Band-Zigzag Based Trend FollowerWe defined new method to derive zigzag last month - which is called Channel-Based-Zigzag . This script is an example of one of the use case of this method.
🎲 Trend Following
Defining a trend following method is simple. Basic rule of trend following is Buy High and Sell Low (Yes, you heard it right). To explain further - methodology involve finding an established trend which is flying high and join the trend with proper risk and optimal stop. Once you get into the trade, you will not exit unless there is change in the trend. Or in other words, the parameters which you used to define trend has reversed and the trend is not valid anymore.
Few examples are:
🎯 Using bands
When price breaks out of upper bands (example, Bollinger Band, Keltener Channel, or Donchian Channel), with a pre determined length and multiplier, we can consider the trend to be bullish and similarly when price breaks down the lower band, we can consider the trend to be bearish.
Here are few examples where I have used bands for identifying trend
Band-Based-Supertrend
Donchian-Channel-Trend-Filter
🎯 Using Pivots
Simple logic using zigzag or pivot points is that when price starts making higher highs and higher lows, we can consider this as uptrend. And when price starts making lower highs and lower lows, we can consider this as downtrend. There are few supertrend implementations I have published in the past based on zigzags and pivot points.
Adoptive-Supertrend-Pivots
Zigzag-Supertrend
Drawbacks of both of these methods is that there will be too many fluctuations in both cases unless we increase the reference length. And if we increase the reference length, we will have higher drawdown.
🎲 Band Based Zigzag Method
Band Based Zigzag will help overcome these issues by combining both the methods.
Here we use bands to define our pivot high and pivot low - this makes sure that we are identifying trend only on breakouts as pivots are only formed on breakouts.
Our method also includes pivot ratio to cross over 1.0 to be able to consider it as trend. This means, we are waiting for price also to make new high high or lower low before making the decision on trend. But, this helps us ignore smaller pivot movements due to the usage of bands.
I have also implemented few tricks such as sticky bands (Bands will not contract unless there is breakout) and Adaptive Bands (Band will not expand unless price is moving in the direction of band). This makes the trend following method very robust.
To avoid fakeouts, we also use percentB of high/low in comparison with price retracement to define breakout.
🎲 The indicator
The output of indicator is simple and intuitive to understand.
🎯 Trend Criteria
Uptrend when last confirmed pivot is pivot high and has higher retracement ratio than PercentB of High. Else, considered as downtrend.
Downtrend when last confirmed pivot is pivot low and has higher retracement ratio than PercentB of High. Else, considered as uptrend.
🎯 Settings
Settings allow you to select the band type and parameters used for calculating zigzag and then trend. Also has few options to hide the display.
Trendoscope
Chatterjee CorrelationThis is my first attempt on implementing a statistical method. This problem was given to me by @lejmer (who also helped me later on building more efficient code to achieve this) when we were debating on the need for higher resource allocation to run scripts so it can run longer and faster. The major problem faced by those who want to implement statistics based methods is that they run out of processing time or need to limit the data samples. My point was that such things need be implemented with an algorithm which suits pine instead of trying to port a python code directly. And yes, I am able to demonstrate that by using this implementation of Chatterjee Correlation.
🎲 What is Chatterjee Correlation?
The Chatterjee rank Correlation Coefficient (CCC) is a method developed by Sourav Chatterjee which can be used to study non linear correlation between two series.
Full documentation on the method can be found here:
arxiv.org
In short, the formula which we are implementing here is:
Algorithm can be simplified as follows:
1. Get the ranks of X
2. Get the ranks of Y
3. Sort ranks of Y in the order of X (Lets call this SortedYIndices)
4. Calculate the sum of adjacent Y ranks in SortedYIndices (Lets call it as SumOfAdjacentSortedIndices)
5. And finally the correlation coefficient can be calculated by using simple formula
CCC = 1 - (3*SumOfAdjacentSortedIndices)/(n^2 - 1)
🎲 Looks simple? What is the catch?
Mistake many people do here is that they think in Python/Java/C etc while coding in Pine. This makes code less efficient if it involves arrays and loops. And the simple code may look something like this.
var xArray = array.new()
var yArray = array.new()
array.push(xArray, x)
array.push(yArray, y)
sortX = array.sort_indices(xArray)
sortY = array.sort_indices(yArray)
SumOfAdjacentSortedIndices = 0.0
index = array.get(xSortIndices, 0)
for i=1 to n > 1? n -1 : na
indexNext = array.get(sortX, i)
SumOfAdjacentSortedIndices += math.abs(array.get(sortY, indexNext)-array.get(sortY, index))
index := indexNext
correlation := 1 - 3*SumOfAdjacentSortedIndices/(math.pow(n,2)-1)
But, problem here is the number of loops run. Remember pine executes the code on every bar. There are loops run in array.sort_indices and another loop we are running to calculate SumOfAdjacentSortedIndices. Due to this, chances of program throwing runtime errors due to script running for too long are pretty high. This limits greatly the number of samples against which we can run the study. The options to overcome are
Limit the sample size and calculate only between certain bars - this is not ideal as smaller sets are more likely to yield false or inconsistent results.
Start thinking in pine instead of python and code in such a way that it is optimised for pine. - This is exactly what we have done in the published code.
🎲 How to think in Pine?
In order to think in pine, you should try to eliminate the loops as much as possible. Specially on the data which is continuously growing.
My first thought was that sorting takes lots of time and need to find a better way to sort series - specially when it is a growing data set. Hence, I came up with this library which implements Binary Insertion Sort.
Replacing array.sort_indices with binary insertion sort will greatly reduce the number of loops run on each bar. In binary insertion sort, the array will remain sorted and any item we add, it will keep adding it in the existing sort order so that there is no need to run separate sort. This allows us to work with bigger data sets and can utilise full 20,000 bars for calculation instead of few 100s.
