what can we do with 90% winrate? | Page 4 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Made a 100% Winrate EA 26 replies

Naustingfx End-Of-Day Trading System Journal @ 68.26% WinRate 34 replies

Winrate >90% - Unbelievable 4 me: omg 8 replies

MACD +60% Winrate 1:1.5 RR 63 replies

Statistical Gambling System (High Winrate, but NO EDGE) 372 replies

Attachments: what can we do with 90% winrate?
Exit Attachments

what can we do with 90% winrate?

  • #61
  • Oct 21, 2015 4:06pm Oct 21, 2015 4:06pm
may be you should code and test the opposite.
with same rr trading the last close direction.

Quoting vasily
risk 3 reward 1 trade the opposite of last close / daily
  • #62
  • Oct 21, 2015 5:03pm Oct 21, 2015 5:03pm
I did a small test just to see where the losses happens, 25 wins in a row apparantly it works great when it's trending, added a momentum indicator and a loss came at the reversal I also noticed that it doesn't trade on all bars, maybe I don't understand the settings, what I want to do is set a range between two dates, for example when we have a ranging market going sideways, but if I do it don't trade much?
Attached Image(s) (click to enlarge)
Click to Enlarge Name: report2.jpg Size: 157 KB
Click to Enlarge Name: visual.png Size: 30 KB
Bulls are stupid Animals!especially when Im short!
  • #63
  • Oct 21, 2015 5:03pm Oct 21, 2015 5:03pm
Just a few thoughts......

I think daily ATR sizing of TP and SL is best.
you have ranging years where the daily ATR of EUR/USD is just about 50 pips a day.
in that case you could have a lower TP of 10 pips and SL of 30 pips.

often at the end of the day price might be in a range already.
you look at 0hrGMT and price is in the bottom half of that range: buy
top half: sell

another possibility for the Asian ranging period before European open:
price is 1.1500......you place 2 oco orders
buy when price drops to 1.1480 - TP 1.1500/SL 1.1420
sell when price rises to 1.1520 - TP 1.1500/SL 1.1580
  • #64
  • Oct 21, 2015 5:41pm Oct 21, 2015 5:41pm
RoughTrader

I would have thought that a trending market would kill this strat?, on strong trends in the daily chart i notice that it rarely pulls back to hit the 0.80 TP

Ranging on the other hand, you encounter red/green/red/green as well as long wicks which would make this more profitable
  • #65
  • Oct 21, 2015 6:09pm Oct 21, 2015 6:09pm
25 win = 8 loss
Quoting roughtrader
I did a small test just to see where the losses happens, 25 wins in a row apparantly it works great when it's trending, added a momentum indicator and a loss came at the reversal I also noticed that it doesn't trade on all bars, maybe I don't understand the settings, what I want to do is set a range between two dates, for example when we have a ranging market going sideways, but if I do it don't trade much? {image} {image}
  • #66
  • Edited 8:29pm Oct 21, 2015 7:21pm | Edited 8:29pm
Quoting MathTrader7
Here is Version 1.10 of the EA that doesn't trade if the body size of the previous daily candle is less or greater than two user specified thresholds.From two above backtests we can observe that a win rate of about 70% can be expected from this strategy. If we add more effective filters we may be able to increase the win rate, but it is not easy to reach to 90% win rate... I guess...
MT7,

One question if I may: why does the position size increase abnormally in the two places indicated by the arrows in my screenshot?
Is the EA running some kind of martingale?

Thanks,
David
Attached Image (click to enlarge)
Click to Enlarge Name: PPI 034.png Size: 11 KB
  • #67
  • Edited 10:40pm Oct 21, 2015 10:27pm | Edited 10:40pm
Quoting hanover
{quote} MT7, One question if I may: why does the position size increase abnormally in the two places indicated by the arrows in my screenshot? Is the EA running some kind of martingale? Thanks, David {image}
Even though I did not write the EA, I believe that the reason is because it utilizes dynamic position sizing due to % risk. If the pips to the SL is small enough, the position size will be significantly increased to keep the $'s risked equal to those positions where the pips to the SL are greater.

