Linear
Forecasting - Holt’s Linear Trend ForecastingHolt's Forecasting method
Holt (1957) extended simple exponential smoothing to allow the forecasting of data with a trend. This method involves a forecast equation and two smoothing equations (one for the level and one for the trend):
Forecast equation: ŷ = l + h * b
Level equation: l = alpha * y + (1 - alpha) * (l + b)
Trend equation: b = beta * (l - l) + (1 - beta) * b
where:
l (or l) denotes an estimate of the level of the series at time t,
b (or b) denotes an estimate of the trend (slope) of the series at time t,
alpha is the smoothing parameter for the level, 0 ≤ alpha ≤ 1, and
beta is the smoothing parameter for the trend, 0 ≤ beta ≤ 1.
As with simple exponential smoothing, the level equation here shows that l is a weighted average of observation y and the one-step-ahead training forecast for time t, here given by l+b. The trend equation shows that b is a weighted average of the estimated trend at time t based on l-l and b, the previous estimate of the trend.
The forecast function is not flat but trending. The h-step-ahead forecast is equal to the last estimated level plus h times the last estimated trend value. Hence the forecasts are a linear function of h.
Linear Regression Trend ChannelThis is my first public release of indicator code and my PSv4.0 version of "Linear Regression Channel", as it is more commonly known. It replicates TV's built-in "Linear Regression" without the distraction of heavy red/blue fill bleeding into other indicators. We can't fill() line.new() at this time in Pine Script anyways. I entitled it Linear Regression Trend Channel, simply because it seems more accurate as a proper description. I nicely packaged this to the size of an ordinary napkin within 20 lines of compact code, simplifying the math to the most efficient script I could devise that fits in your pocket. This is commonly what my dense intricate code looks like behind the veil, and if you are wondering why there is no notes, that's because the notation is in the variable naming. I excluded Pearson correlation because it doesn't seem very useful to me, and it would comprise of additional lines of code I would rather avoid in this public release. Pearson correlation is included in my invite-only advanced version of "Enhanced Linear Regression Trend Channel", where I have taken Linear Regression Channeling to another level of fully featured novel attainability using this original source code.
Features List Includes:
"Period" adjustment
"Deviation(s)" adjustment
"Extend Method" option to extend or not extend the upper, medial, and lower channeling
Showcased in the chart below is my free to use "Enhanced Schaff Trend Cycle Indicator", having a common appeal to TV users frequently. If you do have any questions or comments regarding this indicator, I will consider your inquiries, thoughts, and ideas presented below in the comments section, when time provides it. As always, "Like" it if you simply just like it with a proper thumbs up, and also return to my scripts list occasionally for additional postings. Have a profitable future everyone!
Time Series ForecastIntroduction
Forecasting is a blurry science that deal with lot of uncertainty. Most of the time forecasting is made with the assumption that past values can be used to forecast a time series, the accuracy of the forecast depend on the type of time series, the pre-processing applied to it, the forecast model and the parameters of the model.
In tradingview we don't have much forecasting models appart from the linear regression which is definitely not adapted to forecast financial markets, instead we mainly use it as support/resistance indicator. So i wanted to try making a forecasting tool based on the lsma that might provide something at least interesting, i hope you find an use to it.
The Method
Remember that the regression model and the lsma are closely related, both share the same equation ax + b but the lsma will use running parameters while a and b are constants in a linear regression, the last point of the lsma of period p is the last point of the linear regression that fit a line to the price at time p to 1, try to add a linear regression with count = 100 and an lsma of length = 100 and you will see, this is why the lsma is also called "end point moving average".
The forecast of the linear regression is the linear extrapolation of the fitted line, however the proposed indicator forecast is the linear extrapolation between the value of the lsma at time length and the last value of the lsma when short term extrapolation is false, when short term extrapolation is checked the forecast is the linear extrapolation between the lsma value prior to the last point and the last lsma value.
long term extrapolation, length = 1000
short term extrapolation, length = 1000
How To Use
Intervals are create from the running mean absolute error between the price and the lsma. Those intervals can be interpreted as possible support and resistance levels when using long term extrapolation, make sure that the intervals have been priorly tested, this mean the intervals are more significants.
The short term extrapolation is made with the assumption that the price will follow the last two lsma points direction, the forecast tend to become inaccurate during a trend change or when noise affect heavily the lsma.
You can test both method accuracy with the replay mode.
Comparison With The Linear Regression
Both methods share similitudes, but they have different results, lets compare them.
In blue the indicator and in red a linear regression of both period 200, the linear regression is always extremely conservative since she fit a line using the least squares method, at the contrary the indicator is less conservative which can be an advantage as well as a problem.
Conclusion
Linear models are good when what we want to forecast is approximately linear, thats not the case with market price and this is why other methods are used. But the use of the lsma to provide a forecast is still an interesting method that might require further studies.
Thanks for reading !
Linear Trailing StopBased on my latest script "Linear Channels"
This is a trailing stop version of the linear channels. Thanks to capissimo for helping me fix several issues with the linear extrapolation part.