However, last loop where we calculate SumOfAdjacentSortedIndices is not replaceable easily. Hence, we only limit these iterations to certain bars (Even though we use complete sample size). Plots are made for only those bars where the results need to be printed.
🎲 Implementation
Current implementation is limited to few combinations of x and fixed y. But, will be converting this into library soon - which means, programmers can plug any x and y and get the correlation.
Our X here can be
Average volume
ATR
And our Y is distance of price from moving average - which identifies trend.
Thus, the indicator here helps to understand the correlation coefficient between volume and trend OR volatility and trend for given ticker and timeframe. Value closer to 1 means highly correlated and value closer to 0 means least correlated. Please note that this method will not tell how these values are correlated. That is, we will not be able to know if higher volume leads to higher trend or lower trend. But, we can say whether volume impacts trend or not.
Please note that values can differ by great extent for different timeframes. For example, if you look at 1D timeframe, you may get higher value of correlation coefficient whereas lower value for 1m timeframe. This means, volume to trend correlation is higher in 1D timeframe and lower in lower timeframes.
HSupertrendLibrary "HSupertrend"
Supertrend implementation based on harmonic patterns
hsupertrend(zProperties, pProperties, errorPercent, showPatterns, patternColor)
derives supertrend based on harmonic patterns
Parameters:
zProperties : ZigzagProperties containing Zigzag length and source array
pProperties : PatternProperties used for calculation
errorPercent : Error threshold for scanning patterns
showPatterns : Draw identified patterns structure on chart
patternColor : Color of the pattern lines to be drawn
Returns:
ZigzagProperties
ZigzagProperties contains values required for zigzag calculation
Fields:
length : Zigzag length
source : Array containing custom OHLC. If not set, array.from(high, low) is used
PatternProperties
PatternProperties are essential pattern parameters used for calculation of bullish and bearish zones
Fields:
base : Base for calculating entry and stop of pattern. Can be CD, minmax or correction. Default is CD
entryPercent : Distance from D in terms of percent of Base in the direction of pattern
stopPercent : Distance from D in terms of percent of Base in the opposite direction of pattern
useClosePrices : When set uses close price for calculation of supertrend breakout
Harmonic Patterns Based SupertrendExtending the earlier implemented concept of Harmonic-Patterns-Based-Trend-Follower , in this script, lets make it work as supertrend so that it is more easier to operate.
🎲 Process
🎯 Derive Zigzag and scan harmonic patterns for last 5 confirmed pivots
🎯 If a pattern is found, bullish and bearish zones are calculated based on parameter Base
🎯 These bullish and bearish zones act as supertrend based on current trade in progress.
🎯 When in bullish mode, bearish zone will only go up irrespective of new pattern forming new low. Similarly when in bearish mode, bullish zones will only come down - this is done to imitate the standard supertrend behaviour.
🎲 Note
Patterns are not created on latest pivot as last pivot will be unconfirmed and moving. Due to this, patterns appear after certain delay - patterns will not be real time. But, this is expected and does not impact the overall process.
Here are few chart captures to demonstrate how it works.
🎲 Settings
Settings are explained in the screenshot below.
Harmonic Patterns Based Trend FollowerEarlier this week, published an idea on how harmonic patterns can be used for trend following. This script is an attempt to implement the same.
🎲 Process
🎯 Derive Zigzag and scan harmonic patterns for last 5 confirmed pivots
🎯 If a pattern is found, highest point of pattern will become the bullish zone and lower point of the pattern will become bearish zone.
🎯 Since it is trend following method, when price reaches bullish zone, then the trend is considered as bullish and when price reaches bearish zone, the trend is considered as bearish.
🎯 If price does not touch both regions, then trend remains unchanged.
🎯 Bullish and bearish zone will change as and when new patterns are formed.
🎲 Note
Patterns are not created on latest pivot as last pivot will be unconfirmed and moving. Due to this, patterns appear after certain delay - patterns will not be real time. But, this is expected and does not impact the overall process.
When new pattern formed
When price breaks over the zones
🎲 Output
🎯 Patterns formed are drawn in blue coloured lines. Due to pine limitation of max 500 lines, older patterns automatically get deleted when new ones come.
🎯 Bullish Zone and Bearish Zone are plotted in green and red colours and the zone will change whenever new pattern comes along.
🎯 Bar colors are changed according to calculated trend. Trend value can be 1 or -1 based on the current trend. You can also find the value in data window.
🎯 For simplicity purpose, input option for selection of specific patterns are not provided and also pattern names are not displayed on the chart.
Auto Harmonic Pattern - Backtester [Trendoscope]We are finally here with the implementation of backtesting tool for Auto-Harmonic-Pattern-UltimateX .
CAUTION: THIS IS NOT A STRATEGY AND SHOULD NOT BE FOLLOWED BLINDLY. WE ENCOURAGE USERS TO UTILISE THIS AS BACKTESTING TOOL FOR BUILDING THEIR STRATEGY BASED ON HARMONIC PATTERNS
This script is based on our premium indicator - Auto-Harmonic-Pattern-UltimateX . In this script, along with implementation of scanning harmonic patterns, we provide various options via settings which enables users to build their own strategy based on harmonic patterns, use them with custom coded filters, backtest them on various tickers and timeframes.
Harmonic Patterns is concept and we can trade harmonic pattern in many ways. While general interest around harmonic patterns is to find reversal zones and use them for short term swing trades. But, using it along trend following strategies can also be very rewarding. Here is one of the educational idea I shared about using harmonic patterns for trend following. These are just few possibilities where users can explore further on how they want to trade this. The settings of this script are crafted in such a way that it enables users to explore all these possibilities.
