| stevehopwoodforex.com https://www.stevehopwoodforex.com/phpBB3/ Print view |
|
| Preemptive Zig Zig https://www.stevehopwoodforex.com/phpBB3/viewtopic.php?t=2792 |
Page 1 of 4 |
| Author: | snailbeard [ Wed Jul 31, 2013 11:15 am ] |
| Post subject: | Preemptive Zig Zig |
During many hours of staring at charts on different time frames, I concur that behind all price movements, even going back decades are Fibonacci ratios and sequences, some of these could be behavioural psychology, but despite that man-made belief, is there truly a natural pattern on long time frames? More importantly is it consistent enough to generate likely reversals and targets? Channels are often easy to spot and also candles whose shape allows use to draw vectors into the future. The challenge has been to percolate information down the time frames. The aim is not to predict the future, more to know where the schisms are, meaning critical levels which are useful for entries and reversals. In some cases the signs on a higher time frame are so strong that there is no effort in understanding what is likely to happen. So the aim is to enter when there is a minimum risk of reversal and maximum chance of continuance. However, all though this is all about a manual strategy, it is more about rules which don't require complex layers of counter and counter/counter rules. The basic philosophy is: Identify where pull-backs and reversals occur by means of Fibonacci ratios and channel boundaries, starting on weekly charts and working down the levels. Let a price movement occur that starts from a 'schism' and confirms a particular direction: Enter a trade and partial close in parts (2 to 5 levels) Move SL to breakeven at the first TP, take some profit. SL is fairly close to the schism which allows for larger lot sizes. So far there is nothing new and many people trade like this already, and the trick will be to feed this down to a time frame where we accurately monitor a schism to within 10 or 20 pips for continuation or reversal, get a large percentage of winning trades. Many people use pivot levels and sometime pivots levels coincide with underlying levels, but a better understanding of the stronger schisms is more useful than the pivot levels. On the higher timeframe we have a very clear idea of market sentiment which we need to be aware of at the 5M and 15M levels because we want to time our entries so well that losses are minimised. In the attached chart we see the big picture very quickly: |
|
| Author: | snailbeard [ Wed Jul 31, 2013 11:21 am ] |
| Post subject: | Re: Preemptive Zig Zig |
The next step is to quantify what we see in simple terms to be accessible on the daily timeframe, then to merge what we see on daily for use on H4 and so on down to one minute, where we look for the W or M or other pattern which will be a signal to open or close a trade or partially close a trade. Finally we should have a shortlist of super support and resistance. We should also have a set of swing vectors which identify the origin of the current swing on different timeframes. We have to know when price has pulled back before we can safely enter a trade in order to avoid starting out into the face of an immediate pull back. I have been able to do this manually on the M1 timeframe at pivots by properly understanding the higher time frames. |
|
| Author: | snailbeard [ Wed Jul 31, 2013 4:00 pm ] |
| Post subject: | Re: Preemptive Zig Zig |
Monthly Fibo channels Finding the correct Fibonacci levels is harder than bunging on the pivot indicator, but with a bit of hunting we can find coincident levels or super support and resistance. Perhaps our software can do the hunting for us. On the attached chart I have gone up to monthly bars to make it easier, there are even super support and resistance levels going back to the 1990s. There are vectors projecting price 3 years in the future. The price is random on small scale charts but geometric patterns constrain the price or pull back into some 'prefered' geometric path. |
|
| Author: | snailbeard [ Wed Jul 31, 2013 5:41 pm ] |
| Post subject: | Re: Preemptive Zig Zig |
After studying the higher time and seeing the price has moved by a typical swing, we can sit on the M5 time frame and watch the drama unfold, at UK lunch time the US is waking up and about to make some moves. On the daily there is quite a bit of ranging and zig-zags help reveal some trend lines forming a triangle and price is moving towards a triangle support so a) look for a bounce b) look for a break out south However, on zooming in to M5 we see price movement with three phases. Since the price has recently swung south we are less enthusiastic about entering in the same direction. It is also the last day of the month and perhaps time to settle accounts. A channel line provides support. On H4 we see a swing of over 500 pips and we are very sceptical about selling. On H4 we also observe the current bar closing outside the Bollinger, which is another reason to be cautious about selling and more positive about buying. Back to M5 and we watch a zig zag forming a 3-phase completion below the M5 Bollinger. So we have both H4 and M5 with price below the Bollinger. There is a high probability of at least M5 reversal possibly extending into other time frames. However, H4 is steep so downward momentum is likely result in another bump from the bottom of H4 Bollinger - it is quite possible that there will be rejection at WS1 with chance to grab some pips back towards the bottom Bollinger. This is very simple to explain but can we translate this into an uncomplicated set of rules? The screenshot show only the M5 view as we sit and twiddle thumbs: |
|
| Author: | snailbeard [ Thu Aug 01, 2013 4:23 am ] |
| Post subject: | Re: Preemptive Zig Zig |
