How can i backtest this strategy?

[MajorDutch]

Hi

I have developed an overnight range breakout strategy for EURUSD and have backtested it manually. So far the trade wins 65% and loses 35% with a strict 1:1 rish reward ratio.

I am very aware that I need to backtest over a large sample size of at least 1000 trades.

Does anyone know how I can backtest this? or would be willing to backtest it for me? The strategy is very simple and involves a open bracket market order at european open.

Thanks in advance.

Simon

 

[intradaybill]

The win rate of 65% is not enough for forex trading because the slippage is too high.

This good be tested on Metatrader. You should visit their site and download the program. Then either learn how to do it yourself or find one of the guys that know the language to do it for you. If you think this is an edge of some kind be aware than having someone else to program it you are basically giving it away.

 

[Hotch]

The win rate of 65% is not enough for forex trading because the slippage is too high.

This is just plain false. It could be possible, but you don't have the information to judge, and people have done brilliantly with much lower win rates.

 

[MajorDutch]

excellent thanks will check out Metatrader backtesting.

 

[tomorton]

The win rate of 65% is not enough for forex trading because the slippage is too high.

QUOTE]


A win rate of 65% on trades with a 1:1 risk:reward is still break-even despite slippage of 30% per trade. Surely nobody in forex has overheads of 30%?

 

[BlackBison]

The win rate of 65% is not enough for forex trading because the slippage is too high.

Ridiculous comment, given he hasn't mentioned stop or target sizes.

Personally I would forward test it with real money over the next month or two and then if it makes money ramp up the bet size..

 

[MajorDutch]

I managed to do some backtesting 63% winners 37% losers strict 1:1 RR + or - 20 pips over a 250 sample size. Max consecutive winners in sample = 11, max consecutive losers in sample = 5. I have not taken into account slippage and spread.

Maybe someone can help me out in understanding slippage. The strategy only involves limit orders to open and close, so surely slippage should be evenly distributed i.e. sometimes I will get filled at 19.8 pips other times 20.2 pips so these should cancel out. I have read reports of some brokers using slippage manipulation so this could come into play. The spread on EURUSD with my broker is 1 pip so this also needs to be considered.

 

[Barramundi]

You've said it involves 'open bracket market order at european open', but later you've said 'involves limit orders to open and close'. Which is it, Limit or Market?

 

[MajorDutch]

Hi Barra

good spot. limit orders only, no market orders.

 

[Barramundi]

Hi Barra

good spot. limit orders only, no market orders.

In that case slippage should not be an issue, but can you guarantee that you will get entered every time?
One of the dangers with limit orders is that you'll get filled on the 37% of losing trades, but only get filled on say half of the 63% of winners (-> 31%). The probabilities then turn from 63% winners/37% losers to having more losers than winners.

 

[MajorDutch]

well in the backtesting all 250 trade set ups were filled. the long and short limit orders to open are set at outset (one cancels other). There is a very small chance neither order will get filled but this is very unlikely. What am I missing here? why wouldnt I get filled on a limit order? if my limit order is by at 20, surely i am going to get filled at 20 when price hits 20. maybe 20.2, maybe 19.9 including slippage.

I just worked out the probability of 11 successive losing trades is approximately 1 in 40,000. So I am thinking I am going to risk 2% per trade to start with.

This is all work in progress I am learning so if anyone has any suggestions please chip in.

 

[Barramundi]

Limit orders guarantee you a price, but not a fill. Market orders guarantee you a fill, but not at a fixed price. So the market could just pass through your order, and if it doesn't come back to pick you up, then you miss out on a fill.

 

[MajorDutch]

ok i understand, I am using my igindex account for this strategy so will see how I get on.

 

[intradaybill]

Ridiculous comment, given he hasn't mentioned stop or target sizes.

Personally I would forward test it with real money over the next month or two and then if it makes money ramp up the bet size..

No you make ridiculous comments because he already said R:R was 1:1.

If you think a 65% win rate with R:R 1:1 will ever make it in forex it is only because you have never traded forex.

65% with 1.5:1 will make it.

 

[the hare]

If you think a 65% win rate with R:R 1:1 will ever make it in forex it is only because you have never traded forex.

What sort of size are you trading ?

 

[MajorDutch]

No you make ridiculous comments because he already said R:R was 1:1.

If you think a 65% win rate with R:R 1:1 will ever make it in forex it is only because you have never traded forex.

65% with 1.5:1 will make it.