In order to know how the indicator work i recommend reading the post on the Linear Channels indicator here
Hope you like it and feel free to leave your suggestions :)
Linear ChannelsIntroduction
I already made an indicator (simple line) that tried to make lines on price such that the results would not repaint and give a good fit to the price, today i publish a channels indicator based on the simple line indicator. The indicator aim to show possible support and resistance levels when the central line posses a low sum of squares with the price, a linear extrapolation was also provided in order to show possible future price positions respective to the channels.
The Indicator
The emphasis parameter of the simple line indicator has been removed, instead we keep length and mult as numerical input parameters. In general length control how persistent the lines are, larger values will create longer lines on average, mult help make the line fit to the price better but might as well affect how spread the channels are as well as the lines average length. When mult > length the lines will fit better the price while when length >= mult the fit might not be the best.
The point parameter allow you to fix the indicator when using it on high market price values or when the indicator exhibit a weird behaviour.
point = false on btcusd
point = true
If the lines still does not fit well enough try to lower length.
I know this might result inconvenient in so many ways but i'am working on simplifying things. Therefore some larger price values might use lower length and use mult instead. For market not using the point parameters a settings of : length > 1 and mult = length*2 might provide a good to go setup.
The channel spreading parameter allow to make spread the channels by a certain factor.
Issues
I'am still not good with line extensions, if it bother you deactivate the extrapolation parameter. Sorry for the inconvenience.
Conclusion
It is possible to make non repainting linear indicators, and i'am working on some of them. While some might argue that price is not linear thus not requiring the use of linear indicators it can still be interesting to use those if they, unlike the linear regression, don't repaints and provide a way to change their directions according to the price trend.
Thanks for reading !
Trading System(Dark)Combo of many useful indicators, contains
1)Regular and Hidden Divergence Buy and Sell signals by scarf
2)Time and Money channels by Lazybear
3)Fibonacci Bollinger Bands by Rashad
4)Linear Regression Curve by ucsgears
Thanks for all the creators for the source codes!
MACD of Linear Regression Slope Indicator I used MACD to find peak and trough points in the Linear Regression Slope
Linear Tendency FollowerLinear Trend Follower follows 'source' trend using lines within a number of periods ('length') using the last n periods source variation divided by 'length' as line slope. It is delayed by 'length' periods.
Momentum Linear RegressionThe original script was posted on ProRealCode by user Nicolas.
This is an indicator made of the linear regression applied to the rate of change of price (or momentum). I made a simple signal line just by duplicating the first one within a period decay in the past, to make those 2 lines cross. You can add more periods decay to made signal smoother with less false entry.
LWMA w/ Color ChangeLinear Weighted MA that changes colors based on slope.
Green = slope up from last bar
Yellow = slope is 0 from last bar
Red = slope down from last bar
This time with the ability to change the period.
LWMA w/ Color ChangeLinear Weighted MA that changes colors based on slope.
Green = slope up from last bar
Yellow = slope is 0 from last bar
Red = slope down from last bar
Function 2 Point Line using UNIX TIMESTAMP V1experimental:
draws a line from 2 vectors(price, time)
update:
reformatted the function,
added automatic detection of the period multiplier by approximation(gets a bit goofy with stocks/week time),
example using timestamp() function.
offsetting is still bugged, i cant find a way around it atm.
Function Linear Decay V2EXPERIMENTAL:
improved range detection (it now locks range when its last formed on the appropriate side improving readability as it doesnt auto adjust when opposite extreme moves)
[STRATEGY] Follow the Janet YellenIn the era of central bank's helicopter money, the market will always be skyrocketing up and up given enough time.
What's the strategy to profit from indices?
Only short the market when its in a state of euphoria /irrational exuberance bubble, or sell when it is confirmed (20% drawdown). Otherwise, you really have no reason not to long at every chance.
Conclusion:
Follow the printing press like a sheep.
[RS]Decay Channel Candles V0EXPERIMENTAL: Experiment using Linear Regression based on %atr for decay(decay option is a mutiplier for the atr).
[RS]Linear Regression Bands V1experiment with linear regression, the purpose was to catch break outs early, but it creates to much visual noise
same as version 0 but with added margin filter and signal to mark entrys
[RS]Average Advance and Decline Lines V0Method to draw linear regression lines from average price advance&decline range
Linear Regression Slope - Version 2Version 2 - Linear Regression Slope. This version will have more freedom on picking your own length for all the Inputs.
One of the main reason I changed it is because, Slope calculation on transition period was not being computed properly. Because the Version 1, looks back the length assigned, and compute the slope based on two candle readings, could be 10 days apart or 50. That was misleading.
Therefore, I changed it to plot daily slope and Smooth it with an EMA.
Linear Regression Curve -
List of All my Indicators - www.tradingview.com
Linear Regression SlopeCorrected Version, for
VERSION - 2 () updated for Custom Length, and fixed some formula glitches.
UCSgears_Linear Regression SlopeThis is version 1 of the Linear Regression Slope. In ideal world the Linear regression slope values will remain same for any time period length. because the equation is y = mx+b, where m is the slope. All I did here is m = y/x
The Main Purpose of this indicator is to see, if the Trend is accelerating or decelerating.
The first Blue bar will caution when a strong trend is losing strength. I will leave the rest for you to explore.
I picked AAPL again, because it does have both up and down trend, in the recent time.
Mistake in the code
Corrected Version -
