My experience on building EA - page 2

[Simon Gniadkowski]

mp88:

you guys are just chatting here,lol, well, the chart is useful but it can be shortened into one short sentence in summary,which i found more direct

win: loss ratio is directly proportional to risk:reward ratio

That statement is correct only for a Strategy that forecasts a winning trade as well as a Coin Toss.  The chart can be used to determine if ones Strategy is as good or better than a Coin Toss,  if it's lose to a Coin Toss it will not make money long term . . . .  anything can win in the very short term,  even a Coin Toss.

[Simon Gniadkowski]

mp88:
that really makes me wonder how you guys trade, or you guys dont trade at all?

Not yet,  still coding up my ideas.  I see no need to test a bad Strategy by "forward testing".

[Alain Verleyen]

mp88:
that really makes me wonder how you guys trade, or you guys dont trade at all?

Before all, I am programmer since 1986. I only begin to trade in late 2011. So my advice is surely more confident in programming question that of trading.

My experience, as programmer in MT4/MT5, is that you can't hope make a winning EA in long term, if you are not able to trade manually. So I had begin to trade manually.

I've never been able to use popular indicator in a winning strategy, so I use mainly chart action price, candlestick pattern & support/resistance.

[Simon Gniadkowski]

angevoyageur:

Before all, I am programmer since 1986. I only begin to trade in late 2011. So my advice is surely more confident in programming question that of trading.

My experience, as programmer in MT4/MT5, is that you can't hope make a winning EA in long term, if you are not able to trade manually. So I had begin to trade manually.

I've never been able to use popular indicator in a winning strategy, so I use mainly chart action price, candlestick pattern & support/resistance.

Bar patterns,  did you see these charts I uploaded to the mql4 Forum just over a year ago ?  https://www.mql5.com/en/forum/138317

[Alain Verleyen]

RaptorUK:
Bar patterns,  did you see these charts I uploaded to the mql4 Forum just over a year ago ?  https://www.mql5.com/en/forum/138317

I didn't see it. Good work :-)

What'is your conclusion ?

[Simon Gniadkowski]

angevoyageur:

I didn't see it. Good work :-)

What'is your conclusion ?

Many/most/all  strategies that depend on bar patters don't work . . .  having said this I haven't tested many/most/all strategies based on bar patterns so I can only apply my conclusion to my own work.  I invite everyone else to look at the charts,  think about what they see and come to their own conclusions ;-)

[Alain Verleyen]

RaptorUK:
Many/most/all  strategies that depend on bar patters don't work . . .  having said this I haven't tested many/most/all strategies based on bar patterns so I can only apply my conclusion to my own work.  I invite everyone else to look at the charts,  think about what they see and come to their own conclusions ;-)

All depend of what you are looking at chart. For me, the 6 charts in your link, ALL show some interesting pattern. You have to be flexible with pattern use, and don't rely on them alone.

Also I think, higher timeframe is better to use with bar/candle pattern (and I know your next objection ;-))

EDIT: I wonder if your chart shifting technique could not be use as a filter ? A signal present in 4, 5 or 6 charts must be stronger that if only in one chart.

[Simon Gniadkowski]

angevoyageur:

All depend of what you are looking at chart. For me, the 6 charts in your link, ALL show some interesting pattern. You have to be flexible with pattern use, and don't rely on them alone.

Also I think, higher timeframe is better to use with bar/candle pattern (and I know your next objection ;-))

EDIT: I wonder if your chart shifting technique could not be use as a filter ? A signal present in 4, 5 or 6 charts must be stronger that if only in one chart.

Had to go out . . .

It may depend on what one is looking for, it may also help confirming on a lower timeframe  or higher timeframe,  there are lots of things that can be tested thoroughly by coding an EA and running it on several instruments for several years,  anything that has merit will so it across instruments and consistently through time,  IMO.

 

Yes,  D1 and H4 are bad timeframes . . . unless you know which D1 and H4 timeframes are the valid ones and why,  but you knew I was going to say this ;-)