stevehopwoodforex.com
https://www.stevehopwoodforex.com/phpBB3/
Print view

Zigzag Bollinger Band
https://www.stevehopwoodforex.com/phpBB3/viewtopic.php?t=3237		 Page 28 of 28
Author:  Lee [ Tue Dec 30, 2014 9:30 am ]
Post subject:  Zigzag Bollinger Band
Snailbeard,

The USD rally against the EUR in particular is a result of diverging central bank policies. The Fed have now ended QE, recent economic data has been positive, and they will most likely begin a slow process of interest rate normalisation by H2 2015. The EZ (Euro Zone) on the other hand faces deflation and will most likely (analyst opinion) announce Fed style QE in Q1 2015. If this is the case we'll see a further sell off as the EUR will be devalued further.

This is why no indicator, volume, oscillators etc. can predict when price will bottom out and a reversal will begin. It will happen at some point, obviously, but only a comparison of the two economies can give the answers. If the above pans out as analysts expect we could see price, in my opinion, down at 1.15.
Author:  snailbeard [ Sat Jan 03, 2015 5:08 pm ]
Post subject:  Asymmetric Price Momentum
Greetings Traders - and may it be a prosperous 2015 filled with lots of pips!

Just before the holidays I was trying to improve automatic stop-loss and target calculations, however I was side tracked by more analysis of zig-zag reversals. Often different forms of analysis can produce the same result but if the results are opposite then we get the equivalent of analysis paralysis.

In the following GBPUSD M15 chart ZB4 has to determine whether the dominant direction is up or down at a potential break-out at 04:15.

ZB4 already indicated a caution in the log file:

04:15 GBPUSD: : * CAUTION * possible bW1W2FiboTrap

which is a hint that we should wait for a confirmed break-out and not trust a simple break-out.
Due to some tinkering with H4 reversals, ZB4 had cancelled the H4 reversal flags due to a group of rising candles which is generally correct if the longer term momentum is up.

Both H1 zig-zags and M15 zig-zags are stepping up. However the downward momentum is 120 pips in 27 bars and the upward momentum is 30 pips in 27 bars. That is a ratio of 4 to 1 almost exactly. Since the daily momentum is up it would be simple to assume that price has pulled-back to the H4 mid-Bollinger prior to resuming the daily trend and the big boys are happy for us to read it that way:
gbpusd-m15-asymmetric-price-momentum-0415-10Oct2014.png
For the short term: the above 14 hour analysis tells us to wait, and I also extended the longer term analysis to include the strength of the weekly trend (which is down) which is used to give more weight to the H4 zigzag reversal.

Despite all these wonderful tweaks this version of ZB4 is still too cautious relative to buggy yet highly profitable version of late November which is still 35% more profitable on GPBUSD.

However, I would need to compare 10 pairs to make a serious comparison.
Author:  snailbeard [ Sun Jan 04, 2015 9:22 am ]
Post subject:  Fibonacci trap
Following on from above we add Fibonacci levels to the recent downward swing and find that the candles are trapped inside a Fibonacci channel. Price will eventually breakout with momentum and only then will we come to know the next move.
gbpusd-m15-H4-reversal-with-fibonacci-0415-10Oct14.png
In the next chart we see that the false breakout had weak momentum but the real breakout moved out of the channel with increasing momentum:
gbpusd-m15-Fibo-BkOut-with-Mmn-10Oct14.png
Author:  snailbeard [ Tue Jan 06, 2015 9:39 am ]
Post subject:  usdjpy - triangular breakout
I woke up this morning to find ZB4 was selling USDJPY. Long term traders might be holding buy positions on USDJPY but yesterday and today already were several opportunities to sell USDJPY.

The first zigzag-Bollinger method I implemented had problems with triangular channel breakouts. Firstly, real-time zigzags are unreliable and secondly the day trade is just as likely to be a pull back as an advance. Also the slope of a triangle would affect the entry direction. In reality the triangle can often be sloping in the opposite direction to the actual breakout.

The following chart shows how there were several opportunities to short USDJPY:
usdjpy-m15-triangular-fibo-entry-05Jan15.png
The first thing to notice is that this mornings breakout is weaker echo of yesterdays breakout. Also note that the volume is below average. Therefore, keeping in mind the long term outlook for USDJPY we don't want to get stuck going the wrong way.

ZB4 could not enter yesterday due to the long-term momentum. However, I am considering an additional entry method based on a combination of the triangular breakout and Fibonacci confirmation of the breakout.

This close up view of 05 January 2015 shows where the first ZB trader could have entered a trade at 'Ozb' but probably in the long term direction leading to a loss, where as ZB4 has reversal detection to avoid going the wrong way.
usdjpy-m15-close-up-TCBO+FBO-5Jan2015.png
There are often false breakouts from channels so just working out that price has crossed a boundary is a good way to have a lot of losing trades. Watch-out for 'Gurus' selling training courses on trading breakouts.