🎲 Components
Chart components of this script is lighter compared to Auto Harmonic Pattern - UltimateX. This is because we want to keep lighter interface in order to support seamless execution of emulator. Since pine strategy framework does most of the things such as calculating profitability, keeping track of trades and results etc, display with respect to - "Closed Trade Stats" are removed from this script and "Open Trade Stats" are made lighter.
🎲 Settings
🎯 Trade Settings : Few important settings under this section are
Due to pine limitations, we will not be able to support both long and short in a same setup. Hence, users need to chose either long or short trade setup.
Entry/Base/Target play important role in defining your strategy.
Confluence is another important factor which lets users use multiple patterns at once as confirmation.
🎯 Zigzag Settings : Zigzag settings determine the size of patterns being formed.
Please note that smaller patterns may not yield very good results and larger patterns may take time to complete trade. Similarly higher depth can cause runtime issues. Recursive zigzag option is alternative to deep search algorithm.
🎯 Filters :
Filters enable users to select trades based on specific conditions. Ability to use external filter even allows writing and using custom filters to be used with this algorithm. Here is a video which explains how this can be done. HOW-TO-Use-external-filters
Pattern filters allow users to pick and chose patterns they want to trade. This can be done either individually or based on category
🎯 Alerts :
Apart from strategy specific alerts, the script also implements customisable alerts via pine alert() function. Alerts can be configured to send upon three conditions
When new pattern is created
When an existing pattern updates entry/stop/target due to safe repaint of D (Only happens when Trail Entry Price is selected)
When a pattern in trade closes either due to hitting stop or target
Important Note: Alerts fired via this method may not match the trades shown on chart as trades which are controlled via pine strategy emulator depends on various other factors such as pyramiding.
Alert template is customisable and users can make use of available placeholders to get dynamic data in alerts. Valid placeholders are
{alertType} - Alert type - New/Update/Close
{id} - Pattern Id
{ticker} - Ticker
{timeframe} - Chart timeframe
{price} - Current price
{patterns} - Identified pattern names
{direction} - Direction - Long/Short
{entry} - Entry Price
{stop} - Stop Price
{target} - Target Price
{orderType} - Limit/Stop - applicable for only New and Update types
{status} - Trade status. Valid values are Pending/Cancelled/Stopped/Success
Template is common for all custom alert types. Hence, updating the template will impact all custom alerts - New/Update/Close
{
"alert" : "{alertType}",
"id" : {id},
"ticker" : "{ticker}",
"timeframe" : "{timeframe}",
"price" : {price},
"patterns" : "{patterns}",
"direction" : "{direction}",
"entry" : {entry},
"stop" : {stop},
"target" : {target},
"orderType" : {orderType}
"status" : {status}
}
Here is a video on how to customise the alerts using templates and placeholders - HOW-TO-Customize-Alerts-With-Placeholders
🎯 Miscellaneous :
These are simple settings to control display and backtest bars. If you are running alerts, we suggest turning of Open Trades and Drawings and limit backtest to minimal value in order to improve efficiency of
🎯 Backtest Engine Parameters :
Default settings are optimised for trend following. Users are encouraged to play around with settings and filters to build strategy out of this tool.
Position sizing is not leveraged. Margin settings makes sure that trades cannot exceed capital.
All measures are taken to avoid repainting. Script does not use request.security and real time bars. This drastically reduces the risk of repainting in scripts.
If you are premium user, please select "Bar Magnifier".
Band Based Trend FilterSimilar to RelativeBandwidthFilter , this script is also a simple trend filter which can be used to define your trading zone.
🎲 Concept
On contrary to reversal mindset, we define trend when price hits either side of the band. If close price hits upper band then it is considered as bullish and if close price hits lower band, then it is considered bearish. Further, trend strength is measured in terms of how many times the price hits one side of the band without hitting other side. Hit is counted only if price has touched middle line in between the touches. This way price walks on the bands are considered as just one hit.
🎲 Settings
Settings are minimal and details can be found in the tooltips against each parameters
🎲 Usage
This can be used with your own strategy to filter your trading/non-trading zones based on trend . Script plots a variable called "Trend" - which is not shown on chart pane. But, it is available in the data window. This can be used in another script as external input and apply logic.
Trend values can be
1 : Allow only Long
-1 : Allow only short
0 : Do not allow any trades
Wolfe Strategy [Trendoscope]Hello Everyone,
Wish you all Merry X-Mas and happy new year. Lets start 2023 with fresh new strategy built on Wolfe Indicator. Details of the indicator can be found here
🎲 Wolfe Concept
Wolfe concept is simple. Whenever a wedge is formed, draw a line joining pivot 1 and 4 as shown in the chart below:
For simplicity, we will only consider static value for Target and Stop. But, entry is done based on breaking the triangle. Revised strategy looks something like this:
🎲 Settings
Settings are simple and details of each are provided via tooltips.
Out of these, the most important one is minimum risk reward ratio. If you set lower risk reward threshold then losing few trades may generate more losses than more winning trades. Similarly higher value will filter out most of the trades and may not work efficiently. Default value set to 1 to make sure optimal risk reward is present before placing trade. Also make note that since the entry bar is always moving towards stop, as and when pattern progress, the RR will also increase. Hence, a pattern which is below RR threshold may become good to trade at certain point of time in future.
🎲 Strategy Parameters
Default strategy parameters are initialised via definition. Margins are set to 100 to disable leveraged trades. Appropriate values are chosen for other parameters. These can be altered based on individual strategy and trading plan.
As the strategy concentrates on the single pattern, number of trades generated are comparatively less. But, there is chance to increase the algorithm further to catch more such patterns on larger scale. Will try to work on them in next versions.
🎲 Pine Strategy limitations
Backtest can only be done on one direction as pine strategy cannot have both long and short open trades together. Hence, it is mandatory to chose either long/short trades in settings.
