Inside Bar Trading Revisited | Page 6 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

3 Bar Fractal with mid-bar and bar on each side 11 replies

Inside bar or outside bar identifier 7 replies

Journal related to: Advanced Inside Bar (IB) Trading 19 replies

Pin bar, pin bar, pin bar 0 replies

Attachments: Inside Bar Trading Revisited
Exit Attachments

Inside Bar Trading Revisited

  • #101
  • Jan 15, 2009 8:58am Jan 15, 2009 8:58am
The pure IB trading method suggests that SL is moved to the BE point after taking 1/2 profit at 1/2 ATR.

Up until now, the EA has been moving the SL to the opening price at this point. Does giving the price more room to move improve the chances of catching the trend again after a small reversal? Well the figures are as follows:

EURUSD 15M, ZERO SL, 3001 trades, 2522 wins (84%), -£460.96
EURUSD 15M, BE SL, 2962 trades, 1315 wins (70%), -£451.00

EURUSD 30M, ZERO SL, 1765 trades, 1500 wins (85%), £9.51
EURUSD 30M, BE SL, 1706 trades, 1224 wins (72%), -£4.63

EURUSD 1H, ZERO SL, 929 trades, 766 wins (82%), -£178.71
EURUSD 1H, BE SL, 904 trades, 639 wins (71%), -£181.30

As expected the win ratio drops, as we're now setting a SL at the BE point for the trade as a whole, rather than the zero point for the second half of the trade. What we don't see is any major change in profit, so we could assume that there are as many trades that carry on to hit the TP as do hit the SL.

As with all the analysis so far we need to revisit them once we've started to use indicators and, hopefully, isolated the majority of losing trades.
  • #102
  • Jan 15, 2009 11:09am Jan 15, 2009 11:09am
Quoting 1ka
jcbwrites, hmm, tricky one....
Well--many thanks for the help, 1ka. Attached is the log file for yesterday. I left in the H1H1_IB indicator lines along with the IBTrader lines to show the correlation--they are both seeing the IBs. One thing I have noticed is a lot of initialize/deinitialize going on--I have been fiddling trying to get it to work so I'm hoping that's the only reason for all those.

BTW--I have between 3 and 5 other EAs running simultaneously with the IBTrader. I have looked for any possible interaction but have yet to see any logged anywhere--maybe I'm missing something there. Today I'll try setting up a completely separate demo account and run only IBTrader in it. That may also tell me something.

My hours are a bit off from the regular forum so I may be a bit slow getting back and forth--plus I'm traveling tomorrow (Friday). I'll check into the forum over the weekend if you discover anything earthshaking!

Once again, many thanks for the help--I'd love to see this thing in action!
Attached File(s)
File Type: txt IB Log 011409.txt   9 KB | 531 downloads
  • #103
  • Jan 15, 2009 12:22pm Jan 15, 2009 12:22pm
Quoting jcbwrites
Well--many thanks for the help, 1ka. Attached is the log file for yesterday. I left in the H1H1_IB indicator lines along with the IBTrader lines to show the correlation--they are both seeing the IBs. One thing I have noticed is a lot of initialize/deinitialize going on--I have been fiddling trying to get it to work so I'm hoping that's the only reason for all those.

BTW--I have between 3 and 5 other EAs running simultaneously with the IBTrader. I have looked for any possible interaction but have yet to see any logged anywhere--maybe I'm missing...
Hi jcbwrites
I was looking at your log and see that your "maxbars" is set to 4 and your baroffset is set to 30. I dont no which broker you use, but if they dont use fractional pips your baroffset setting is set quite high. I had the same problem as you with no trades being opened. I changed my baroffset to 3 and increased my maxbars setting and the EA seems to be working good now for me. Not sure if that is your problem, but thought I would add my 2 cents worth. Hope that helps

