The One trade a day 'so simple its embarrassing system'

[phil_trade]

Haha!

Well many thanks for trying! I presume your backtesting tool takes into account the difference between summer and winter i.e. that april to October I use GMT+1 for the 0900-1300 signal?

Im confident the small sample I have is correct but its just a little too small really. The last two months data were actually gathered through 'forward testing' this, infact the last month or so has been on a live account too and I'm in profit at the moment despite October being the poorest performing month so far.

Again thank you for trying though!

Out of interest what was the SL/TP that gave you that 'best' performance!?

I ran my test from april to October, so no issue with summer/winter time I guess.
I play with start hour from 12 to 14 and delta hour from 3 to 5 to test a large range of param.

How did your trade stay live ? only some tick or several minutes ? If several minutes, my M1 data must be ok ?

My best result was TP 15 SL 70

[FlyHighDrifter]

Hmmm interesting. M1 should be fine, I have not noticed them closing out super fast so far.. but then backtesting from the chart I never had to go below a 5m chart to determine my result so I guess in theory its possible that many closed in less than a minute but I would think that is very unlikely.

I have not gathered the duration of the trades ( it was just taking too long). I would say the trades usually close out in the first hour, often able to just grab 10 - 15 pips in the first 15 minutes. This does make the entry time critical..since going live I have been using a SL10 and a tp of 10-12 and at the start of this week on monday I lost a trade that should have been a 12 pip win just because I entered 30 seconds early!! I was always surprised by how often I noticed that price would then reverse after the initial 10-15 pip move in my direction. I guess that's the beauty of hard manual backtesting - you get a feel for things, and I suspect this is why my backtest results seemed to suffer as I used TP's much larger than 15 or so.

The next work on my on my list is to check through the entire data set I have with an opposing direction trade at around 13:15ish.

Even when using the larger stops (ATR) which was typically between 30 and 40 pips, most of the time the trade would be closed within a couple of hours.

[Chrispook31]

Loving the simplicity of this system FlyHigh - so easy to log on for a set time rather than aimlessly staring at the charts for hours.

I'm going to get involved in this on demo account.

Look forward to hearing more

[phil_trade]

Hmmm interesting. M1 should be fine, I have not noticed them closing out super fast so far.. but then backtesting from the chart I never had to go below a 5m chart to determine my result so I guess in theory its possible that many closed in less than a minute but I would think that is very unlikely.

I have not gathered the duration of the trades ( it was just taking too long). I would say the trades usually close out in the first hour, often able to just grab 10 - 15 pips in the first 15 minutes. This does make the entry time critical..since going live I have been using a SL10 and a tp of 10-12 and at the start of this week on monday I lost a trade that should have been a 12 pip win just because I entered 30 seconds early!! I was always surprised by how often I noticed that price would then reverse after the initial 10-15 pip move in my direction. I guess that's the beauty of hard manual backtesting - you get a feel for things, and I suspect this is why my backtest results seemed to suffer as I used TP's much larger than 15 or so.

The next work on my on my list is to check through the entire data set I have with an opposing direction trade at around 13:15ish.

Even when using the larger stops (ATR) which was typically between 30 and 40 pips, most of the time the trade would be closed within a couple of hours.

With such small profit by trade, broker's datafeed is very important. Refer to gnbnzd strange move on alpari UK last week ! I guess it's a part of our different result ?

[FlyHighDrifter]

Philippe,
Yea I guess you're right. I have an alpari UK demo so I shall have a look at the last few weeks just to see how different it might have been with this system.
At the beginning I nearly started to manually test this with my Alpari account but I always intended to use IG to trade live due to the lower spreads so thought it made sense to backtest on their data.

Ofcourse on the Alpari UK the signals are from 1100-1500 with the trade being taken at 1500 due to the two hour offset.


Chris,
Many thanks - I hope it continues to perform the way it has! Although I am profitable live with this so far, I would recommend the demo for now!

Fly

[FlyHighDrifter]

Thank goodness it won today... didnt want the first result after me posting this to be negative!

[FlyHighDrifter]

With such small profit by trade, criminal's datafeed is very important. Refer to gnbnzd strange move on alpari UK last week ! I guess it's a part of our different result ?

I have manually looked back over my alpari UK charts and and compared 6 weeks from 1st October to 14th November. The results were not exactly the same but almost... there were only two differences, one where a 10 pip profit was missed (only 9.9 in alpari) so it turned to a loser and one where a doji had formed between the 0900 and 1300 time frame (1100-1500 in alpari).. the alpari signalled a different trade direction that actually made it more favourable and scored a win over my IG signals.

I guess there will be differences when the movements are this small but surely where the differences lose on one trade, in theory they could win on another?

From this small comparison the results would have still been profitable - infact more so.

I can't think of any other reasons why your tests would be so different.

[johnlucas]

Hi,

I have coded a "SuperSimpleEA", just to be in line with this thread , that is supposed to automate the logic explained in post 1.
Time is GMT + 0 (winter UK time).

I have backtested some years of data using this ea. 2012 looks very good...but 2011, 2010, 2009...are not acceptable.

While it is perfectly possibile that this EA is a piece of garbage, if you like, give it a go.

It let's you test the logic faster (if it is correct in its assumptions),

cheers,

John.

[FlyHighDrifter]

Johnlucas - thank you so much, that is great! This is useful to hear... Whilst my tests were good I am concerned it may (or may not be) just a blip rather than the normal performance.

I'm still getting to grips with EA's I shall have a real good play with this

This is why I love this forum, thanks for the input so far everyone!