Since pyramiding is limited to 1, there is possibility of a pattern not generating trade even though the entry conditions are met. They are just based on pine limitations and not necessarily mean patterns are not good for placing trades.
Channel Based Zigzag [HeWhoMustNotBeNamed]🎲 Concept
Zigzag is built based on the price and number of offset bars. But, in this experiment, we build zigzag based on different bands such as Bollinger Band, Keltner Channel and Donchian Channel. The process is simple:
🎯 Derive bands based on input parameters
🎯 High of a bar is considered as pivot high only if the high price is above or equal to upper band.
🎯 Similarly low of a bar is considered as pivot low only if low price is below or equal to lower band.
🎯 Adding the pivot high/low follows same logic as that of regular zigzag where pivot high is always followed by pivot low and vice versa.
🎯 If the new pivot added is of same direction as that of last pivot, then both pivots are compared with each other and only the extreme one is kept. (Highest in case of pivot high and lowest in case of pivot low)
🎯 If a bar has both pivot high and pivot low - pivot with same direction as previous pivot is added to the list first before adding the pivot with opposite direction.
🎲 Use Cases
Can be used for pattern recognition algorithms instead of standard zigzag. This will help derive patterns which are relative to bands and channels.
Example: John Bollinger explains how to manually scan double tap using Bollinger Bands in this video: www.youtube.com This modified zigzag base can be used to achieve the same using algorithmic means.
🎲 Settings
Few simple configurations which will let you select the band properties. Notice that there is no zigzag length here. All the calculations depend on the bands.
With bands display, indicator looks something like this
Note that pivots do not always represent highest/lowest prices. They represent highest/lowest price relative to bands.
As mentioned many times, application of zigzag is not for buying at lower price and selling at higher price. It is mainly used for pattern recognition either manually or via algorithms. Lets build new Harmonic, Chart patterns, Trend Lines using the new zigzag?
BinaryInsertionSortLibrary "BinaryInsertionSort"
Library containing functions which can help create sorted array based on binary insertion sort.
This sorting will be quicker than array.sort function if the sorting needs to be done on every bar and the size of the array is comparatively big.
This is created with the intention of using this to solve a bigger problem posted by @lejmer. Wish me luck!!
binary_insertion_sort(sortedArray, item, order)
binary insertion sort - inserts item into sorted array while maintaining sort order
Parameters:
sortedArray : array which is assumed to be sorted in the requested order
item : float|int item which needs to be inserted into sorted array
order : Sort order - positive number means ascending order whereas negative number represents descending order
Returns: int index at which the item is inserted into sorted array
update_sort_indices(sortIndices, newItemIndex)
adds the sort index of new item added to sorted array and also updates existing sort indices.
Parameters:
sortIndices : array containing sort indices of an array.
newItemIndex : sort index of new item added to sorted array
Returns: void
get_array_of_series(item, order)
Converts series into array and sorted array.
Parameters:
item : float|int series
order : Sort order - positive number means ascending order whereas negative number represents descending order
Returns:
get_sorted_arrays(item, order)
Converts series into array and sorted array. Also calculates the sort order of the value array
Parameters:
item : float|int series
order : Sort order - positive number means ascending order whereas negative number represents descending order
Returns:
Electrocardiogram ChartThis is an attempt to develop alternative visualisation of financial charts. This script also makes use of new pine feature types which represents User Defined Object Types. You can refer to below documentation to understand more about this feature:
www.tradingview.com
www.tradingview.com
🎲 Structure of new chart components
🎯Instead of candles/bars, this type of chart contains Electrocardiogram blocks which resembles the heartbeat signals on electrocardiogram.
Body color of the block is defined by the open and close prices of the bar. If close is greater than open, body is green. Otherwise, the body is painted red.
Border color of the block is defined by the close prices of current and previous bar. If the close of current bar is greater than that of last bar, then the border color is green. Otherwise, border color is painted red.
🎯Inside each blocks there will be 5 connecting lines called the signal lines.
open-open
open-firstPeak(high or low of the bar whichever comes first)
firstPeak-secondPeak(high or low of the bar whichever comes last)
secondPeak-close
close-close
🎯 Color of the signal lines are determined by which among the high/low of the bar comes last. If highest part of the bar reached after reaching the lowest part of the bar, then signal lines are coloured green signifying bullish sentiment towards the end of bar. If lowest part of the bar reached after reaching the highest part of the bar, then signal lines are coloured red signifying bearish sentiment towards the end of bar.
Pictorial examples here:
🎲 Limitations with pinescript implementation
Since, pinescript can only use maximum 500 lines and each block will take 1 box and 5 lines, it is not possible to display more than 100 bars.
Each block of new Electrocardiogram chart will take the space of 7 bars of candlestick chart. Due to this, the alignment of regular OHLC candles is not inline with the new chart type. Background highlighting is done for the part of the OHLC candles where Electrocardiogram blocks are plotted so that it helps users to map the bars manually
Thanks to @theheirophant for suggestion of name :)
taLibrary "ta"
Collection of all custom and enhanced TA indicators. Same as enhanced_ta. But, removed all the displays to make it faster.
ma(source, maType, length)
returns custom moving averages
Parameters:
source : Moving Average Source
maType : Moving Average Type : Can be sma, ema, hma, rma, wma, vwma, swma, highlow, linreg, median
length : Moving Average Length
Returns: moving average for the given type and length
atr(maType, length)
returns ATR with custom moving average
Parameters:
maType : Moving Average Type : Can be sma, ema, hma, rma, wma, vwma, swma, highlow, linreg, median
length : Moving Average Length
Returns: ATR for the given moving average type and length
atrpercent(maType, length)
returns ATR as percentage of close price
Parameters:
maType : Moving Average Type : Can be sma, ema, hma, rma, wma, vwma, swma, highlow, linreg, median
length : Moving Average Length
Returns: ATR as percentage of close price for the given moving average type and length
bb(source, maType, length, multiplier, sticky)
returns Bollinger band for custom moving average
Parameters:
source : Moving Average Source
maType : Moving Average Type : Can be sma, ema, hma, rma, wma, vwma, swma, highlow, linreg, median
length : Moving Average Length
multiplier : Standard Deviation multiplier
sticky : - sticky boundaries which will only change when value is outside boundary.