Hi Intraday Bill

I am interested to understand why you think a 65% win rate with R:R 1:1 will not work in forex. I am not trying to stir up an argument here I am genuinely interested, I believe I have a good grasp of probability theory.

by my calculations if this is a true edge at 63% winners 37% losers at R:R of 1:1 +/- 20 pips then my expectation is +5.1 pips per trade. Therefore if I trade 250 days per year I can expect to make +1275 pips per annum, at a reasonable order size this could be very profitable for me. There is also the possibility that I could trade the edge on multiple instruments.

By the way I got filled on the trade this morning +20.0 pips result, no slippage

 

[Faceman78]

Hi Intraday Bill

I am interested to understand why you think a 65% win rate with R:R 1:1 will not work in forex. I am not trying to stir up an argument here I am genuinely interested, I believe I have a good grasp of probability theory.

by my calculations if this is a true edge at 63% winners 37% losers at R:R of 1:1 +/- 20 pips then my expectation is +5.1 pips per trade. Therefore if I trade 250 days per year I can expect to make +1275 pips per annum, at a reasonable order size this could be very profitable for me. There is also the possibility that I could trade the edge on multiple instruments.

By the way I got filled on the trade this morning +20.0 pips result, no slippage

The world is full to the brim of haters, detractors and general numpties.

Forget them... have full belief in what you are doing, and if you have done the maths and know what you are talking about.... dont waste time on people who will just try to bring you down.

good luck.

 

[Barramundi]

Hi Intraday Bill

I am interested to understand why you think a 65% win rate with R:R 1:1 will not work in forex. I am not trying to stir up an argument here I am genuinely interested, I believe I have a good grasp of probability theory.

by my calculations if this is a true edge at 63% winners 37% losers at R:R of 1:1 +/- 20 pips then my expectation is +5.1 pips per trade. Therefore if I trade 250 days per year I can expect to make +1275 pips per annum, at a reasonable order size this could be very profitable for me. There is also the possibility that I could trade the edge on multiple instruments.

By the way I got filled on the trade this morning +20.0 pips result, no slippage

I don't and won't speak for Billy Boy, but here are my two bobs worth. Personally, I wouldn't attempt to trade a system with a 65% success rate at 1:1. Based on probabilities it would be fine - for every 3 trades you have a net winner or thereabouts. You've obviously got a good grasp of those issues. The reason would be that it doesn't give much room for error in terms of human mistakes or the risk of being hit by unexpected news events and suffering slippage or missing fills on good trades. What are you going to do if you have 6 straight losses - many would step back, having lost confidence in their system, only to watch the next few, which they also feared would be losers, hit their profit targets. These are the sort of reasons I wouldn't trade it. But if you think you can execute it very well, and actually duplicate the outcomes from extensive backtesting, then by all means go for it.

Personally, I stick with high probability trades that offer more than I am risking. Last night was fairly typical, I had 8 trades opportunities. The worst of these was an AUD/JPY long that went sideways for hours and that I scratched for break even. All the others hit my first target or went a lot further. I didn't manage to get into all of them, but as they were all quality trades, I didn't end up with any losers and thus came out well ahead. Point is I can underperform my system and still make a dollar.

 

[MajorDutch]

ok now I am getting you, trading a 65% edge with 1:1 RR agreed requires excellent execution. I just purchased a book on programming EAs into MT4 so I am hoping to totally automate this strategy also optimise the SL and TP and see how it works on different instruments.

It sure sounds like you have an excellent trading edge however I am yet to find an edge that I am able to trade with high hit rate and at least 1.5:1 RR. If you can give me any pointers in finding a better edge I would be most grateful.

re - your point on 6 losses in a row, this for me would come down to money management. If I risked 2% per trade my account would be sitting on a 12% loss, whilst not ideal I would like to think I would trade through this. 10 losing trades would occur approximately once every 10,000 trades so I should be able to ride out the storm. If I risked 5% per trade though I could be in for a wipeout!

 

[Barramundi]

It sure sounds like you have an excellent trading edge however I am yet to find an edge that I am able to trade with high hit rate and at least 1.5:1 RR. If you can give me any pointers in finding a better edge I would be most grateful.

Even the AUD/JPY long I mentioned, I was correct with the move as it happened, just got bored with the sideways action.

I like to target big moves off large timeframes, but time the entries off smaller timeframes. That way I can keep my stops tight, but still get into big winners. I wouldn't be able to program it, targeting the areas is too discretional. None of it is rocket science tho, just use a number of common tools.