100% Mechanical PinBar system

Nothing wrong with an indicator finding all the pins, even the stupid ones we glance over as not important. That should be the beauty of an EA, someone could combine it with other indicators to filter out some pins. Have an ATR indicator to filter out those crazy small pins that are more like an indicision in the market, not a real pin. Have conditions, like, 100ma over 200ma by minimum 7 pips for long on 15min chart. Only alert to pin that forms on correct side of MA's and the correct way, not upside down. Also make the EA alert when the pin forms and touches the 20ma and 50ma for confluence.. You get the idea. You could really filter out a massive amount of crap pins and trades and just stick to the best ones with a good EA.

First thing I'd do is stop thinking of the pin as a typical pin, just make it a long wick candle indicator that alerts if a candle has a wick longer than say 35% of the total length. From there someone could build on it.

Hi Pasopvirpot,
I think it is possible to develop a profitable mechanical system based on pinbars.
Most people seem to use pinbars for either very long term position trading or short to medium term swing trades. Generally the best reversal pins occur around support or resistance levels and these are seen as almost the holy grail of price action. It is quite difficult to create a mechanical way of determining precisely where the major levels are. It is simple enough to do it by eye and drawing a line but to do the same thing on a computer requires you to tell the computer which swing highs and swing lows to consider, sounds easy enough but doing so results in a little too much curve fitting. The results tend to be more of a zone of price, so some discretion is still required.
So leaving that problem aside for now, the other main challenge is filtering out all of the dross pinbars from the good ones. I think Juicy t was correct when he said MAs ATR etc is the way to go.
Sorry I have no MT4 coding experience, but have set up a strategy to run on my platform.
I have attached a PDF with all the charts and results I am about to mention.
The 1st Chart is the 60 minute chart for the GBP futures and shows the 3 types of pinbars, bullish in green, bearish in red.
The dark green dot is a simple pinbar
The light green dot is a hammer pinbar
The green coloured candle is a bullish pinbar.
The red dots are the reverse of the above
You will notice there are a total of 18 pinbars and visually you can see 6 definitely would have failed, the remainder possibly may have worked.
The second chart shows the filters I have used and the plain language formula used for a pinbar reversal buy is

Hull Moving Average (Close, 9) < Hull Moving Average (Close, 25) And a Pinbar Bullish Or a pinbar up Or a pinbar hammer And Range > Avg True Range (21) * 0.75 And StochK (14, 3) > StochD (14, 3 , 3) And StochD (14 , 3 , 3) <20

This could be further filtered by using, for example, only the pinbar bullish, which is the most infrequent of the pinbars.

The reversal pinbar sell is opposite to the above and stoichastics are >84

1) 2 Hull moving averages to determine slope/trend
2) One of the 3 types of pinbars must be formed
3) Range of the pinbar must be greater than 75% of the ATR (21)
4) Stochastic D must be in an oversold condition and Stoichastic K must have crossed D to indicate a change of direction.

The filtered signals when applied to the chart are shown as a red Diamond above the pinbar for a sell and
A green diamond below the pinbars, for a buy.
This filtering has now cut the number of visible signals on the chart, down from 18 to 4, which visually look OK

The next chart shows the strategy placed on the chart. The target is 140 pips and stop-loss is opposite side of the pinbar
Green filled in box, is a winning trade. Pink filled in box, is a loosing trade
This simple strategy gives the results shown in the 1st table
12 trades in 2 months 6 winners 6 losers so 50% win rate Profit Factor 4.75 profit $4144 at ($6.25/pip) =663 pips = 331pips/month.

This can be further improved, as shown on the next chart and table, by adding a trailing stop based on a target of 35 ticks. Once the trade is in profit by 35 pips, the trailing stop kicks in and becomes a 3 bar low trail (for long 35tgt, for short 65tgt)
You will note the failed trade on the previous chart becomes a winning trade and the results now become.
12 trades in 2 months 8 winners 4 losers so 66.7% win rate Profit Factor 7.7 profit $5231 at ($6.25/pip) =836 pips = 418 pips/month.

The next table shows the back-test over the last 16 months and it still produces respectable result even with a 14 trade losing streak in the middle.

The next chart and table show a benchmark comparison of the back test without the filters and the results are:

113 trades in 2 months 21 winners 92 losers so 18.6% win rate Profit Factor 0.99 loss $-125 at ($6.25/pip) = -20 pips = -10 pips/month.

So at first glance it would appear filtering pinbars can be effective and potentially profitable.

These tests have just been on reversal pinbars with only 1 lot per trade, if we had been trading two lots with the second lot as a runner the last trade would still be going from entry on the 7th Jan 2011 at 1.5425, until today where it closed at 1.5994 i.e. +569 pips
Additionally there is multi-lot/ multi position trading to consider plus pinbar continuation signals, which are possibly safer and more profitable than reversal pinbars.
To filter the continuation pinbars the same filters can be utilised but the stoichastics configuration and values would need to be adjusted and I would probably look to include the EMA slope from a higher timescale as a further confirmation of trend.
I appreciate this is not a comprehensive test. I have not applied it to other pairs or other time frames; however these results should provide an interesting springboard for further analysis.
Sorry I can’t help with MT4 coding but hope this is of some help.

