始めましょう

線形回帰

linreg-areas (like fib system just more accurate) So this system is a different approach to form a daily fib like system. Its based on linear regression and its slope , we calculate the slope of the line and we put it in MTF format that is without security and non repainting (int2=1440 min) As you in graph the coin in this case BTC has cycle from low to high similar to the fibs system. the lowest level is bellow -0.2 , next level is -0.1 ,0,0.1 and above 0.2 is consider the highest with this system its easy to find the highs and the lows as the coin go between the levels either up or down like the fibs system. one can use this together with fibs system in order to make analysis more accurate. also it easy to set signals once the slope go between levels either up or down . another advantage for this system compare to fibs system is the ability to reduce the MTF to lower frame let say 4 hour or even less and then by the same logic to create system of high and lows to the most smaller frames (but if you do that you need in code to change the levels setting numbers so it fit the best ) follow the direction of the slope and you will find the road that you wish:))
Pineスクリプト™ インジケーター
RafaelZioniの投稿
4
24HLRThis indicator is 24 hours Linear regression line . with upper SD and lower SD (can be manipulated ) The idea is that the coin will bounce up or down around the LR line (red) when you go lower frame of 4h up to1 hour charts then you can see more easy the direction of the coin , its potential low or potential high , where the LR lead the way . In this example we look on btc at 4hours chart and see that its in a cross road either to hell or to escape (if break the line going down upward) and here in 1D graph
Pineスクリプト™ インジケーター
RafaelZioniの投稿
7
Regression Line (OLS) w/Residual Bands - Beasley SavagePlots linear regression "line of best fit" for a given source and length. Also plots the standard deviation of residuals as upper and lower bands. For some reason it's not plotting the lines exactly where the values are, so I wouldn't recommend using it as a purely visual indicator, but it could be used in strategies. A quick implementation of a strategy buying when the close crossed over the lower band and selling when close crossed under the upper band resulted in ~60-70% trade accuracy, not taking into account commission/slippage. If anyone knows how to fix the lines not plotting accurately please let me know :)
Pineスクリプト™ インジケーター
BeasleySavageの投稿
Periodic ChannelThis indicator try to create a channel by summing a re-scaled and readapted sinusoidal wave form to the price mean. The length parameter control the speed of the sinusoidal wave form, this parameter is not converted to a sine wave period for allowing a better estimation, higher length's work better but feel free to try shorter periods. The invert parameter invert the sinusoidal wave. Each bands represent possible return points, the higher the band the higher the probability. Inverted sin wave exemple The performance of the indicator is subjective to the main estimation (blue line), select the parameter that best fit the blue line to the price. Best ragards
Pineスクリプト™ インジケーター
alexgroverの投稿
22
Fractal Breakout V2Version 2 of my fractal pattern aid ( Version 1 ). I added a bouncing line between the high and low trend lines, connecting consecutive extreme points. I also chased down a pesky bug in the slope calculation...and for now I have disabled the ability to change resolution basis for extreme detection (e.g. 30m on a 1hr chart). For fun, I added some shading to make it more apparent at a glance what is happening, but if you find it gimmicky, there's an option to turn that off. I am inexperienced with pattern recognition, so please send feedback if you have any ideas that would make this more useful. Thanks! Lemrin
Pineスクリプト™ インジケーター
Lemrinの投稿
11
Fractal BreakoutFirst of all, huge credit to synapticEx , whose brilliant use of the security function inspired me to figure out a way to get quasi-shape boundaries automatically drawn on a chart. This study draws upper and lower trend lines, based on configurable fractal*** reversal detection, calculates slope from the last two upper or lower reversal points, and then extends a dotted line along the same slope...until the next upper (or lower) reversal occurs. If the high (or low) breaks this extension, the dotted line becomes solid to aid visibility. Reversal detection is configurable to use any number of ticks, but probably four to eight will work best. I made the inclusion of volume in the reversal logic optional (off by default) and left the existing SMA input found in synapticEx's code intact, albeit with a lower default. With the addition of trend lines, I found volume hindered identification of reversals, although I could try various other filters than the SMA included originally. I have also left intact the very nice ability to change the period and use the requested period identify reversals, courtesy of synapticEx. This could be used in a strategy, as the values plotted are actual values that are available to include in logic and do not include knowledge of the future. However , information is not available until the floor of half the number of ticks used in reversal detection (I then offset by that number to line things up visually). Having never heard of it until now, I just Googled the Bill Williams Alligator strategy, which looks interesting, so maybe I could see how this could be ported to that. ***As I typed this, I remembered that while making reversal detection configurable, I changed the detection logic simply to look for highest (or lowest) of the desired length of ticks. I don't know whether this is not strictly fractal anymore, but if desired, with a little work, I could make it require consecutive, consistent changes before and after each reversal again. Here are a few screenshots from hourly ticks, using the "current" (hourly) period, with and without volume, and playing with the number of points used to identify reversals. Not using volume Using volume
Pineスクリプト™ インジケーター
Lemrinの投稿
8