We need the second stage more than the first and once you understand what is really happening you can remove the triangle. Looking at the bars between Ozb and TCBO notice the M15 bars that close under the Fibonacci 38.2 levels. Both the 14:00 bar and 14:45 close under this level. We could take either of these bars as prelude to a short. However, the later bar is after two rising bars, so better to delay entry by a few minutes to confirm price is moving the right way.
On the following chart we can put a new set of short-range Fibonacci levels which show price stepping down. At 15:05 price has moved out of the previous Fibonacci channel in a downward direction:
USDJPY-M1-entry-timing-1505-05Jan2015.png
Author:  snailbeard [ Tue Jan 06, 2015 11:28 am ]
Post subject:  USDJPY - A closer look at today's breakout
There was a potential breakout this morning at around 0030 - not much use unless you're in an Asian time zone. ZB4 could potentially trade this although I suspect it does not like the proximity of the D1 Bollinger band (purple) and prefers to wait for a confirmed break of this line. ZB4 finally sets-up an entry at 04:15 CET - still too early for most of us and this short-duration short only lasts a few hours with an exit at 07:35 at or very close to the lower H4 Bollinger band (lime green):
usdjpy-m15-close-up-tcbo+fbo+zb4-entry-06Jan2015.png
So this might be considered an 'echo' trade of yesterday's break, but both the set-up and range are reduced in time and pips. The Fibonacci levels and the triangular channel lines complement each other. In both cases there are usually multiple alternative sets of Fibonacci levels and trend lines. So finding the correct set by visual inspection is not difficult but programming multiple searches for channels and matching Fibonacci levels is better done in a high level language and not in MQL and still not trivial.

In the following chart I have been more specific about connecting the triangular channel with the Fibonacci levels this helps to produce a better break-out plan:
USDJPY-M15-detailed-triangle-andFibonacci-lines-05Jan2015.png
Some traders use pending orders to confirm the breakout: we need to allow enough room to confirm the break-out but too much room and we'll miss the boat. ZB4 uses virtual pending orders to keep an eye on the spread in case our broker wants to eat our profits by widening the spreads (when price really starts to move). It is common to put pending orders at a sufficient gap either side of the channel so I have added an alternative rising break at 119.710.

In the next M1 chart we can see that an autotrader cannot afford to wait for the next M15 bar to close, it has to monitor each tick using a VPO (virtual pending order). When the minute bars close down below the 61.8 we might consider the break-out to be confirmed. Also the note the rising volume:
USDJPY-M1-by-M1-timing-0037-06Jan2015.png
Author:  snailbeard [ Mon Jan 12, 2015 7:08 am ]
Post subject:  Mapping PipMasher into ZB4
I spent a couple of days looking at PipCruiser's PipMasher idea and code as I thought it might add additional entry signals. ZB4 suffers from over-cautiousness at pull-backs so the idea was to run this in parallel with ZB4 to get complementary signals. Many people already trade a variation of Pipmasher based on moving averages. Pipmasher computes trends on MN1, W1 and D1 based on candle open and close, unfortunately this does not always reflect accurately the trends on those time-frames and using candle pairs or triplets would give better results.

The following chart shows a long term down trend and in September the D1 candles moved back towards the Middle Bollinger band and then price moves back with the long term trend:
mapping-pipmasher-into-ZB4-Sep2014.png
So pipmasher has the potential to reenter the trend earlier and perhaps ride it for longer. So what does ZB4 think is happening at the time that PM is triggering, this would correspond to about 10am to 11 am.

Firstly what about ZB4's trend flags on different time frames:

iW2Direction: -1, iW1Direction: 1, iD1Ma10Direction: 0

The daily direction is updated regularly and it doesn't reflect the long term trend until 16:00:

GBPUSD_245_02092014.txt:1474: 15:30 GBPUSD: : PBRE iDailyDirection: 1
GBPUSD_245_02092014.txt:1479: 15:45 GBPUSD: : PBRE iDailyDirection: 1
GBPUSD_245_02092014.txt:1489: 16:00 GBPUSD: : PBRE iDailyDirection: -1

So that means ZB4 was looking for buy opportunities until 16:00 and then started looking for sell opportunities - too late in the day.

Although ZB4 is looking for Buy entries in the morning it is constantly blocked:

10:00 WARNING: (iDailyDirection > 0) && bM15H4TrendDown
10:00 PBRE Buy entry dropped because bD1ZzReversingDown || bH4ZzReversingDown : 0, 1
10:00 PBRE Buy entry dropped because bAdxH4Sell
10:00 PBRE Buy entry dropped because (iWpCrossPriceAction < 0): bCross1Wp: 1 , bCross2Wp: 0
10:00 PBRE Buy entry dropped because M15ZzNoBuyFlag
10:00 PBRE Buy entry dropped because 3M15CandleBias DOWN
10:00 PBRE Buy entry dropped because bM15BarFallingBodyTops

So ZB4 has 7 downward indicators but is constrained by the daily direction which is always lagging a reversal. Therefore I just need to reverse the daily direction based a reasonable rule...