I am positive that there is no Martingale type position sizing in the EA.

May GREEN pips continually flow into your account and your life.
  • #68
  • Edited 1:38am Oct 22, 2015 1:28am | Edited 1:38am
This is version 2 of Retracer_Extender EA for testing purposes.

Updates :

  1. ATR filter with input ATR period added, that can be set to avoid trading if previous candle range is less that ATR x Min Multiplier, or larger than ATR x Max Multiplier.
  2. Max Lots input that can be set to avoid trading large volumes if SL is too small. In such case Max Lots is used instead of calculated lots.
  3. The EA uses previous candle close instead of (high/low) as indicated by roughtrader.

Notes :

  1. The EA can be used to trade in the SAME direction as previous candle, or Opposite Direction (original method)
  2. The EA can be set to close open trade at end of day, or keep them until stopped or reach target
  3. I tried different settings but unfortunately I didn't notice any real edge

Ooops : that was an old version.
New Version is uploaded now. Sorry.

Attached File(s)
File Type: ex4 Retracer_Extender_02.ex4   12 KB | 682 downloads
No guts, no glory
  • #69
  • Oct 22, 2015 1:29am Oct 22, 2015 1:29am
This EA gives good results to daily chart
eur usd
but I'm not sure of the result
I dont have no idea code and the ea
Can I learn ,The idea of yours in this matter

Regards.
Attached File(s)
File Type: mq4 Break Out [Previous Candle] High Low.mq4   4 KB | 690 downloads
File Type: ex4 Break Out [Previous Candle] High Low.ex4   10 KB | 508 downloads
  • #70
  • Oct 22, 2015 2:36am Oct 22, 2015 2:36am
Quoting MathTrader7
Here is Version 1.10 of the EA that doesn't trade if the body size of the previous daily candle is less or greater than two user specified thresholds. {image}{image}{file} From two above backtests we can observe that a win rate of about 70% can be expected from this strategy. If we add more effective filters we may be able to increase the win rate, but it is not easy to reach to 90% win rate... I guess...

can you give me the results in a form of x and o
where x is a losing trade and o is a wining trade
  • #71
  • Oct 22, 2015 2:36am Oct 22, 2015 2:36am
Quoting SmoothTrader
Even though I did not write the EA, I believe that the reason is because it utilizes dynamic position sizing due to % risk.
If the pips to the SL is small enough, the position size will be significantly increased to keep the $'s risked equal to those positions where the pips to the SL are greater. I am positive that there is no Martingale type position sizing in the EA. May GREEN pips continually flow into your account and your life.
SmoothTrader,

Many thanks for your reply. That suggests an (occasional) vast deviation in the SL size.

The plot doesn't look like a standard martingale (the pos sizing increases don't coincide with major equity falls), but I was curious as those two 'spikes' seemed to be way bigger than the sizes for the other 600-odd trades.

Best wishes to you also.
David
  • #72
  • Oct 22, 2015 3:31am Oct 22, 2015 3:31am
Hi all.

Here is one with 90% win rate...
Any ideas why it only works in backtesting and with a low spread ?

Attached Image (click to enlarge)
Click to Enlarge Name: 57Mill.gif Size: 13 KB




Attached File(s)
File Type: ex4 !SpeedTrap V3.0.ex4   46 KB | 1,033 downloads
  • #73
  • Oct 22, 2015 3:58am Oct 22, 2015 3:58am
Quoting Mingary
Hi all. Here is one with 90% win rate... Any ideas why it only works in backtesting and with a low spread ? {image} {file}
In my experience, any EA which has low SL (2-4 pips, or 20-40 points), and identically low TP values will give artificial results, since the ticks are simulated. It is even worst for trailing stops, especially if the distance is low.
In my opinion backtests should be initially done on open prices only to see if there is any edge, than you can dive into other details.