Its a new day and watching how GBPJPY price action unfolds - we started a potential reversal yesterday and taking a look at the daily chart: there are obvious channel lines and strong support. There are railroad tracks crossing back into the Bollinger bands. The overall daily outlook is price action upwards towards the upper channel. |
|
| Author: | snailbeard [ Wed Aug 07, 2013 12:25 pm ] |
| Post subject: | Re: Preemptive Zig Zig |
Using the historic zig zag could be applied to other strategies: After adding a multi-pair zig-zag tracker and putting it in to one EA which was about to open a sell position: we can look at the distance between the previous high and low: 12:25:17 2013.07.02 00:00 Snailbeard2.3.2 GBPUSD,H1: H4: ZZnearestHighBarID: 28, dPriceAtZZHigh: 1.54358, dZZHighToPrice: -0.02206 12:25:17 2013.07.02 00:00 Snailbeard2.3.2 GBPUSD,H1: H4: ZZnearestLowBarID: 9, dPriceAtZZLow: 1.52501, dZZLowToPrice: -0.00349 In terms Fibonacci, we expect pull backs of about 23%, 38%, 50%, 77% on certain timeframes but we also have swings between certain weekly pivots. We can speculate that since the price has moved 220 pips since the last significant pull back, unless we are in a powerful move then we should NOT continue to sell until we get the pull back. Another example of a losing trade which could have been avoided: 12:25:17 2013.07.09 00:00 Snailbeard2.3.2 GBPUSD,H1: ZZnearestHighBarID: 21, dPriceAtZZHigh: 1.52718, dZZHighToPrice: -0.03203 12:25:17 2013.07.09 00:00 Snailbeard2.3.2 GBPUSD,H1: ZZnearestLowBarID: 10, dPriceAtZZLow: 1.49927, dZZLowToPrice: -0.00412 12:25:17 2013.07.09 00:00 Snailbeard2.3.2 GBPUSD,H1: open #2 sell stop 0.19 GBPUSD at 1.48623 sl: 1.49623 tp: 1.47745 ok In this instance a main swing of 320 pips has completed with an insufficient retrace of just 41 pips. There is a need to parameterise the thresholds for the major and minor swings. |
|
| Author: | snailbeard [ Thu Aug 15, 2013 4:24 pm ] |
| Post subject: | Re: Preemptive Zig Zig |
What does a successful strategy require? 1) There has to a trigger event examples: a price crossing a threshold: this could be a pivot level or Bollinger line or a derivative of price (e.g a moving average) crossing something else. There are lots of strategies which stop here. 2) We need a qualifier or several qualifiers Lets say we have decided to go for price crossing a pivot level as our primary event Then our qualifier needs to be checked at the time price crosses a threshold. If things a moving quickly then it would better to work the other way round, which means we have to determine our qualifier has a bias for a particular direction. However, there are two ways that we can qualify a trigger event: 1) Positive reinforcement 2) Negative inhibit During my experiments with NB10 I learned that we have some simple logic to get us into a trade but not enough qualifiers to inhibit bad trades. The human visual cortex is so powerful that we can quickly see several coincident properties which would suggest to us a reversal when simple rules propose a trade in same direction. However, to express this human experience in logic turns out to be either very long-winded or computationally expensive. In addition, we can end with a dozen rules to inhibit a trade and thereby end up missing a high proportion of valid trades. So I am currently experimenting with a multi-time-frame qualifier: Firstly, there is the dumb-ass trigger event based on M1 to M15/M30. Secondly there is higher time frame bias/ inhibitor. For simplicity, I am using 3 to 4 Fast moving averages to make a primary direction event. Then, I have a pre-analysed M15 & H1 Bollinger perspective: If there is a recent high price outside M15 & H1 upper Bollinger then trading up is inhibited and vice versa for the opposite Bollinger. However, if the trigger event is down then we accept. Exit strategy Often underestimated is the importance of following a trade into profit and preventing it reversing into a loss. I have been experimenting with partial exits: My understanding in this area could be flawed as I have not had much success so far, it does not work well with NB10, the best results (NB10) were with larger stop-losses and smaller lot sizes, This becomes obvious when you understand the mechanics of the strategy. The partial exit might work better if the entry method is closer the recent price barrier so that reversals in price are less strong than with other strategies. However, this an area for much experimentation. In the next post I look at the dumb-ass trigger events. |
|
| Author: | snailbeard [ Thu Aug 15, 2013 4:53 pm ] |
| Post subject: | Re: Preemptive Zig Zig |
In this chart there is a 4x timeframe moving average trend indicator. The dumb-event rule to enter a trade a) All vertical dots turn to the same color after there was at least one dot of the opposite colour. b) We could also consider using the same logic to exit a running trade We have the option of just using 3 timeframes say M1, M5 and M15 There are occasions when this trigger event is the start of a big long swing, with no significant reversal, but most of the time it is just noise. The last vertical red line is the most interesting line and experienced traders can visualise a couple of trend lines linking the highs and/or lows. The imaginary triangle formed by these imaginary trend lines are used by break-out traders. Observe that the Bollinger contracts to a narrow tube and the price never reverses above the bend. Such types of entries with tighter stop losses mean bigger bet sizes or better win to loss ratios! This is a precursor to a move of over 100 pips which never spikes back across the entry point. Using an early move to break even and say 20 pip partial exits is an alternative to a trailing stop-loss or profit locking approach. However, we can only catch this kind of move by understanding the situation on higher time frames. If we only view the higher time frame our entry logic is lagging. If we only look at the M1 chart we don't know why this could be the right trigger. |
