Pin Bar System

[pocrel]

I dont think its going to be possible to get accurate auto pivots for commodities (i will test index in a bit).
Obviously to auto generate the pivot values i need to get the open, close, high and low from the prev day (market hours). So in theory i should be able to look at livecharts, get these values and find them on an hourly chart of yesterday. This isnt working and i bet your charts are different again. An example

SLM Livecharts (WTI)
Open 98.95 99.06
Close 98.95 98.97
High 100.44 100.48
Low 95.23 95.24

This is assuming WTI markrt times that pivots are calculated at are 00.00 GMT to 21.00 gmt. If anyone can correct these times please let me know.

[edit] The format of the little chart above is messed but you get the idea.

take the high on the chart and the low (they are both very close to livecharts few points in it)

and the close is 1030pm on smartlive for wti again very similar to livcharts

same rules for brent

 

[pgrv]

Hi Mata

I think there is a problem with this code, regarding expiry of the orders. Orders will not expire correctly


int et;
// if(et>0)
// {
et = TimeCurrent()+(PERIOD_M30*60)*BarsTillExpire;
// }
// else
// {
// et=0;
// }


Thank you, this is great thread


/Peter

All, I know some of you have been having some problems with the EA. One of the common ones is that it is deleting orders as they should be executed. This is down to it calculating lots that would put you over your margin. What it should have done is look at how much margin per lot is required. I have fixed that now but it will still cause problems if you have a very small account and lotsize is close to the minimum.
....

 

[Mata Nui]

take the high on the chart and the low (they are both very close to livecharts few points in it)

and the close is 1030pm on smartlive for wti again very similar to livcharts

same rules for brent

Done:
Backtest 04/04/2011 to today with 2.5% risk.

Using the new 'More Accurate' pivots BRENT makes just less then 3% PROFIT using M30 and the system rules.
Using the Old auto pivots BRENT makes just less than 20% PROFIT same as above.

As i keep saying the EA works a LOT better on M15 on pretty much every instrument i have tested.

Using the new 'More Accurate' pivots BRENT makes 12% PROFIT using M15 and the system rules.
Using the Old auto pivots BRENT makes 57% PROFIT same as above.

Whatever way you look at it . Its profit

 

[Mata Nui]

More tests:
WTI looses on M15 whichever pivots are used but looses twice as much with the new pivots.
On M30 it wins whichever pivots are used and wins twice as much with the new pivots.

FTSE Highs/lows are nothing like my chart so i guess they are quoting cash but SLM use a futures contract.

 

[pocrel]

Done:
Backtest 04/04/2011 to today with 2.5% risk.

Using the new 'More Accurate' pivots BRENT makes just less then 3% PROFIT using M30 and the system rules.
Using the Old auto pivots BRENT makes just less than 20% PROFIT same as above.

As i keep saying the EA works a LOT better on M15 on pretty much every instrument i have tested.

Using the new 'More Accurate' pivots BRENT makes 12% PROFIT using M15 and the system rules.
Using the Old auto pivots BRENT makes 57% PROFIT same as above.

Whatever way you look at it . Its profit

have you been closing half at 1 to 1 and leaving stop below the pin and then letting it ride to the next pivot level?

what were the old pivot derived from?

 

[pocrel]

More tests:
WTI looses on M15 whichever pivots are used but looses twice as much with the new pivots.
On M30 it wins whichever pivots are used and wins twice as much with the new pivots.

FTSE Highs/lows are nothing like my chart so i guess they are quoting cash but SLM use a futures contract.

Smartlive is using cash market, so pivots need to be drived from these market times so that needs to be from 8 to 430pm

if it takes prices outside these times then its wrong

 

[Mata Nui]

have you been closing half at 1 to 1 and leaving stop below the pin and then letting it ride to the next pivot level?

what were the old pivot derived from?

As per the system for these results. half off at 1x, stop left at the pin point, target is next pivot. I have set this as default now so every initial test is as per exact system rules.

I dont get the difference in profitability but reading back on the posts there are an awfull lot of times when we are using discretion in closing early etc.

 

[pocrel]

Trade alert order to short eurgbp

 

[Allistah]

Can you post a chart that points to the exact bar you got in and about at what point on the bar you got in at? It looks like you get in on the third bar of the pair and this confuses me because you have to wait unilateral that third bar closes to make sure the pattern is valid.

I so want to understand this so any help is appreciated.

-Alli

 

[Mata Nui]

have you been closing half at 1 to 1 and leaving stop below the pin and then letting it ride to the next pivot level?

what were the old pivot derived from?

Sorry, missed the last question. the old pivots are derived from the day candles, the new ones are from 1/2 hour and market open & close parameters.

 

[pocrel]

Can you post a chart that points to the exact bar you got in and about at what point on the bar you got in at? It looks like you get in on the third bar of the pair and this confuses me because you have to wait unilateral that third bar closes to make sure the pattern is valid.

I so want to understand this so any help is appreciated.

-Alli

hi alli

the pin bar is the last bar to close 130pm to 2pm and this has given the signal to short shown by the red arrow on the chart

not sure what your really talking about but have a look over the thread its all spelled out

 

[pocrel]

Sorry, missed the last question. the old pivots are derived from the day candles, the new ones are from 1/2 hour and market open & close parameters.

so the only different with brent and wti should be the close time? to derive the pivots?

 

[Mata Nui]

Hi Mata

I think there is a problem with this code, regarding expiry of the orders. Orders will not expire correctly


int et;
// if(et>0)
// {
et = TimeCurrent()+(PERIOD_M30*60)*BarsTillExpire;
// }
// else
// {
// et=0;
// }


Thank you, this is great thread


/Peter

You are indeed correct. Thank you, fixed now. Hope it doesnt mess up my backtests. What the error would have done is prevent any orders expriring.

 

[Allistah]

Ok, going to work now so I'll be watching this close as I learn whats up. Thanks for being patient with me guys..

-Alli

 

[Mata Nui]

so the only different with brent and wti should be the close time? to derive the pivots?

Yes, I would guess the opens would be the same, The high and low will obviously be the same (apart from any broker differences).

 

[pocrel]

Trade alert order to short eurgbp

short eurgbp now

 

[pocrel]

trade alert looking to short usdjpy

 

[pocrel]

short eurgbp now

out of eurgbp for +0

no momentum given it enough time

plenty more to look at now

 

[megamuel]

While i check todays trades can someone test this opt set for me please?

Initital settings:
EUR/USD
15 Min
04/04/2011 to 13/05/2011

Then load up the set. Once its done for that date range can you please post your % gain or loss and test other date ranges. SLM data starts to get messed up after about 2 weeks so i can only reliably backtest a month.

Hi Mata, I loaded up the set and ran the backtest and here are the results I got (attached). Are they similar to yours? Look good but once you try a year back test this happens....

 

[megamuel]

Meant to be this one...