trades chops, even a 20 in a row, but they should be smaller than the winners. Thats why its important not to cut the fewer winners, but to let them be bigger in avarage(positive curve). Currently i am testing 4pip range. Anyhow, i'm testing this now real time using a demo so i can have some idea how much slippage during a news spike etc...

I havent been able to test it longer than begining of this year, and even then, i dont know how reliable data it is. So take it with grain of salt :o
Here are some backtest pictures:
Attached Image(s) (click to enlarge)
Click to Enlarge Name: summary.png Size: 65 KB
Click to Enlarge Name: cumu.png Size: 43 KB
Click to Enlarge Name: dd.png Size: 62 KB
Click to Enlarge Name: daily.png Size: 52 KB
Click to Enlarge Name: weekly.png Size: 52 KB
Click to Enlarge Name: monthly.png Size: 52 KB
Attached Image
  • #13
  • Sep 6, 2010 11:24am Sep 6, 2010 11:24am
Quoting johto
I dont think this is curve fitted, It have ran it on a same values on differend pairs, and it seems to have positive profit curve on all of them.
Just do a 10-year portfolio test and you'll know if it's curve fitted or not.

If you need help, I can do it for you but I seriously doubt the results will look like the curve you have (unless of course you really have found something very significant). Easiest way to test if the simulator is correct is to run one/couple of days through simulator and the same period live and check if the results match 100%.

t. fellow finn.
  • #14
  • Sep 6, 2010 11:29am Sep 6, 2010 11:29am
Quoting johto
You are still locked in thinking in "time". Time has no importance in profits, only price matters!
My point exactly

As I said I watch all TF and nothing is important to me apart from price
Time hides Nothing
  • #15
  • Sep 6, 2010 11:31am Sep 6, 2010 11:31am
Quoting CindyXXXX
My point exactly

As I said I watch all TF and nothing is important to me apart from price
Do you really mean time in itself does not matter? If you do, then check historical volatility and high/low during the day for different weekdays and hours and you might be surprised.

Time really does matter, only not in the sense many newbies are using it.
  • #16
  • Sep 6, 2010 11:34am Sep 6, 2010 11:34am
Quoting mikkom
Just do a 10-year portfolio test and you'll know if it's curve fitted or not.

If you need help, I can do it for you but I seriously doubt the results will look like the curve you have (unless of course you really have found something very significant). Easiest way to test if the simulator is correct is to run one/couple of days through simulator and the same period live and check if the results match 100%.

t. fellow finn.

Morjens. Yeah, i will be running it some weeks to see if it makes any sense in real time
  • #17
  • Sep 6, 2010 11:42am Sep 6, 2010 11:42am
johto: good results, average ratio is 1:2 risk/reward

what kind of strategy you used? trend following or trade fading?
  • #18
  • Sep 6, 2010 11:47am Sep 6, 2010 11:47am
Trend following. System picks a side and follows it until it changes to the other side.

Heres a picture showing Win/Loss hitrate:

http://img.skitch.com/20100906-m6ara...kcceraattx.jpg
  • #19
  • Sep 6, 2010 11:58am Sep 6, 2010 11:58am
johto: i used range bars at ES e-mini trades. Its better to see some trends at range bars than time bars. And entry is much earlier at range bars cuz all bars have the same range size, but time bar have different size of pips.

But i also like tick charts trading ES. 233 or 500 ticks.
  • #20
  • Sep 6, 2010 1:16pm Sep 6, 2010 1:16pm
Quoting mikkom
Do you really mean time in itself does not matter? If you do, then check historical volatility and high/low during the day for different weekdays and hours and you might be surprised.

Time really does matter, only not in the sense many newbies are using it.

Can you show us some example, please?
0 traders viewing now