Ray
  • #104
  • Jan 15, 2009 12:28pm Jan 15, 2009 12:28pm
Quoting sparkky2thou
Hi jcbwrites
I was looking at your log and see that your "maxbars" is set to 4 and your baroffset is set to 30. I dont no which broker you use, but if they dont use fractional pips your baroffset setting is set quite high. I had the same problem as you with no trades being opened. I changed my baroffset to 3 and increased my maxbars setting and the EA seems to be working good now for me. Not sure if that is your problem, but thought I would add my 2 cents worth. Hope that helps

Ray
Hey--thanks Sparky--have no idea what any of these inputs do so I just took them as they were set in the EA--I'll give it a try
  • #105
  • Jan 15, 2009 12:57pm Jan 15, 2009 12:57pm
Quoting jcbwrites
Hey--thanks Sparky--have no idea what any of these inputs do so I just took them as they were set in the EA--I'll give it a try
Er, <has quick look at code>

MaxBars is the number of bars the EA will 'go-back' to find an IB. Since the count starts from the actual bar that started the IB (the leftmost bar) then it needs to be at least 3 before anything will work (and yes, I need to change this so it starts from 1 and not 3 :-) ). Six is a good figure as it allows up to four bars to be considered for the breakout.

BarOffset is the number of pips that are added to the high and low of the IB. The breakout and subsequent order will only happen when the price crosses this value. If you are on a 4-digit account, you should change this to 3. I suspect this is the problem you can see.

Good spot by Sparky, thanks.

====

Just checked your log and this is indeed the problem. Your point is 0.0001 rather than my brokers which is 0.00001.

06:04:33 IBTrader EURUSD,M30: IBTrader has been initialised. Point is 0.00010000, Spread is 2, Stop Level is 3
  • #106
  • Jan 15, 2009 1:02pm Jan 15, 2009 1:02pm
Quoting 1ka
The pure IB trading method suggests that SL is moved to the BE point after taking 1/2 profit at 1/2 ATR.

Up until now, the EA has been moving the SL to the opening price at this point. Does giving the price more room to move improve the chances of catching the trend again after a small reversal? Well the figures are as follows:

EURUSD 15M, ZERO SL, 3001 trades, 2522 wins (84%), -£460.96
EURUSD 15M, BE SL, 2962 trades, 1315 wins (70%), -£451.00

[font=Courier New]EURUSD 30M, ZERO SL, 1765 trades,...
Thanks 1ka for the update EA doing .during 2nd half means pa already in a trend & hence there should be good room for the movements & finding proper sl pair specific is not any good job.After over 1 years of trading still I am very unsuccessfull in using trading with correct sl & hence to take more risk.

anyway what I am doing from the day I look into the insider bar from dibs thread are :
just trying to recognise one IB
1. Probable IB: As new current bar forming & within the ranges previous bar (obviously is greater than new) though should not be correct one as still to close it & by this time it may not be any IB .Such forms can be meet any time frames & almost every few minutes.Just place 1 buy stop above 3pips +spread and one sell stop below 3pips +spread from the high & low values of the previous bar & not sl. Once triggered any after some profit set sl very close to the price order >1pips above /below to confirm the profit or I activate swiss_army EA to do that> when certain profit usually I close it or keep on ts and or tp.Intween If the price go back the triggered price I simply close it .The other pending order also delete after one triggered .so free trade without any or minimum risk.So far used in M5 to m30 tfs, anytime & all with profit .here only scalping for minimum profit 100% sitting before computer& acting fast without much lust for profit & no follow up of the trade as they may be much profitable.
2.Pure IB : After one good size bar there are 1 or multiple insider bars then also do as above & in that case open pending orders in relation to the previous good size bar and usually for the trades for bigger profit .

So far this works very good.
Just for exchange of views trading insider bar.
fontu
Not understand?better not follow my analyses!doubt clearer first, risks ++
  • #107
  • Edited Jan 16, 2009 12:54am Jan 15, 2009 1:31pm | Edited Jan 16, 2009 12:54am
For those folks who are trying this out, then here is the latest EA. Nothing much has changed except I've tweaked trend trailing and turned it on by default.