Returns: Bollinger band with custom moving average for given source, length and multiplier
bbw(source, maType, length, multiplier, sticky)
returns Bollinger bandwidth for custom moving average
Parameters:
source : Moving Average Source
maType : Moving Average Type : Can be sma, ema, hma, rma, wma, vwma, swma, highlow, linreg, median
length : Moving Average Length
multiplier : Standard Deviation multiplier
sticky : - sticky boundaries which will only change when value is outside boundary.
Returns: Bollinger Bandwidth for custom moving average for given source, length and multiplier
bpercentb(source, maType, length, multiplier, sticky)
returns Bollinger Percent B for custom moving average
Parameters:
source : Moving Average Source
maType : Moving Average Type : Can be sma, ema, hma, rma, wma, vwma, swma, highlow, linreg, median
length : Moving Average Length
multiplier : Standard Deviation multiplier
sticky : - sticky boundaries which will only change when value is outside boundary.
Returns: Bollinger Percent B for custom moving average for given source, length and multiplier
kc(source, maType, length, multiplier, useTrueRange, sticky)
returns Keltner Channel for custom moving average
Parameters:
source : Moving Average Source
maType : Moving Average Type : Can be sma, ema, hma, rma, wma, vwma, swma, highlow, linreg, median
length : Moving Average Length
multiplier : Standard Deviation multiplier
useTrueRange : - if set to false, uses high-low.
sticky : - sticky boundaries which will only change when value is outside boundary.
Returns: Keltner Channel for custom moving average for given souce, length and multiplier
kcw(source, maType, length, multiplier, useTrueRange, sticky)
returns Keltner Channel Width with custom moving average
Parameters:
source : Moving Average Source
maType : Moving Average Type : Can be sma, ema, hma, rma, wma, vwma, swma, highlow, linreg, median
length : Moving Average Length
multiplier : Standard Deviation multiplier
useTrueRange : - if set to false, uses high-low.
sticky : - sticky boundaries which will only change when value is outside boundary.
Returns: Keltner Channel Width for custom moving average
kpercentk(source, maType, length, multiplier, useTrueRange, sticky)
returns Keltner Channel Percent K Width with custom moving average
Parameters:
source : Moving Average Source
maType : Moving Average Type : Can be sma, ema, hma, rma, wma, vwma, swma, highlow, linreg, median
length : Moving Average Length
multiplier : Standard Deviation multiplier
useTrueRange : - if set to false, uses high-low.
sticky : - sticky boundaries which will only change when value is outside boundary.
Returns: Keltner Percent K for given moving average, source, length and multiplier
dc(length, useAlternateSource, alternateSource, sticky)
returns Custom Donchian Channel
Parameters:
length : - donchian channel length
useAlternateSource : - Custom source is used only if useAlternateSource is set to true
alternateSource : - Custom source
sticky : - sticky boundaries which will only change when value is outside boundary.
Returns: Donchian channel
dcw(length, useAlternateSource, alternateSource, sticky)
returns Donchian Channel Width
Parameters:
length : - donchian channel length
useAlternateSource : - Custom source is used only if useAlternateSource is set to true
alternateSource : - Custom source
sticky : - sticky boundaries which will only change when value is outside boundary.
Returns: Donchian channel width
dpercentd(useAlternateSource, alternateSource, length, sticky)
returns Donchian Channel Percent of price
Parameters:
useAlternateSource : - Custom source is used only if useAlternateSource is set to true
alternateSource : - Custom source
length : - donchian channel length
sticky : - sticky boundaries which will only change when value is outside boundary.
Returns: Donchian channel Percent D
oscillatorRange(source, method, highlowLength, rangeLength, sticky)
oscillatorRange - returns Custom overbought/oversold areas for an oscillator input
Parameters:
source : - Osillator source such as RSI, COG etc.
method : - Valid values for method are : sma, ema, hma, rma, wma, vwma, swma, highlow, linreg, median
highlowLength : - length on which highlow of the oscillator is calculated
rangeLength : - length used for calculating oversold/overbought range - usually same as oscillator length
sticky : - overbought, oversold levels won't change unless crossed
Returns: Dynamic overbought and oversold range for oscillator input
oscillator(type, length, shortLength, longLength, source, highSource, lowSource, method, highlowLength, sticky)
oscillator - returns Choice of oscillator with custom overbought/oversold range
Parameters:
type : - oscillator type. Valid values : cci, cmo, cog, mfi, roc, rsi, stoch, tsi, wpr
length : - Oscillator length - not used for TSI
shortLength : - shortLength only used for TSI
longLength : - longLength only used for TSI
source : - custom source if required
highSource : - custom high source for stochastic oscillator
lowSource : - custom low source for stochastic oscillator
method : - Valid values for method are : sma, ema, hma, rma, wma, vwma, swma, highlow, linreg, median
highlowLength : - length on which highlow of the oscillator is calculated
sticky : - overbought, oversold levels won't change unless crossed
Returns: Oscillator value along with dynamic overbought and oversold range for oscillator input
multibands(bandType, source, maType, length, useTrueRange, sticky, numberOfBands, multiplierStart, multiplierStep)
multibands - returns Choice of oscillator with custom overbought/oversold range
Parameters:
bandType : - Band type - can be either bb or kc
source : - custom source if required
maType : Moving Average Type : Can be sma, ema, hma, rma, wma, vwma, swma, highlow, linreg, median
length : - Oscillator length - not used for TSI
useTrueRange : - if set to false, uses high-low.