My 2 cents....
  • #74
  • Edited 4:23am Oct 22, 2015 4:09am | Edited 4:23am
Quoting Mingary
Hi all. Here is one with 90% win rate... Any ideas why it only works in backtesting and with a low spread ? {image} {file}
Simulated ticks by the MT4 strategy tester represents some patterns in tick speeds, tick jumps, etc. which makes some newbie EA developers think that they had found holy grail.

Edit: Here is a back test with real ticks downloaded from Dukascopy history for this EA:
Attached Image (click to enlarge)
Click to Enlarge Name: BT_Speed.png Size: 47 KB

Which means with real ticks there were only two trade patterns and both losers!
Trading is the hardest way to make easy money...
  • #75
  • Oct 22, 2015 4:15am Oct 22, 2015 4:15am
Quoting hanover
{quote} MT7, One question if I may: why does the position size increase abnormally in the two places indicated by the arrows in my screenshot? Is the EA running some kind of martingale? Thanks, David {image}
Hi Honover,

There is no Martingale!
As SmoothTrader mentioned, that is because of the lot size resolution set by the broker. For example, if lot step size is 0.01, then there are situations that given a stop loss distance, it is not possible to calculate a lot size that makes the trade risk equal (even approximately) to the desired risk.

Best,
Math
Trading is the hardest way to make easy money...
  • #76
  • Oct 22, 2015 4:15am Oct 22, 2015 4:15am
Quoting MathTrader7
{quote} Simulated ticks by the MT4 strategy tester represents some patterns in tick speeds, tick jumps, etc. which makes some newbie EA developers think that they had found holy grail.

Exactly. Thats why an EA should be tested in live market conditions for at least 2 weeks, than backtested on the same 2 weeks to check the results. The results of live tests and backtest must be identical, otherwise the EA simulation is not valid.
  • #77
  • Oct 22, 2015 4:16am Oct 22, 2015 4:16am
Quoting GoldTheHun
{quote} Exactly. Thats why an EA should be tested in live market conditions for at least 2 weeks, than backtested on the same 2 weeks to check the results. The results of live tests and backtest must be identical, otherwise the EA simulation is not valid.
I totaly agree with you!
Trading is the hardest way to make easy money...
  • #78
  • Oct 22, 2015 6:13am Oct 22, 2015 6:13am
Quoting MathTrader7
{quote} I totaly agree with you!
you could have a point, I will start running it manually on demo starting monday, just opened a fresh demo account, so no other trades have been made on it
Bulls are stupid Animals!especially when Im short!
  • #79
  • Oct 22, 2015 6:16am Oct 22, 2015 6:16am
Quoting abokwaik
This is version 2 of Retracer_Extender EA for testing purposes. Updates : ATR filter with input ATR period added, that can be set to avoid trading if previous candle range is less that ATR x Min Multiplier, or larger than ATR x Max Multiplier. Max Lots input that can be set to avoid trading large volumes if SL is too small. In such case Max Lots is used instead of calculated lots. The EA uses previous candle close instead of (high/low) as indicated by roughtrader. Notes : The EA can be used to trade in the SAME direction as previous candle, or Opposite...
Thanks abokwaik
Bulls are stupid Animals!especially when Im short!
  • #80
  • Oct 22, 2015 6:24am Oct 22, 2015 6:24am
Quoting MathTrader7
{quote} Hi Honover, There is no Martingale! As SmoothTrader mentioned, that is because of the lot size resolution set by the broker. For example, if lot step size is 0.01, then there are situations that given a stop loss distance, it is not possible to calculate a lot size that makes the trade risk equal (even approximately) to the desired risk. Best, Math
would it be possible to make an EA without the filters, I would like it to test trade on all bars for now, that way I can go in between certain dates and test what works,
for example, there are sudden long dips in the graph, so I only wouild like to test that area.
Bulls are stupid Animals!especially when Im short!
2 traders viewing now