Note that when a trend is moving it will set the SL to 'chase' the trend and push the TP out. The idea is that as long as it's trending it will continue onwards until it reverses and hits the SL.

This will result in many OrderModify calls and I'm not sure how brokers respond to that (do they care?).

A potential side effect, and I'm guessing about this, is that if the MetaTrader client only allows a single request to be in action at any one time then these calls may fail, as will any other requests, such as closing an order. I suspect these calls are queued and actioned in order, so it may be OK.

I need to read up on this and see what the effects are.


==========

Oh, and DON'T use this on a live account. It WILL lose you money
Attached File(s)
File Type: ex4 IBTrader.ex4   12 KB | 386 downloads
  • #108
  • Jan 15, 2009 1:35pm Jan 15, 2009 1:35pm
Quoting 1ka
The pure IB method uses a 30 minute timescale. What effect do other timescales have?

Running the EA we get:

EURUSD 1M 1267 trades, 912 wins (71%), -£134.85
EURUSD 5M 2694 trades, 2109 wins (78%), -£391.19
EURUSD 15M 2179 trades, 1743 wins (79%), -£393.72
EURUSD 30M 783 trades, 658 wins (84%), £57.19
EURUSD 1H 880 trades, 716 wins (81%), -£189.15
EURUSD 4H 282 trades, 117 wins (84%), £223.01
[FONT=Courier...
unless u used SL = TP, otherwise, your 80% win ratio is not meaningful when you consider multiple time frame.

in other words, 4h's 80% win is very different from 80% win in 30m, in terms of both price action and number of pips.

just my 1 cent.

p.s. therefore, i think the longer TF, the more significant the IB is (more likely to win.....)
  • #109
  • Jan 15, 2009 1:37pm Jan 15, 2009 1:37pm
Quoting fontu
So far this works very good.
Just for exchange of views trading insider bar.
fontu
Thanks fontu, it's good to read how other people are trading with IBs.
  • #110
  • Jan 15, 2009 1:39pm Jan 15, 2009 1:39pm
Quoting 1ka
For those folks who are trying this out, then here is the latest EA....
Thank you for all your hard work and willingness to share. I am running this EA on a demo account at the moment, but it has only been for a couple days. I will share my results with you when I have a few more days of history.
  • #111
  • Edited 2:01pm Jan 15, 2009 1:49pm | Edited 2:01pm
Quoting charvo

just my 1 cent.
Every cent is welcome 99,999,999 more and I'll be a $ millionaire

The TP values are based around the average bar size, so will be relative to the timeframe they are on. They will naturally be larger on the higher timeframes, as will the SL.

Even so, the EA does not show any major change in the win ratio at higher timeframes. What I do expect to change is that when we start to use trending indicators these will be more accurate at higher timeframes so result in less losing trades and increasing the win ratio.
  • #112
  • Jan 15, 2009 2:36pm Jan 15, 2009 2:36pm
There have been posts in the forum and the chat room about using the RSI indicator to improve the win ratio or the profit taken. The idea is that if the RSI is above 50 then we only Buy and if it's below 50 we Sell.

I ran a number of tests with RSI periods of 8, 13, 18 & 23. I also looked at introducing bands 60/40 and 70/30 rather than just 50/50.

To me the results are inconclusive. I can see some subtle changes in win ratios and some changes in profit/loss. However, I can't see an overriding trend that would suggest the RSI indicator is weeding out only bad trades.

Attached are some grabs to show the results.
Attached Images
  • #113
  • Jan 15, 2009 3:20pm Jan 15, 2009 3:20pm
Quoting 1ka
So to recap.

........