After adding an extra function 'updateD1TrendByLastD1andW2()' to recalculate the daily direction:

10:00 GBPUSD: : updateD1TrendByLastD1andW2(): , iW2Direction: -1, iW1Direction: 1, bWpDirUp: 0, bWpDirDown: 1, iOldDailyDirection: 1, iNewDailyDirection: -1

ZB4 gets an updated daily direction according to the D1 re-trending philosophy and should now be able look for sell opportunities or we might turn this into a sell signal.

Since ZB4 already had these flags available it only took an hour of coding and testing to add this small function.
Author:  snailbeard [ Tue Jan 13, 2015 5:50 am ]
Post subject:  ZB4 with Pipnibbler retrend test result
This was a very useful exercise as it forced me visit parts of the code which I haven't looked at for months which uncovered some quirks especially with the stop-loss and take-profit. These values can be adjusted several times but the TP should never be less than the SL and only after calculating the final SL can ZB4 calculate the correct lot size.

For clarity I'll refer to ZB4's version of pipmasher as pipnibbler because the implementations of the pullback-retrend idea are very different and there wont be that much of an overlap with Pipmasher's pending orders.

So ZB4 has a number of entry methods and from the log file we can see which one gives the trigger:

Code: Select all

10:00 GBPUSD: :   bPriceBelow: M15 & H1 LBB: breakout?
10:00 GBPUSD: :  Trade because:   bBreakH4ORFB
10:00 GBPUSD: :  Trade because: (bHasIntraDaySlope && bBreakH4ORFB) : -1
[/size]

For those who have not seen earlier posts: 'bBreakH4ORFB' means the open H4 bar opened then reversed a bit then price went back past the initial open price and price is moving in the direction of the daily-direction.

The following chart shows how the trade unfolds on an M15 chart, the trigger is close to the Pipmasher pending order level:
ZB4-with-pipnibbler-retrend-entry-02Sep14.png
Author:  snailbeard [ Wed Jan 21, 2015 9:13 am ]
Post subject:  Swiss Cheese
So you think I am referring to recent events at the SNB. I'm quite sure if I had been trading any CHF pairs at the time my account could have been wiped out. The daily trend at the time was up for GBPCHF. However, the daily swing was more or less complete with price bumping against the upper D1 Bollinger band so only a badly timed trade would still be going up.

If I understand correctly, as soon as the announcement was made all the big banks bought as many Swiss Francs as they could, and many of these same banks are also liquidity providers, so does this result in a conflict of interest? Should retail investors and brokers be compensated for actions of the big banks?

Was it irresponsible of the SNB to give no prior notice of an important announcement or is it just banks behaving badly again?

So what happens if in the future I hold some buy trades on the Euro and the ECP has an unscheduled announcement that the Euro zone is reverting to national currencies?
I don't have any confidence that the Euro is sustainable, it seems that France, Italy and Spain are likely to destroy it at some point in the future.

I'm sure many people are reflecting on what could have happened to their account balance last week if they were unlucky enough to be holding the wrong trades.

In view of these issues it is worth reflecting on the long term viability of retail trading.
Or perhaps one should never buy a currency which is being kept artificially away from a lower value?
Author:  snailbeard [ Wed Jan 21, 2015 10:25 am ]
Post subject:  FxOpen crappy data
Just another nail in the coffin of us retail traders is unreliable data feeds from brokers.

I have been experimenting with a faster entry method for daily trades based on the previous day and Fibonacci levels. I have often referred to Alpari as 'Fat-spread Alpari' but unfortunately they were also a victim of the banks and now we must lament their passing and perhaps hope for a resurrection in a few months time, despite the poor spreads the data feed quality has improved over time.

On the other hand I have often praised FxOpen for their thin spreads on the positive side they sometimes have zero spreads but on the down side is their crappy data with missing bars which makes a mess of some calculations requiring minute by minute updates. It means I need to write extra code to land on the bar that I was first aiming for.

At this point I would like to acknowledge Nanningbob's long term trading methods, which are immune to crappy data and the shenanigans of banks and brokers. Although I have been negative about very slow moving averages, sometimes this would have saved many NB traders from going against the fundamentals.

Nevertheless, I am still trying to make an EA which produces a higher rate of signals by daily outlook than ZB4 does at the moment.

For reference here is a clean data feed from Alpari UK:
AplariUK-gbpusd-m1-no-missing-M1-bars-03Nov2014.png
Perhaps I should also mention that GlobalPrime's data feed for this same snapshot was also good!
(when they first started I had to post a number of screen shots like the FxOpen one but I have not had any problems recently, I am happy that GP was not affected by the SNB chaos)

And here is the crappy data feed with missing M1 bars from FxOpen:
FxOpen-Missing-Bars-gbpusd-m1-03Nov2014.png
I hope this and similar posts embarrasses them into improving the quality of their data!
Author:  peterfan [ Mon Aug 03, 2015 8:16 am ]
Post subject:  Zigzag Bollinger Band
Thanks for bringing this point up.
All times are UTC Page 28 of 28

Powered by phpBB® Forum Software © phpBB Limited