sticky : - for sticky borders which only change upon source crossover/crossunder
numberOfBands : - Number of bands to generate
multiplierStart : - Starting ATR or Standard deviation multiplier for first band
multiplierStep : - Incremental value for multiplier for each band
Returns: array of band values sorted in ascending order
mbandoscillator(bandType, source, maType, length, useTrueRange, stickyBands, numberOfBands, multiplierStart, multiplierStep)
mbandoscillator - Multiband oscillator created on the basis of bands
Parameters:
bandType : - Band type - can be either bb or kc
source : - custom source if required
maType : Moving Average Type : Can be sma, ema, hma, rma, wma, vwma, swma, highlow, linreg, median
length : - Oscillator length - not used for TSI
useTrueRange : - if set to false, uses high-low.
stickyBands : - for sticky borders which only change upon source crossover/crossunder for band detection
numberOfBands : - Number of bands to generate
multiplierStart : - Starting ATR or Standard deviation multiplier for first band
multiplierStep : - Incremental value for multiplier for each band
Returns: oscillator currentStates - Array containing states for last n bars
Next Pivot Projection [Trendoscope]Still experimental. Extending further on the divergence backtest results - in this script we try to project next 2 pivots (including one unconfirmed pivot)
🎲 Previous experiments
1. Divergence-Backtester
2. Divergence-Backtester-V2
🎲 Additions
Apart from collecting the stats on number of occurrences of HH, HL, LH, LL - this script also keeps track of average ratio for each levels and average bars.
Based on these data, we try to calculate the next pivot projections including possible bar and price.
Cloud covering the candles indicate historical levels of average HH, HL, LH, LL projections.
Hover on projection labels to find more details in tooltips.
🎲 Overall method in a nutshell
🎲 Going bit deeper
🎯 Unconfirmed Pivot and its projection - Last pivot of the zigzag is always unconfirmed. Meaning, it can potentially repaint based on further price movements. But, projection of the unconfirmed pivot will not change as it will be based on previous two pivots - both of which are confirmed.
🎯 Next Pivot Projection - Next pivot is projected based on last two pivots - which include last unconfirmed pivot. Hence, these projections can potentially repaint based on the last pivot repaint.
🎯 Historical projections displayed as cloud - Historical projection values are displayed as cloud around pivots.
A cloud above represents area from average lower high range to average higher high range. Cloud color is green if average ratio of pivot high is more than 1. Red Otherwise.
A cloud below represents area from average higher low range to average lower low range. Cloud color is red if average ratio of pivot high is more than 1. Green otherwise
Divergence Backtester - V2Further attempts to study divergence impact on price in shorter terms.
Previous study can be found here:
In this script, we are trying to gather the stats based on last two pivot state together. For example, Individual table of Pivot High Projection is as explained below:
But, by looking at the bigger picture, we can further estimate following things regarding the current unconfirmed pivot and the new pivot which is yet to be formed.
strategyLibrary "strategy"
Library containing few key calculations for strategy involving leveraged limit and stop orders
getQty(entry, stop, riskPercentage)
calculate qty and leverage based on entry and stop price for given risk percentage.
Parameters:
entry : Entry Price
stop : Stop Price
riskPercentage : risk percentage per trade
Returns: - Quantity based on the risk and calculated leverage on position including existing positions
bracketOrder(entry, stop, target, maxLeverage, isLimitOrder, riskPercentage)
Calculates position size based on risk and creates bracket orders for given entry/stop/target
Parameters:
entry : Entry Price
stop : Stop Price
target : Target Price
maxLeverage : Maximum leverage allowed
isLimitOrder : if true, places limit order for entry, else places stop order.
riskPercentage : risk percentage per trade
Returns: orderPlaced - true if orders successfully placed, false otherwise.
order(entry, stop, maxLeverage, isLimitOrder, riskPercentage)
Calculates position size based on risk and creates order for given entry/stop
Parameters:
entry : Entry Price
stop : Stop Price
maxLeverage : Maximum leverage allowed
isLimitOrder : if true, places limit order for entry, else places stop order.
riskPercentage : risk percentage per trade
Returns: orderPlaced - true if orders successfully placed, false otherwise.
Auto Harmonic Pattern - Extreme [Trendoscope]Publishing free to use scripts always brought me more value - be it in terms of more followers, or helped me to engage with people having right ideas. This has helped me immensely over past few years of my presence in tradingview. To all the coders out there who is starting to work on pinescript - my suggestion is to publish as many open source/free to use scripts as possible. It will always be rewarding experience and help you grow in a way you can never imagine.
Today, I realised that I have accumulated 10,000+ followers in tradingview just by learning and publishing what I learn. As token of gratitude to all our followers and supporters, we would like to gift this indicator. This is also tribute to all the people who have positively influenced our presence in tradingview and encouraged us to reach this level. Unfortunately cannot publish it as open source due to having proprietary code in it and fear of being misused.
I will not talk about the subject harmonic patterns as it is very generic and information is available online in many places. You can also refer to my open source libraries and scripts related to harmonic patterns. Some of them are here:
Libraries
eHarmonicpatternsExtended
eHarmonicpatterns
harmonicpatterns
Indicators
Harmonic-Patterns-Custom-Candles
Manual-Harmonic-Patterns-With-interactive-inputs
Manual-Harmonic-Projections-With-interactive-inputs
Auto-Harmonic-Patterns-V2
There are few more among open source and free scripts. But, will not mention them here as they are deprecated.