Tonkar

i appreciate the work being done here. but i fail to understand whats the need for a mecahnical system ? in my view mechanical system is required when there are complex calculations and you cant monitor the market 24x5.

with pinbars, all you need to do is spend 1 min on a chart (each pair daily, H4) and you identify 10-15 pairs and spend some 15 min everyday in the morning and visually scan the charts. you can add other criteria like location, round numbers, trend lines, SR lines etc. trading just pin bars will not be profitable if you dont consider other facts too

See my chart.... 15min time frame. That's why.

on lower tf, you get too many false signals...

i would trade pins on H4 and above

It depends on your conditions you set for the EA. You could set a lot of conditions that filter the crap, then once met, alert you to a pin on lower time frames, but it may not work at all on higher time frames. This is why an EA needs to be either adjustable or made with a particular time frame in mind.

I have traded pins/long wick candles off most time frames for many years, if I didn't think it could be done, I wouldn't waste my time thinking about it. I appreciate your feelings on the time frame and you are welcome to trade how you like, but lets move forward with trying to develop something if the EA writer still wants to. Lets start with first picking up ALL the long wick candles, bad ones and all.

Juicyt

Yes timeframes below 4H is basically random data, patterns only begin to give you the upper hand from 4H and above.

Had to learn this the hard way

if it happens on lower timeframes, it will happen on higher ones.

I have updated the first post with version 1.4 of the EA. Several improvements have been made, it now uses market orders only, it draws arrows for qualifying bars, takeprofit can be set using a risk reward ration.

Please have a look and give me your feedback. Also let us know if you find any setting worth looking at.

Next development will probably be trend filters.

See below an image of trades on AUD/USD with risk:reward 1:1

I only took a glance and could already see many pins missed. What are the requirements for a pin?

Default setting is to only identify the best pinbars according to my criteria.

Changing the following will identify a lot more pinbars:

1. Set PB_NewHighOrLow to false to identify all pinbars without looking for newer highs/lows

2. Try setting PB_OpenFromTip and PB_CloseFromTip to 0.50 to relax the wick criteria.

If there are clear pinbars it's still not identifying let me know the timeframe, date and time of the bar so I can have a look.

The EA has been updated to V1.5.

You can now use RSI, Stochastics and ADX filters. I have also added descriptions to the different settings to explain them better.

With regards to these filters, you can set max and min levels. E.g. for RSI you can set it to take long trades between 55-100 (with the trend) or set it to take long trades when 0-30 (counter trend).

With regards to the timframes you have to enter the minutes of the timeframe e.g. 15 is 15 minute, 240 is 4H timeframe.

Please let me know if you experience any problems.

Please share any settings and backtests you find profitable.

I've been looking at and trading pins for many years and I can only give you this advice if you want to move forward... Think different than most people, but still use what works for most.

Use sup/res (this means you need an indicator to display this and it can be done if you use monthly OHLC) Use possibly fibs (once again you need an indicator but can be done) and use maybe MA's for confluence as dynamic sup/res. If you do this, loosen up on your pin criteria, you will do well. I would start at monthly OHLC as sup/res, use a zigzag indicator to help and draw some fibs on the monthly zig to zags then drop down and use pins that hit those lines with confluence. Manage to the next sup/res line. You can trade this on 15min well up to 1hr well. 4hr possible well. If you want to filter then use the positions of the MA's to determine a clear trend and not consolidation.

Or do nothing above but load it up with lagging indicators to filter out the trades and miss the good ones and wonder why it doesn't work consistently.

If you don't think what I am writing will work please let me know why. Or if it is too hard and you don't want to do that much work, let me know as I can only imagine it would be hard work.

Juicyt

Hi,

I've tested the pin bar ea 3 ver. looks interesting, just that the stop loss needs to be adjusted as i'm using the default settings.

cheers..................

Good job with your first EA.... You should really consider to upload the mq4 file, no one can help you if we don't have the code.

If somebody really wanted to steal your code they can do it easily, there are loads of ex4 decompilers out there, your code aint safe even in ex4 format.

About the code, try not to put too much settings into your EA, the more settings the bigger the chance for curve fitting when optimizing.

I really agree with juicyt, and thanks for taking your time to explain

Thanks Sim, but somehow I feel as though it's falling upon deaf ears.

Juicyt

Nicely put and very true...

Thanks for this advice JuicyT.

What would you use for your stop loss?

It depends on your sup/res, but I would put it behind the sup/res line that price just bounced off. Not too close as it is an area, not a hard and fast line, but behind it none the less.

Juicyt