  1. TP setting does matter, an initial TP at 1/2 ATR gets more consistent wins.

Have you ever tried to use TP of 1/2 size of outside bar?

for example, the set-up has 2 bars, 1 inside bar (the later bar), the other is the "outside bar" (the earlier bar). if the outside bar is 50 pips high (High()-Low()), then use 50*1/2 = 25 pips as the TP.

the reason behind is that often big IB bar breakout is led by a long "outside bar" movement that's often followed by a big retrace or further big move.
  • #114
  • Jan 15, 2009 4:18pm Jan 15, 2009 4:18pm
Quoting charvo
Have you ever tried to use TP of 1/2 size of outside bar?

for example, the set-up has 2 bars, 1 inside bar (the later bar), the other is the "outside bar" (the earlier bar). if the outside bar is 50 pips high (High()-Low()), then use 50*1/2 = 25 pips as the TP.

the reason behind is that often big IB bar breakout is led by a long "outside bar" movement that's often followed by a big retrace or further big move.
Hi charvo,

I hear what you are saying and it might be worth explicitly trying that at some point although I think the results will be similar to the ATR we use. With the ATR we're averaging the height (High-Low) of the last 20 bars. I ran a test a while back that showed initial TP at 0.5 ATR was better than initial TP at full ATR (better meaning a higher win ratio). The pure IB thread has followed suit and also recommends 0.5 ATR. All this means though is that an average breakout continues for half a bar before retracement to the SL.

Thanks and keep the ideas coming.
  • #115
  • Jan 15, 2009 4:39pm Jan 15, 2009 4:39pm
26 wins from 31 trades across various currencies (84%). I'd like to post the results but can't attach HTML. If people are interested I could zip them but this not as convenient.

Actually, when we take 1/2 profit the trade splits and is counted as two trades. This is in keeping with the strategy tester which reports trades the same way.

Whilst it looks like it is inflating the win ratio, it's probably the right thing to do as the second trade could fail (although we have set a SL to avoid this). We ran some tests recently where the SL was at the trade BE and saw some of the second half trades report a loss and lower the win ratio.
  • #116
  • Jan 15, 2009 5:03pm Jan 15, 2009 5:03pm
Quoting 1ka
For those folks who are trying this out, then here is the latest EA....
1ka--finally got the IBTrader working on usdchf but only one trade so far. I changed the maxbars to 5 and the baroffset to 3 at the suggestion of one other user. Still not working on the eurusd. I'll try this one and see what happens next week. Thanks for all your hard work--this has a lot of potential.
  • #117
  • Jan 15, 2009 5:10pm Jan 15, 2009 5:10pm
Here are the results for this whole day by the default settings, (previous version).
Attached Image (click to enlarge)
Click to Enlarge Name: Anto.jpg Size: 134 KB
  • #118
  • Jan 15, 2009 5:12pm Jan 15, 2009 5:12pm
Quoting 1ka
26 wins from 31 trades across various currencies (84%). I'd like to post the results but can't attach HTML. If people are interested I could zip them but this not as convenient.

Actually, when we take 1/2 profit the trade splits and is counted as two trades. This is in keeping with the strategy tester which reports trades the same way.

Whilst it looks like it is inflating the win ratio, it's probably the right thing to do as the second trade could fail (although we have set a SL to avoid this). We ran some tests recently where the SL was at...

Can you please post your results and currencies
  • #119
  • Jan 16, 2009 12:53am Jan 16, 2009 12:53am
Quoting anto
Can you please post your results and currencies
As requested. Take these with a pinch of salt, as the saying goes.

A bug meant that some trades were not getting closed after setting the SL to zero BE. This was fixed in the afternoon.

Trend trailing was switched on in the evening.
Attached File(s)
File Type: zip Statement.htm.zip   3 KB | 342 downloads
  • #120
  • Jan 16, 2009 1:00am Jan 16, 2009 1:00am
Quoting jcbwrites
--this has a lot of potential.
It does, but only if we can weed out those losing trades. They lose big time (in comparison) and wipe out the profit.

If anyone can find a consistent method that identifies these bad trades, without affecting the good ones too much, then post here.

Thanks.
1 trader viewing now