Now coming to the present indicator.
Setting are made simpler to enhance usability. This can also help new traders learning harmonic patterns. Snapshot of the settings can be found here:
Components of the script are as below
Please note, I may not be able to make any further update to this script. But, I promise never to degrade or reduce the quality of any free to use or open source script I ever published.
Extreme Trend Reversal Points [HeWhoMustNotBeNamed]Using moving average crossover for identifying the change in trend is very common. However, this method can give lots of false signals during the ranging markets. In this algorithm, we try to find the extreme trend by looking at fully aligned multi-level moving averages and only look at moving average crossover when market is in the extreme trend - either bullish or bearish. These points can mean long term downtrend or can also cause a small pullback before trend continuation. In this discussion, we will also check how to handle different scenarios.
🎲 Components
🎯 Recursive Multi Level Moving Averages
Multi level moving average here refers to applying moving average on top of base moving average on multiple levels. For example,
Level 1 SMA = SMA(source, length)
Level 2 SMA = SMA(Level 1 SMA, length)
Level 3 SMA = SMA(Level 2 SMA, length)
..
..
..
Level n SMA = SMA(Level (n-1) SMA, length)
In this script, user can select how many levels of moving averages need to be calculated. This is achieved through " recursive moving average " algorithm. Requirement for building such algorithm was initially raised by @loxx
While I was able to develop them in minimal code with the help of some of the existing libraries built on arrays and matrix , I also thought why not extend this to find something interesting.
Note that since we are using variable levels - we will not be able to plot all the levels of moving average. (This is because plotting cannot be done in the loop). Hence, we are using lines to display the latest moving average levels in front of the last candle. Lines are color coded in such a way that least numbered levels are greener and higher levels are redder.
🎯 Finding the trend and range
Strength of fully aligned moving average is calculated based on position of each level with respect to other levels.
For example, in a complete uptrend, we can find
source > L(1)MA > L(2)MA > L(3)MA ...... > L(n-1)MA > L(n)MA
Similarly in a complete downtrend, we can find
source < L(1)MA < L(2)MA < L(3)MA ...... < L(n-1)MA < L(n)MA
Hence, the strength of trend here is calculated based on relative positions of each levels. Due to this, value of strength can range from 0 to Level*(Level-1)/2
0 represents the complete downtrend
Level*(Level-1)/2 represents the complete uptrend.
Range and Extreme Range are calculated based on the percentile from median. The brackets are defined as per input parameters - Range Percentile and Extreme Range Percentile by using Percentile History as reference length.
Moving average plot is color coded to display the trend strength.
Green - Extreme Bullish
Lime - Bullish
Silver - range
Orange - Bearish
Red - Extreme Bearish
🎯 Finding the trend reversal
Possible trend reversals are when price crosses the moving average while in complete trend with all the moving averages fully aligned. Triangle marks are placed in such locations which can help observe the probable trend reversal points. But, there are possibilities of trend overriding these levels. An example of such thing, we can see here:
In order to overcome this problem, we can employ few techniques.
1. After the signal, wait for trend reversal (moving average plot color to turn silver) before placing your order.
2. Place stop orders on immediate pivot levels or support resistance points instead of opening market order. This way, we can also place an order in the direction of trend. Whichever side the price breaks out, will be the direction to trade.
3. Look for other confirmations such as extremely bullish and bearish candles before placing the orders.
🎯 An example of using stop orders
Let us take this scenario where there is a signal on possible reversal from complete uptrend.
Create a box joining high and low pivots at reasonable distance. You can also chose to add 1 ATR additional distance from pivots.
Use the top of the box as stop-entry for long and bottom as stop-entry for short. The other ends of the box can become stop-losses for each side.
After few bars, we can see that few more signals are plotted but, the price is still within the box. There are some candles which touched the top of the box. But, the candlestick patterns did not represent bullishness on those instances. If you have placed stop orders, these orders would have already filled in. In that case, just wait for position to hit either stop or target.
For bullish side, targets can be placed at certain risk reward levels. In this case, we just use 1:1 for bullish (trend side) and 1:1.5 for bearish side (reversal side)
In this case, price hit the target without any issue:
Wait for next reversal signal to appear before placing another order :)
ahpuhelperLibrary "ahpuhelper"
Helper Library for Auto Harmonic Patterns UltimateX. It is not meaningful for others. This is supposed to be private library. But, publishing it to make sure that I don't delete accidentally. Some functions may be useful for coders.
insert_open_trades_table_column(showOpenTrades, table_id, column, colors, values, intStatus, harmonicTrailingStartState, lblSizeOpenTrades)
add data to open trades table column
Parameters:
showOpenTrades : flag to show open trades table
table_id : Table Id
column : refers to pattern data
colors : backgroud and text color array
values : cell values
intStatus : status as integer
harmonicTrailingStartState : trailing Start state as per configs
lblSizeOpenTrades : text size
Returns: nextColumn
populate_closed_stats(ClosedStatsPosition, bullishCounts, bearishCounts, bullishRetouchCounts, bearishRetouchCounts, bullishSizeMatrix, bearishSizeMatrix, bullishRR, bearishRR, allPatternLabels, flags, rowMain, rowHeaders)
populate closed stats for harmonic patterns
Parameters:
ClosedStatsPosition : Table position for closed stats
bullishCounts : Matrix containing bullish trade stats
bearishCounts : Matrix containing bearish trade stats
bullishRetouchCounts : Matrix containing bullish trade stats for those which retouched entry
bearishRetouchCounts : Matrix containing bearish trade stats for those which retouched entry
bullishSizeMatrix : Matrix containing data about size of bullish patterns
bearishSizeMatrix : Matrix containing data about size of bearish patterns
bullishRR : Matrix containing Risk Reward data of bullish patterns
bearishRR : Matrix containing Risk Reward data of bearish patterns
allPatternLabels : array containing pattern labels
flags : display flags
rowMain : Pattern header data
rowHeaders : header grouping data
Returns: void
get_rr_details(patternTradeDetails, harmonicTrailingStartState, disableTrail, breakEvenTrail)
calculate and return risk reward based on targets and stops
Parameters:
patternTradeDetails : array containing stop, entry and targets
harmonicTrailingStartState : trailing point
disableTrail : If set, ignores trailing point
breakEvenTrail : If set, trailing does not go beyond breakeven.