|
| Author: | snailbeard [ Thu Aug 15, 2013 8:21 pm ] |
| Post subject: | Re: Preemptive Zig Zig |
For this M15 chart, I knocked up a version of the Bollinger Indicator to show the H1 Bollinger on the M15 chart. The trend line comes from linking zig-zag peeks on a higher time frame (H1). The Zig-Zig trend lines often form important channel boundaries. If we restrict our view to current M15 bar and ignore the recent history, we can say: a) Candle is sitting below M15 Mid Bollinger (suggests downward bias) b) Candle is closer to H1 upper B than lower B (?which direction is price swinging?) c) Candle is not outside M15 or H1 Bollinger (Not a potential zig-zag candle) Conclusion We are probably inside a swing, but we don't know where More information is required: Locate the recent high/low by stepping backwards: After stepping back 3 H1 bars we find a high-close outside the top Bollinger ( a potential swing event) We can do more by back-stepping bars until we find the corresponding low zig zag From this, we can calculate the price moved in the previous swing (about 90 pips) We can also calculate the emerging swing (20 pips so far in a downward direction). Note 1: In a strong trend price can ride outside the Bollinger Note 2: H4 is Red (downward therefore not a strong upward trend) Note 3: Current bar is well inside the Bollinger so not a strong trend Conclusion Inhibit upward trades, as the swing is down Enable downward trades because price has moved down from zig zag (Price has moved back into the Bollinger from an extreme). Now the question arises: Do we wait for a trigger event then do back-stepping analysis or do we do bar by bar analysis and store results for x-bars? What kind of data can we make quickly available? Starting with the zig-zag - the general algorithm for locating zig-zags requires back-stepping anyway, so lets wait until we have a trigger event. What do we need to know about most recent and previous zig-zag? a) high or low swing b) price value c) time or number of back bars or both since current bar d) pip swing between each pair of zig-zags e) is it a full zig-zag (outside Bollinger) or a minor zig-zag (inside or at the mid Bollinger) Note: in a strong trend the reversal can be short-lived |
|
| Author: | snailbeard [ Thu Sep 26, 2013 11:33 am ] |
| Post subject: | Advanced MTF BB & ZZ & BBZZRP |
Moving on to a more advanced analysis, I am now combining Bollinger Bands and Zig Zags from four time frames. Further to this, my EA does some 'Relative Position' analysis, which stores the relationship between the points of the zig-zag and where they are relative to the BB. I have modified the BB and ZZ indicators so that I can see the higher timeframes on an M15 chart. There are two main aspects to the analysis: a) Do the last 3 or 4 points of a zig-zag have a bias up or down? - meaning higher-highs or lower-lows or ranging b) Do the Bollinger Bands act as support or resistance and conversely does a crossing of a BB indicate an event of significance, but by crossing we might mean the body of a candle is fully contained on one side of a band (pre-trigger analysis) as opposed to a trigger event which is price movement across a threshold in a particular direction. So the difficulty with Zig-Zags is that the most recent point is likely to be recalculated, therefore it is better not analyse a zig-zag until several bars have passed but not enough to be looking at new zig-zag point. In the following chart we have cases of both a reversal and a break-out at Bollinger bands so we look for clues then try to identify a bias in the market. At this zoom, we cannot see the Daily Bollinger Band or the Daily ZigZag. Also we can only see the bottom half of the H4 and H1 Bollinger bands. We can see the whole of the M15 Bollinger Bands and Zig-Zag drawn in Black. H1 in White, and H4 in Lime ( D1 lines are in Purple). From a manual trading perspective: price is below the middle of most BBs - therefore general bias is to sell. When the ZZ glances the H4 Mid-BB and falls away we are inclined to enter a sell trade. This is easy to see, but expressing it in computational terms requires a lot more effort. Nevertheless, I am in the process of doing this and have generated alerts based on the relationship a combination of these things. I have marked up where manual rules and EA logic can result in the same entries. In this case we are using H4 BB as main bias or channel. We can include H1 BB & ZZ but to keep it simple I'm just going to say the M15 ZZ is in a range formation, but if you ignore the spikes and look at the close there is a slight downward bias. The pre-trigger set-up 'alert' is when the previous candle is at a mid-Bollinger and the actual trigger would be when the price crosses the line. So the trigger could be a pending order, however, POs have limitations like the price is not far enough away so it would be rejected, so that leaves us with the other option which is essentially a virtual pending order because our script or EA has to check each tick and then decide whether or the price has been crossed. This is very different to the opposite approach which is to pick a timeframe and then work backwards from the current candle open price and the BB ZZ analysis each time. Nevertheless, I am still recalculating most things on 15 minute cycles. With this arrangement we have an alert from an EA and we can choice whether or not to a) manually trade the trigger event b) load a script to trade on a price cross event |
|
| All times are UTC | Page 1 of 4 |
|
Powered by phpBB® Forum Software © phpBB Limited |
|