Returns: nextColumn
Micro ZigzagMicro zigzag is created based on similar concepts as that of zigzag but by using lower timeframe intra-bar data. The lines join candle's high/low points but also depict how the price movement within the candle happened. That is, if the high of the candle is reached first, pivot from previous candle join the high first and then low and vice versa.
The output can also be viewed as advanced line chart.
🎲 Process
🎯 For every bar identify whether high came first or low by using lower timeframe data.
🎯 If high came before low, add high as high pivot first and then low as low pivot. If otherwise, add low as lower pivot first and then high as higher pivot.
🎯 When adding pivot, check if the last pivot is in the same direction as the new one. If yes, replace existing pivot if the new one goes beyond it. Ignore otherwise.
🎯 If the last pivot is of different direction as that one new one, then simple add the new pivot.
SubCandles-V2Further development on the sub-candle concept defined in the earlier script. SubCandle
This time instead of concentrating only on the last part, we are dividing the main candle into three parts.
First Sub-Candle - Covers the candle movement from open to first of highest/lowest point. First sub-candle tells how fast/slow the initial movement took place and which direction it went.
Second Sub-Candle - Is always a full bodied candle. Most important think here is the thickness of the sub-candle which tells how quickly or slowly the price went from one side of the peak to other.
Third Sub-Candle - Similar to first sub-candle but covers the last part. This part is similar to what we are depicting in the earlier version of Sub-Candle script.
Thickness of the sub-candle is based on relatively how much time each sub-candle took to complete within the main candle. Few interpretation can be:
If the first or third sub-candle are the thicker ones and if the thicker sub-candle has long wick, then it may mean strong rejection.
If the first or third sub-candle are thicker ones and if the thicker sub-candle has full body, it may mean higher volatility within the sub-candle.
If the second sub-candle is thicker than the rest - it can mean strong bias towards the direction of second sub-candle
Thinner second sub-candle may mean a dump/pump. May need to check third sub-candle to see if the dump/pump are followed by pump/dump
Also adding the screen shot reference below.
Note: This can repaint within bar. Use it as part of bar replay. Will try to add more option to select a particular candle without bar reply soon
Let me know what you think?
SubCandleI created this script as POC to handle specific cases where not having tick data on historical bars create repainting. Happy to share if this serves purpose for other coders.
What is the function of this script?
Script plots a sub-candle which is remainder of candle after forming the latest peak.
Higher body of Sub-candle refers to strong retracement of price from its latest peak. Color of the sub-candle defines the direction of retracement.
Higher wick of Sub-candle refers to higher push in the direction of original candle. Meaning, after price reaching its peak, price retraced but could not hold.
Here is a screenshot with explanation to visualise the concept:
Settings
There is only one setting which is number of backtest bars. Lower timeframe resolution which is used for calculating the Sub-candle uses this number to automatically calculate maximum possible lower timeframe so that all the required backtest windows are covered without having any issue.
We need to keep in mind that max available lower timeframe bars is 100,000. Hence, with 5000 backtest bars, lower timeframe resolution can be about 20 (100000/5000) times lesser than that of regular chart timeframe. We need to also keep in mind that minimum resolution available as part of security_lower_tf is 1 minute. Hence, it is not advisable to use this script for chart timeframes less than 15 mins.
Application
I have been facing this issue in pattern recognition scripts where patterns are formed using high/low prices but entry and targets are calculated based on the opposite side (low/high). It becomes tricky during extreme bars to identify entry conditions based on just the opposite peak because, the candle might have originated from it before identifying the pattern and might have never reached same peak after forming the pattern. Due to lack of tick data on historical bars, we cannot use close price to measure such conditions. This leads to repaint and few unexpected results. I am intending to use this method to overcome the issue up-to some extent.
Relative Bandwidth FilterThis is a very simple script which can be used as measure to define your trading zones based on volatility.
Concept
This script tries to identify the area of low and high volatility based on comparison between Bandwidth of higher length and ATR of lower length.
Relative Bandwidth = Bandwidth / ATR
Bandwidth can be based on either Bollinger Band, Keltner Channel or Donchian Channel. Length of the bandwidth need to be ideally higher.
ATR is calculated using built in ATR method and ATR length need to be ideally lower than that used for calculating Bandwidth.
Once we got Relative Bandwidth, the next step is to apply Bollinger Band on this to measure how relatively high/low this value is.
Overall - If relative bandwidth is higher, then volatility is comparatively low. If relative bandwidth is lower, then volatility is comparatively high.
Usage
This can be used with your own strategy to filter out your non-trading zones based on volatility. Script plots a variable called "Signal" - which is not shown on chart pane. But, it is available in the data window. This can be used in another script as external input and apply logic.
Signal values can be
1 : Allow only Long
-1 : Allow only short
0 : Do not allow any trades
2 : Allow both Long and Short
drawcandlesLibrary "drawcandles"
simple utility to draw different candles using box and lines. Quite useful for drawing candles such as zigzag candles or MTF candles
draw(o, h, l, c, oBar, cBar)
draws candles based on ohlc values
Parameters:
o : Open Price
h : High Price
l : Low Price
c : Close Price
oBar : Open Time
cBar : Close Time
Returns: void
